# Signal processing device and method, encoder and encoding method, decoder and decoding method and programme

FIELD: physics, computer engineering.

SUBSTANCE: present invention relates to signal processing means. An encoder sets an interval including 16 frames as interval section to be processed, outputs high-frequency band encoded data to obtain the high-frequency band component of an input signal and low-frequency band encoded data obtained by encoding the low-frequency band signal of the input signal for each section to be processed. In this case, for each frame, a coefficient used in estimation of the high-frequency band component is selected and the section to be processed is divided into continuous frame segments including continuous frames from which the coefficient with the same section to be processed is selected. In addition, high-frequency band encoded data are produced which include data including information indicating the length of each continuous frame segment, information indicating the number of continuous frame segments included in the section to be processed and a coefficient index indicating the coefficient selected in each continuous frame segment.

EFFECT: improved sound quality with frequency band expansion.

23 cl, 51 dwg

The technical field to which the invention relates

The present invention relates to an apparatus and signal processing method, signal processing, encoder and method of encoding, the decoder and decoding method, and program, and more specifically, to a device and signal processing method, signal processing, encoder and method of encoding, the decoder and decoding method, and program for reproducing a music signal with improved sound quality by extending the frequency range.

The level of technology

Recently expanded services for the distribution of music, designed for music data distribution via the Internet. Service distribution distributes music, as music data, encoded data obtained by encoding music signals. As a method of encoding a music signal is commonly used method of encoding in which the file size of encoded data is compressed to reduce the bit rate, in order to save time during startup.

This way of encoding music signals can be widely divided into encoding method such as MP3 (MPEG (expert Group in the field of moving images), (Audio sound level 3) (international standard ISO/IEC 11172-3) and this encoding, as a non-AA� (high-performance MPEG4 AAC) (international standard ISO/IEC 14496-3).

In the encoding presented MP3, remove the signal component of the high frequency band (below called high band) above approximately 15 kHz or higher in the musical signal, which is almost imperceptible to humans, and encode the low frequency band (below, called the low band component of the rest of the signal. Therefore, the method of encoding is referred to as the method of coding with the removal of the high-frequency band. This type of encoding with the removal of the high-frequency band allows to suppress the file size of encoded data. However, since the sound in the high band to some extent may be perceived by the person, if the sound receiving and outputting decoded from music signal obtained by decoding encoded data, there is a loss of sound quality, so that one loses the sense of realism of the original sound, and the sound's quality, such as blurring of the sound.

In contrast, in the method of coding presented NON-AAS, allocate specific information component of the signal from the high-frequency band and encode this information in conjunction with the component of the low-frequency band signal. The code below is called the encoding characteristics of the high-frequency band. Since the encoding characteristics� the high-frequency band only encode the information of the characteristics of the component of the high-frequency band signal, as information about the component of the high-frequency band signal, decreases the deterioration of sound quality, and can be improved coding efficiency.

When decoding data encoded by the encoding characteristics of the high-frequency band, decode component of the low-frequency band signal and information characteristics, and a component of the high-frequency band signal component derived from the low-frequency band signal and information characteristics after decoding. Accordingly, a technology that extends the frequency band component of the high-frequency band signal, forming component of the high-frequency band signal component of the signal from the low-frequency band, called expansion technology bands.

As an example of application of the method of extension of the strip, after decoding the data encoded by the encoding by removing the high-frequency band, and perform subsequent processing. In the post-processing component of the high-frequency band signal lost when encoding, generate the decoded low-frequency component of the signal strip, expanding, thus, the frequency band component of the low-frequency band signal (see Patent document 1). Way to expand the bandwidth of prior art is called below expands�the access lanes in accordance with Patent document 1.

In the method of extension of the strip in accordance with the Patent document 1, the device performs an evaluation of the power spectrum (below, respectively, is called the frequency envelope of the high-frequency band) to the high-frequency band of the power spectrum of the input signal, by setting the component of the low-frequency band signal after decoding, as an input signal, and forms a component of the high-frequency band signal having a frequency envelope of the high frequency band component of the signal from the low-frequency band.

Fig.1 illustrates an example of a power spectrum of the low-frequency band after decoding, as an input signal, and the frequency of the envelope estimates the high-frequency band.

Fig.1 on the vertical axis illustrates the power of the logarithm, and the horizontal axis illustrates the frequency.

The device defines the bandwidth in the lower band component of the signal the high-frequency band (below referred to as the initial strip the extension from a kind of coding system for input and information, such a sampling rate, frequency, bits, etc. (below called the information side). Further, the device divides the input signal, as a component of the low-frequency band signal, multiple signals podology. The device has multiple signals popolos after the split, ie, gets the average value of the groups concerned (below called the capacity of the group) in the time direction of each cardinality signals popolos on the side of the lower band, lower than the bandwidth of the beginning of the extension (below, simply called the low-frequency band side). As shown in Fig.1, in accordance with the device, it is assumed that the average value of the respective capacities of the group of signals of multiple popolos on the side of the lower band represents the power, and the point that makes the frequency of the lower end of the strip of the expansion rate, is your starting point. The device performs an evaluation of a primary straight line with a given slope passing through a start point, as the frequency envelope of the high-frequency band higher than the band of the expansion (below, simply called the high-frequency band side). In addition, the position power of the initial point in the direction can be adjusted by the user. The device generates each of the plurality of signals podology on the side of the high-frequency band of the plurality of signals podology on the side of the low-frequency band as the estimation of the frequency of the envelope on the side of the high-frequency band. The device comprises multiple received signals podology on the side of the high-frequency band with each other, the receiving component�you of the high-frequency band signals, and summarizes components of the low-frequency band signals with each other to output the summed component signal. Therefore, the music signal after the expansion of the frequency band close to the original musical signal. However, it is possible to generate the musical signal with the best quality.

Means for expanding the strip, disclosed in Patent document 1, has the advantage consisting in the fact that the bandwidth can be extended to a musical signal after decoding encoded data, taking into account different ways of encoding with the removal of the high frequency bands and encoded data with different speeds in bits per second.

The list of references

Patent document

Patent document 1: Laid out in the application for Japanese patent No. 2008-139844

Disclosure of the invention

The problem solved by the invention

Accordingly, the method of extension of the strip, disclosed in Patent document 1 may be improved compared to that of the estimated frequency envelope on the side of the high-frequency band is a primary straight line with a given slope, i.e., the shape of the frequency envelope is fixed.

In other words, the power spectrum of a musical signal has different forms and often appears of the music signal in which the frequency envelope on the side in�pass band, evaluated by way of extension of the strip as disclosed in Patent document 1, deviates significantly.

Fig.2 illustrates an example of the original power spectrum of a musical signal (offensive music signal) that has a fast change over time, such as when strong single stroke on a drum.

In addition, in Fig.2 also illustrates the frequency envelope on the side of the high-frequency band, measured at the input signal, by setting the component of the signal on the side of attacking the low-frequency band signal with respect to a music signal used as an input signal, using the method of extension of the strip, as disclosed in Patent document 1.

As shown in Fig.2, the power spectrum of the original side of the high-frequency band attacking a musical signal has an essentially flat shape.

In contrast, assessment of the frequency of the envelope on the side of the high-frequency band has a predetermined negative slope, and even if the frequency be adjusted so that it had a capacity close to the original power spectrum, the distinction between power and the original spectrum becomes significant as the frequency becomes high.

Accordingly, in the method of extension of the strip disclosed in Patent document 1 estimated frequency bending�sort of on the side of the high-frequency band cannot reproduce the frequency envelope of the original side of the high-frequency band with high accuracy. Therefore, if the sound from the music signal after the expansion of the frequency band will be formed and displayed, the purity of the sound in the audience will be lower than the original sound.

In addition, the encoding characteristics of the high-frequency band, such as non-AAS, etc., as described above, the frequency envelope on the side of the high-frequency band is used as information characteristics of encoded components of the high-frequency band signal. However, it is necessary to reproduce the frequency envelope of the original side of the high-frequency band with high precision on the side of decoding.

The present invention was made taking account of such circumstances and provides a music signal having a better quality sound as a result of widening of the frequency band.

The solution of tasks

The device signal processing, in accordance with the first aspect of the present invention includes: a demultiplexing module that demultiplexes input encoded data into data including information on a segment including frames in which the same coefficient as the coefficient used in the formation of the high-frequency band signal, is chosen at the area to be treated and including a plurality of frames, and information about the ratio for obtaining a friction coefficient close to�the one selected in frames of the segment, and encoded data of the low-frequency band; a module for decoding the low-frequency band, which decodes encoded data of low-frequency bands for the formation of the low-frequency band signal; a selection module that selects a coefficient for the frame to be processed, of the plurality of coefficients based on said data; a module for computing power podology the high-frequency band, which calculates the power podology the high-frequency band signal for podology the high-frequency band in each of popolos constituting the high-frequency band signal of the frame to be processed, on the basis of the signal podology the low-frequency band of each of popolos constituting the low-frequency band signal of the frame to be processed, and the selected coefficient; and the module of formation of the high-frequency band signal, which generates a high-frequency band signal for a frame to be processed, based on the capacity podology the high-frequency band and signal podology the low-frequency band.

The area to be processed is divided into segments so that the provisions of the frames adjacent to each other, in which selected different factors are set as boundary positions of the segments, and information indicating the length of each segment, set�introduced as information on the segments.

The area to be processed is divided into several segments having the same length so that the length of the segment was the largest, and information indicating the length and information indicating whether the selected coefficient before and after each boundary position of the segments is set as information on the segments.

When the same coefficient is selected for several segments in a row, the data can include one item of information on the coefficient for obtaining the coefficient selected in several segments in a row.

The data generated for each site you want to process, using the system from the first system and the second system having a smaller amount of data, wherein in the first system area to be processed is divided into segments so that the provisions of the frames adjacent to each other, in which selected different factors are set as boundary positions of the segments, and information indicating the length of each of the segments is set as information on the segments, and in the second system area to be processed is divided into multiple segments of equal length so as to segment length was greatest, and information indicating the length and information indicating a difference between the selected coefficients before and after the boundary�about the position of the segments, set as information on the segment, wherein the data further includes information indicating whether the data received by the first system or the second system.

The data may further include information re-use, indicating whether the coefficient of the original frame on the site subject to treatment, with a coefficient of a frame directly before the original frame, and when the data includes information reuse, pointing to the coincidence of two factors, the data do not include information on the ratio of the initial plot segment to be processed.

When the assigned mode that reuses rate information, the data includes information re-use, and when the appointed mode in which the reuse of the information ratio is prohibited, the data do not include information re-use.

The signal processing method or program according to the first aspect of the present invention include the following steps: demultiplexing input coded data on the data including information on a segment including frames in which the selected one and the same coefficient as a coefficient used in Fort�the probing of the high-frequency band signal, on the section to be processed including a plurality of frames, and information about the coefficient for obtaining the coefficient selected for the segment frames, and encoded data of the low-frequency band; decode the encoded data of the low-frequency band to obtain the low-frequency band signal; selects a coefficient of a frame to be processed, of the plurality of coefficients based on said data; calculate power podology the high-frequency band signal podology the high-frequency band for each of popolos constituting the high-frequency band signal of the frame to be processed, on the basis of the signal podology the low-frequency band for each podology constituting the low-frequency band signal of the frame to be processed, and the selected coefficient; and form a high-frequency band signal of the frame to be processed, based on the capacity podology the high-frequency band and signal podology the low-frequency band.

In the first aspect of the present invention, input encoded data demultiplexer on data including information on a segment including frames in which the same coefficient as a coefficient used for generating the high-frequency band signal, on the site subject to treatment, including �you a lot of frames and information on the coefficient for obtaining the coefficient selected in the specified segment frames, and encoded data of the low frequency band encoded data to decode the low-frequency band to obtain a low-frequency signal, a coefficient of a frame to be processed, is selected from the set of coefficients on the basis of these data, power podology the high-frequency band signal podology the high-frequency band for each podology constituting the high-frequency band signal in a frame to be processed is calculated on the basis of the signal podology the low-frequency band of each podology constituting the low-frequency band signal of the frame to be processed and the selected coefficient, and the high-frequency band signal of the frame to be processed, form-based power podology the high-frequency band and signal podology the low-frequency band.

The device signal processing in accordance with the second aspect of the present invention includes: a separation module popolos, which generates signal podology the low-frequency band from the set popolos on the side of the low-frequency band of the input signal and the signal podology the high-frequency band from the set popolos on the side of the high-frequency band input signal; a calculation module pseudomona�ti podology the high-frequency band, which calculates pseudomodest podology the high-frequency band, representing the appraised value of the power signal podology the high-frequency band, signal-based podology the low-frequency band and a predetermined coefficient; a selection module that selects any one factor from the set of coefficients for respective frames of the input signal by comparing the power podology the high-frequency band signal podology the high-frequency band and pseudomodest podology the high-frequency band; and a generation module that generates data including information on a segment having frames in which the selected one and the same coefficient on the section to be processed having a plurality of frames of the input signal, and information on the coefficient for obtaining the coefficient selected in frames of the specified segment.

Generation module divides the area to be treated, into segments so that the provisions of the frames adjacent to each other, in which selected various factors that are set as boundary positions of the segments, and establish information that indicates the length of each segment, as the information about the segment.

Generation module divides the area to be treated, into several segments of equal length so that the length with�of Genta was the greatest, and information indicating the length and information indicating whether the selected coefficient before and after the boundary positions of the segments are set as information on the segments.

Generation module generates data that includes one item of information on the coefficient for obtaining the coefficient selected in several segments in a row, when the same coefficient is selected in several segments in a row.

Generation module generates data for each site to be treated, in the system from the first system and the second system having a smaller amount of data, wherein in the first system, the section to be processed is divided into segments so that the provisions of the frames adjacent to each other, in which selected various factors that are set as boundary positions of the segments, and information indicating the length of each of the segments is set as information on the segments, and in the second system area to be processed is divided into several segments of equal length, to the length of the segment was the largest, and the information indicating the length and information indicating whether the selected coefficient before and after a boundary position of the segments is set as information on the segments.

The data may further include information�ation, indicates whether the data of the first system or the second system.

Generation module generates data, which includes information re-use, indicating whether the coefficient of the initial frame of the area to be treated, with a coefficient of a frame directly before the initial frame, and when the information re-use, indicating that the coefficients are the same, included in the data data are formed, which are not included in the rate of the initial plot segment to be processed.

When you have selected re-use the information about the coefficient generation module generates data, which includes information re-use, and when the appointed regime of prohibition of re-use of information on the coefficient generation module generates data, which includes information re-use.

The signal processing method or program in accordance with the second aspect of the present invention include the following steps: forming a signal podology the low-frequency band from the set popolos on the side of the low-frequency band of the input signal, and the signal podology the high-frequency band from the set popolos on the side of the high-frequency band input signal; calculate pseudoman�there podology the high-frequency band, representing the appraised value of the power signal podology the high-frequency band, signal-based podology the low-frequency band and a predetermined coefficient; choose either set of coefficients for respective frames of the input signal by comparing the power podology the high-frequency band signal podology the high-frequency band and pseudomodest podology the high-frequency band; and generate the data including information on a segment having frames in which the selected one and the same coefficient on the section to be processed having a plurality of frames of the input signal and the coefficient for obtaining the coefficient selected in frames of the segment.

In a second aspect, the present invention provides a signal podology the low-frequency band consisting of a plurality of popolos on the side of the low-frequency band of the input signal, and the signal podology the high-frequency band consisting of a plurality of popolos on the side of the high-frequency band of the input signal, compute pseudomodest podology the high-frequency band as estimates of signal power podology the high-frequency band based on the signal podology the low-frequency band and a predetermined coefficient, any of the plurality of coefficients for the respective frame� input signal is chosen by comparing the power podology the high-frequency band signal podology the high-frequency band and pseudomodest podology the high-frequency band, and form information on a segment having frames in which the same coefficient is selected on the section to be processed having a plurality of frames of the input signal and the coefficient for obtaining the coefficient selected in frames of the segment.

The decoder, in accordance with a third aspect of the present invention includes: a demultiplexing module that demultiplexes input encoded data to data including information on a segment including frames in which the same coefficient as a coefficient used in the formation of the high-frequency band signal, on the section to be processed including a plurality of frames, and information about the coefficient for obtaining the coefficient selected in frames of the segment, and encoded data of the low-frequency band; a module for decoding the low-frequency band, which decodes encoded data of low-frequency bands for the formation of the low-frequency band signal; a selection module that selects a coefficient of a frame to be processed, of the plurality of coefficients based on said data; a module for computing power podology the high-frequency band, which calculates the power podology the high-frequency band signal podology the high-frequency band d�I each podology, includes a high-frequency band signal of the frame to be processed, on the basis of the signal podology the low-frequency band of each podology constituting the low-frequency band signal of the frame to be processed and the selected coefficient; the module of formation of the high-frequency band signal, which generates a high-frequency band signal of the frame to be processed, based on the capacity podology the high-frequency band and signal podology the low-frequency band; and a synthesis module that synthesizes the low-frequency band signal and the high-frequency band signal to generate an output signal.

Method of decoding according to the third aspect of the present invention includes stages: demultiplexing input coded data on the data including information on a segment including frames in which the same coefficient as a coefficient used in the formation of the high-frequency band signal, on the site subject to treatment, including a plurality of frames, and information about the coefficient for obtaining the coefficient selected in frames of the segment, and encoded data of the low-frequency band, decodes the encoded data of the low-frequency bands for the formation of the low-frequency band signal, selects a coefficient of a frame�, want to process, of the plurality of coefficients based on the data, calculate the power podology the high-frequency band signal podology the high-frequency band of each of popolos, including the high-frequency band signal of the frame to be processed, on the basis of the signal podology the low-frequency band of each podology, including the low-frequency band signal of the frame to be processed and the selected coefficient, form the high-frequency band signal of the frame to be processed, based on the capacity podology the high-frequency band and signal podology the low-frequency band, and synthesize the low-frequency band signal and the high-frequency band signal to generate an output signal.

In the third aspect of the present invention, input encoded data demultiplexer on data including information on a segment including frames in which the same coefficient as a coefficient used in the formation of the high-frequency band signal, on the section to be processed including a plurality of frames, and information about the coefficient for obtaining the coefficient selected in frames of the segment, and encoded data of the low frequency band encoded data to decode the low-frequency band for the bulk� low-frequency signal, the coefficient of a frame to be processed, is selected from the set of coefficients on the basis of these data, power podology the high-frequency band signal podology the high-frequency band of each podology constituting the high-frequency band signal in a frame to be processed is calculated on the basis of the signal podology the low-frequency band of each podology constituting the low-frequency band signal of the frame to be processed and the selected coefficient, and the high-frequency band signal of the frame to be processed, is formed on the basis of power podology the high-frequency band and signal podology the low-frequency band, and synthesize the low-frequency band signal and the high-frequency band signal to generate an output signal.

The encoder in accordance with the fourth aspect of the present invention includes: a separation module popolos, which generates signal podology the low-frequency band from the set popolos on the side of the low-frequency band of the input signal, and the signal podology the high-frequency band from the set popolos on the side of the high-frequency band input signal; a module for calculating pseudomodest podology the high-frequency band, which calculates pseudomodest podology the high-frequency band, representing the estimated value of mo�ity of signal podology the high-frequency band, on the basis of the signal podology the low-frequency band and a predetermined coefficient; a selection module that selects any of the plurality of coefficients for respective frames of the input signal by comparing the power podology the high-frequency band signal podology the high-frequency band and pseudomodest podology the high-frequency band; module encoding the high-frequency band, which generates encoded data of the high-frequency band by encoding information on a segment having frames in which the selected one and the same coefficient on the section to be processed including a plurality of frames of the input signal, and information about the factor intended for obtaining the coefficient selected in frames of the segment; the module encoding the low-frequency band, which encodes the low frequency signal band of the input signal and generates encoded data of the low-frequency band; and a multiplexing module that generates an output code string by multiplexing encoded data of the low-frequency band and high frequency encoded data strip.

A method of encoding in accordance with a fourth aspect of the present invention includes: forming a signal podology the low-frequency band from the set popolos on the low frequency side of the floor�input signal si and the signal podology the high-frequency band from the set popolos on the side of the high-frequency band of the input signal, calculate pseudomodest podology the high-frequency band, representing the appraised value of the power signal podology the high-frequency band, signal-based podology the low-frequency band and a given rate, choose the set of coefficients for respective frames of the input signal by comparing the power podology the high-frequency band signal podology the high-frequency band and pseudomodest podology the high-frequency band, and generate encoded data of the high-frequency band by encoding information on a segment including frames in which the selected one and the same coefficient on the section to be processed including a plurality of frames of the input signal, and information on the coefficient for obtaining the coefficient selected in frames of the segment, encode the low-frequency band signal of the input signal to form coded data of the low-frequency band, and form the output code string by multiplexing encoded data of the low-frequency band and high frequency encoded data strip.

In a fourth aspect of the present invention provide a signal podology the low-frequency band from the set popolos on the side of the low-frequency band of the input signal and the signal podology the high-frequency band �W many popolos on the side of the high-frequency band of the input signal, calculate pseudomodest podology the high-frequency band, representing the appraised value of the power signal podology the high-frequency band, signal-based podology the low-frequency band and a given rate, choose the set of coefficients for respective frames of the input signal by comparing the power podology the high-frequency band signal for podology the high-frequency band and pseudomodest podology the high-frequency band, forming the coded data of the high-frequency band by encoding information on a segment including frames in which the selected the same ratio, and information ratio, for obtaining the coefficient selected in frames of the segment, encode the low-frequency band signal of the input signal to form coded data of the low-frequency band, and form the output code string by multiplexing encoded data of the low-frequency band and high frequency encoded data strip.

The results of the invention

In accordance with a first embodiment of the fourth variant implementation is possible to reproduce a music signal with high sound quality by extending the bandwidth.

Brief description of the drawings

Fig.1 shows the example on �the expedition of the energy spectrum of the low-frequency band after decoding the input signal and the envelope of the frequency estimation of high-frequency bands.

Fig.2 shows a view illustrating an example of the initial energy spectrum of a musical signal of attack in accordance with the rapid change in time.

Fig.3 shows a block diagram illustrating an example functional configuration of the device extension of the frequency band in the first embodiment of the present invention.

Fig.4 shows a block diagram of the sequence of operations illustrating an example of a process of expanding the frequency band expansion unit bandwidth of Fig.3.

Fig.5 is a view illustrating a layout of the energy spectrum of the signal supplied to the device extension of the frequency band of Fig.3, and placing a bandpass filter on the frequency axis.

Fig.6 is a view illustrating an example illustrating the frequency characteristics of the vocal area and power spectrum estimation the high-frequency band.

Fig.7 is a view illustrating an example of the energy spectrum of the signal supplied to the device extension of the frequency band of Fig.3.

Fig.8 is a view illustrating an example of a vector power after the rise of the input signal of Fig.7.

Fig.9 shows a block diagram illustrating an example functional configuration of the device of the ratio study, designed to explore the factor used in the schema formation�Oia of the high-frequency band signal in the device extension of the frequency band of Fig.3.

Fig.10 shows a block diagram of the sequence of operations describing an example of the process of learning a coefficient device study of the coefficient of Fig.9.

Fig.11 shows a block diagram illustrating an example functional configuration of the encoder in the second embodiment of the present invention.

Fig.12 shows a block diagram of the sequence of operations describing an example of the encoding process by the encoder of Fig.11.

Fig.13 shows a block diagram illustrating an example functional configuration of the decoder in the second embodiment of the present invention.

Fig.14 shows a block diagram of the sequence of operations describing an example of processing of decoding by the decoder of Fig.13.

Fig.15 shows a block diagram illustrating an example functional configuration of the device of the ratio study, designed to explore a representative vector used in the encoding scheme of the high-frequency band of the encoder of Fig.11, and the coefficient estimates of the decoded power podology the high-frequency band used in the circuit of the decoder high-frequency band of the decoder of Fig.13.

Fig.16 shows a block diagram of the sequence of operations describing an example of the process of learning a coefficient device study of the coefficient of Fig.15.

Fig.17 shows, illust�yuushi example encoded strings on the output of the encoder of Fig.11.

Fig.18 shows a block diagram illustrating an example functional configuration of the encoder.

Fig.19 shows a block diagram of the sequence of operations describing the processing of the encoding.

Fig.20 shows a block diagram illustrating an example functional configuration of the decoder.

Fig.21 shows a block diagram of the sequence of operations describing the decoding process.

Fig.22 shows a block diagram of the sequence of operations describing the encoding process.

Fig.23 shows a block diagram of the sequence of operations describing the decoding process.

Fig.24 shows a block diagram of the sequence of operations describing the encoding process.

Fig.25 shows a block diagram of the sequence of operations describing the encoding process.

Fig.26 shows a block diagram of the sequence of operations describing the encoding process.

Fig.27 shows a block diagram of the sequence of operations describing the encoding process.

Fig.28 is a view showing an example of a device configuration of the learning coefficient.

Fig.29 shows a block diagram of the sequence of operations describing the process of learning coefficient.

Fig.30 is a view describing the decrease in the encoding of the string index of the coefficient.

Fig.31 is a view describing the mind�isenia amount of coding the row index of the coefficient.

Fig.32 is a view describing the decrease in the encoding of the string index of the coefficient.

Fig.33 shows a block diagram illustrating an example functional configuration of the encoder.

Fig.34 shows a block diagram of the sequence of operations describing the encoding process.

Fig.35 shows a block diagram illustrating an example functional configuration of the decoder.

Fig.36 shows a block diagram of the sequence of operations describing the decoding process.

Fig.37 is a view describing the decrease in the encoding of the string index of the coefficient.

Fig.38 shows a block diagram illustrating an example functional configuration of the decoder.

Fig.39 shows a block diagram of the sequence of operations describing the encoding process.

Fig.40 shows a block diagram illustrating an example functional configuration of the decoder.

Fig.41 shows a block diagram of the sequence of operations describing the decoding process.

Fig.42 shows a block diagram illustrating an example functional configuration of the encoder.

Fig.43 shows a block diagram of the sequence of operations describing the encoding process.

Fig.44 shows a block diagram illustrating an example functional configuration of the decoder.

Fig.45 shows a block diagram of the sequence of Opera�rd, describing the decoding process.

Fig.46 shows a diagram describing the recirculation index of the coefficient.

Fig.47 shows a block diagram of the sequence of operations describing the encoding process.

Fig.48 shows a block diagram of the sequence of operations describing the decoding process.

Fig.49 shows a block diagram of the sequence of operations describing the encoding process.

Fig.50 shows a block diagram of the sequence of operations describing the decoding process.

Fig.51 shows a block diagram illustrating an example hardware configuration of a computer that executes the program processing in which the present invention is applied.

The implementation of the invention

Variant implementation of the present invention will be described with reference to the drawings. In addition, a description is performed in the following sequence.

1. The first variant of implementation (when the present invention is applied to the expansion unit bandwidth),

2. The second variant of implementation (when the present invention is applied to the encoder and the decoder),

3. The third variant of implementation (when the index factor included in encoded data of the high-frequency band),

4. The fourth variant of implementation (when the difference between the index of the coefficient and pseudoman�STU podology high frequency bands include the encoded data of the high-frequency band),

5. The fifth variant of implementation (when the index ratio is chosen, using the estimated value).

6. Sixth variant of implementation (when the area ratio is common),

7. Seventh variant of implementation (when the encoding of the string index of the coefficient decreases in the time direction using a method of variable length),

8. Eighth variant of implementation (when the encoding of the string index of the coefficient decreases in the time direction using the fixed-length),

9. A ninth variant of implementation (when you choose any of the method or variable length method of a fixed length),

10. The tenth variant of implementation (when carry out recycling information using variable method),

11. Eleventh variant of implementation (when the recycling of information is performed by using method with a fixed length).

1. The first variant of implementation

In the first embodiment of the handle, which extends the frequency range (below called the handling of the extension of frequency band) to a component of the low-frequency band signal after decoding obtained by decoding encoded data using the encoding with the removal of the high-frequency band.

An example of functional co�figuration in the device extension of the frequency band

Fig.3 illustrates an example functional configuration of the device extension of the frequency band in accordance with the present invention.

The device 10 expansion band performs the processing of expanding the frequency band in relation to the input signal, by setting the component of the low-frequency band signal after decoding, as an input signal, and outputs the resulting signal after the process of expansion of bandwidth, as output signal.

The expansion device 10 frequency bands includes a filter 11 low-frequency circuit 12 delays the bandpass filter 13, the circuit 14 for calculating the magnitude characteristics, scheme 15 the power of podology the high-frequency band, the circuit 16 forming the high-frequency band, the filter 17 high frequencies and the adder 18 of the signal.

The filter 11 of the low frequency filters the input signal in a predetermined cutoff frequency and delivers the component of the low-frequency band signal, which is a component of the low-frequency band signal, the signal after filtering in the circuit 12 delays.

Since the circuit 12 delays synced to the summation of each other component of the low-frequency band signal from the filter 11 of the low frequency component of the signal and the high-frequency band, which will be described below, it performs the delay welcomenote of the low-frequency band signal at a certain time and a component of the low-frequency band signal fed into the adder 18 of the signal.

Bandpass filter 13 includes bandpass filters 13-1-13-N, having a bandwidth that is different from each other. Bandpass filter 13-i ((≤i≤N)) transmits a signal in a predetermined bandwidth of the input signal and delivers the missed signal, one of a plurality of signals podology in scheme 14 for calculating the magnitude characteristics and circuit 16 forming the high-frequency band signal.

Circuit 14 for calculating the magnitude characteristics calculates one or more values of characteristics, using at least one of a plurality of signals popolos and the input signal from the bandpass filter 13, and delivers the calculated values of the characteristics in scheme 15 the power of podology the high-frequency band. Here the magnitude characteristics represent information representing the feature of the input signal, as a signal.

Scheme 15 the power of podology the high-frequency band calculates the evaluation value of the power podology the high-frequency band, which represents the signal strength podology the high-frequency band for each podology the high-frequency band based on one or more values of characteristics from the circuit 14 for calculating the magnitude characteristics, and delivers the calculated value of evaluation in scheme 16 for the formation of the high-frequency band signal.

Scheme 16 the formation of si�Nala generates the high-frequency band component of the high-frequency band signal, which is a component of the high-frequency band signal based on the plurality of signals popolos from the bandpass filter 13, and the value estimates for a set of power values podology the high-frequency band from the circuit 15 of the power podology the high-frequency band, and supplies the generated high signal component in the filter 17 high frequency.

The filter 17 of the high frequency component filters of the high-frequency band signal from the circuit 16 forming the high-frequency band signal, using the cut-off frequency corresponding to the cutoff frequency of the filter 11 of low frequency, and delivers the filtered component of the high-frequency band signal in the adder 18 of the signal.

The signal adder 18 adds the component of the low-frequency band signal from the circuit 12 delays and component of the high-frequency band signal from the filter 17 of the high-frequency band and outputs the summed components as output signal.

In addition, in the configuration in Fig.3, for receiving the signal podology, apply a bandpass filter 13, but is not limited to this. For example, you can apply the filter to the separation of the bands, disclosed in Patent document 1.

In addition, similarly, in the configuration shown in Fig.3, the adder 18 of the signal used for the signal synthesis podology, but are not limited to this. For example, can be used with�neticesi filter strips, disclosed in Patent document 1.

The process of expanding the bandwidth of the device extension of the frequency band

Next, with reference to the block diagram of the sequence of operations shown in Fig.4, will be described processing of expanding the frequency band performed by the expansion unit bandwidth of Fig.3.

In step S1, the filter 11 low frequency filters the input signal with a predetermined cutoff frequency and delivers the component of the low-frequency band signal, the signal after filtering to the circuit 12 delays.

The filter 11 of a low frequency may set a frequency as the cutoff frequency. However, in a variant implementation of the present invention, the filter of low frequencies can be set according to the frequency at the lower end of the strip beginning of the extension by setting the reference frequency in the band of expansion, which is described below. Therefore, the filter 11 delivers the low frequency component of the low-frequency band signal, which is a component of the signal more than the low-frequency band than the band of the expansion to the circuit 12 delays the signal quality after filtering.

In addition, the filter 11 of a low frequency can set the optimal frequency as the cutoff frequency, in response to the parameter encoding, such as encoding method with the removal of the high-frequency band, or the speed� transmission bits, etc. of the input signal. As a parameter encoding, for example, can be used information side used in the method of extension of the strip disclosed in Patent document 1.

In step S2, the circuit 12 performs the delay only delay component of the low-frequency band signal by a certain delay time from the filter 11 low frequency and delivers the delayed component of the low-frequency band signal in the adder 18 of the signal.

In step S3, the bandpass filter 13 (bandpass filters 13-1-13-N) divide the input signal into multiple signals popolos and feeds each of the plurality of signals popolos after the split in the circuit 14 for calculating the magnitude characteristics and circuit 16 forming the high-frequency band signal. In addition, the processing of dividing an input signal using a bandpass filter 13 will be described below.

In step S4, the circuit 14 for calculating the magnitude characteristics calculates one or more values of characteristics using at least one of the plurality of signals popolos from the bandpass filter 13 and the input signal, and delivers the calculated values of the characteristics in scheme 15 the power of podology the high-frequency band. In addition, the process of calculation for the magnitude characteristics of the circuit 14 for calculating the magnitude characteristics will be described in detail below.

In step S5, the circuit 15.�Ki power podology the high-frequency band calculates the evaluation value of a plurality of power values podology the high-frequency band based on one or more values of characteristics and delivers the calculated value evaluation in scheme 16 for the formation of the high-frequency band signal from the circuit 14 for calculating the magnitude characteristics. In addition, the process of calculating the value of the power podology the high-frequency band by means of the circuit 15 of the power podology the high-frequency band will be described in detail below.

In step S6, the circuit 16 signal generates the high-frequency band component of the high-frequency band signal based on the plurality of signals podology from the bandpass filter 13 and the evaluation value of a plurality of power values podology the high-frequency band from the circuit 15 of the power podology the high-frequency band, and delivers the component formed of the high-frequency band signal in the filter 17 high frequency. In this case, a component of the high-frequency band signal is a component signal of a higher frequency band than the band of the expansion. In addition, processing for forming a component of the high-frequency band signal of the circuit 16 for the formation of the high-frequency band signal will be described in detail below.

In step S7, the filter 17 removes high frequency noises such as noise sampling, in the low-band component included in the high-frequency band signal resulting from filtering a component of the high-frequency band signal, from the circuit 16, the signal vysokochistom�Oh bands, and delivers the component of the high-frequency band signal in the adder 18 of the signal.

In step S8, the adder 18 summarizes signal component of the low-frequency band signal from the circuit 12 delays and component of the high-frequency band signal from the filter 17 of the high frequency with each other, and outputs the total components, as output signal.

In accordance with the above-mentioned processing, the bandwidth can be extended to a component of the low-frequency band signal after decoding.

Next will be presented a description of each process step S3-S6 of the flowchart of the sequence of operations shown in Fig.4.

Description of the process performed by bandpass filter

First will be described the processing by the bandpass filter 13 in step S3 in the flowchart of the sequence of operations of Fig.4.

In addition, for convenience of explanation, as described below, it is assumed that the number N of bandpass filter 13 is represented by N=4.

For example, it is assumed that one of the 16 popolos obtained by dividing the Nyquist frequency of the input signal into 16 parts, is a strip of the expansion, and each of the 4 popolos a lower band than the band of the beginning of the extension 16 popolos, represents each bandwidth bandpass filters 13-1-13-4.

Fig.5 illustrates the layout to�each frequency axis for each bandpass filter bandpass filters for 13-1-13-4.

As shown in Fig.5, if it is assumed that the index of the first podology from the high-frequency band in the frequency band (podology) lower frequency band than the band began the extension is sb, the second index podology is a sb-1, and the index of the I-th podology is a sb(I-1). Each of bandpass filters 13-1-13-4 assign each podporou in which the index is sb to sb-3 among popolos lower band, which is lower than the original strip of the extension as bandwidth.

In the present embodiment, the implementation of each bandwidth bandpass filters 13-1-13-4 4 is defined podology among 16 popolos obtained by dividing the Nyquist frequency of the input signal into 16 parts, but is not limited to this and may be a 4 set podology 256 popolos obtained by dividing the Nyquist frequency of the input signal into 256 parts. In addition, each bandwidth bandpass filters 13-1-13-4 may differ from each other.

A description of the process performed by the circuit for calculating the magnitude characteristics

Next will be presented a description of the processing performed by the circuit 14 for calculating the magnitude characteristics in step S4 of the flowchart of the sequence of operations shown in Fig.4.

Circuit 14 for calculating the magnitude characteristics of rasschityva�t one or more values of characteristics, used so that the circuit 15 of the power podology the high-frequency band calculates the value of power podology the high-frequency band by using at least one of a plurality of signals podology from the bandpass filter 13 and the input signal.

In more detail, the circuit 14 for calculating the magnitude characteristics calculates, as the magnitude characteristics, signal strength podology (power podology (below is called power podology the low-frequency band)) for each podology among signals 4 popolos bandpass filter 13, and delivers the calculated value of signal power podology in scheme 15 the power of podology the high-frequency band.

In other words, the circuit 14 for calculating the magnitude characteristics calculates the power power (ib,J) podology the low-frequency band in a given time frame J of the 4 signals x (ib,n) podology, which is supplied from the bandpass filter 13, using the following Equation (1). Here ib is an index podology, and n expressed as a discrete time index. In addition, the number of samples of one frame is expressed as FSIZE, and power is expressed in decibels.

Equation 1

Accordingly, power power (ib,J) podology the low-frequency band received by the circuit 14 for calculating the magnitude characteristics, is supplied to the circuit 15 of the power podology the high-frequency band, as the magnitude characteristics.

A description of the process performed by the scheme of assessment of the capacity podology the high-frequency band

Next will be presented a description of the processing performed by the circuit 15 of the power podology the high-frequency band in step S5 block with�status of the sequence of operations in Fig.4.

Scheme 15 the power of podology the high-frequency band calculates the evaluation value of the power podology (power podology the high-frequency band) for the band, band frequency extension), which expand after podology (bands start to diverge), the index is sb+1, based on the 4 values of power podology supplied from the circuit 14 for calculating the magnitude characteristics.

Thus, if the circuit 15 of the power podology the high-frequency band believes that the index podology maximum bandwidth for the frequency band expansion is equal to eb (eb-sb), assess capacity podology against podology in which the index is sb+1 to eb.

In-band frequency extension value assessment power_{est}(ib,J) of capacity podology, an index which represents an ib is expressed by the following Equation (2), using 4 values of power power (ib,J) podology supplied from the circuit 14 for calculating the magnitude characteristics.

Equation 2

Here, in Equation (2), the coefficients a_{ib}(kb) and B_{ib}are coefficients having a value different for the respective podology ib. The coefficients a_{ib}(kb), B_{ib}are rates set accordingly to obtain the corresponding values with respect to different input signals. In addition, the coefficients a_{ib}(kb), B_{ib}also charged to the optimal value by changing podology sb. Conclusion A_{ib}(kb), B_{ib}will be described below.

In Equation (2), the value of evaluation mo�ness podology the high-frequency band calculated by linear primary combination, using the power of each of the plurality of signals podology from the bandpass filter 13, but is not limited to this and, for example, it can be calculated using a linear combination of a plurality of power values popolos the low-frequency band of the frames before and after a temporary J frame, and can be calculated using a nonlinear function.

As described above, the value of the cardinality estimation podology the high-frequency band, calculated in the evaluation scheme 15 power podology the high-frequency band, is fed into a circuit 16 for the formation of the high-frequency band signal, which will be described later.

A description of the process performed by the scheme of formation of the high-frequency band signal

Next will be presented a description of the processing performed by the circuit 16 forming the high-frequency band signal in step S6 in the flowchart of the sequence of operations of Fig.4.

Scheme 16 the formation of the high-frequency band signal calculates the value of power power (ib,J) podology the low-frequency band of each podology based on Equation (1) described above from the set of signals podology supplied from the bandpass filter 13. Scheme 16 the formation of the high-frequency band signal gets the value of G (ib,J) of the gain according to Equation 3 described below, using many of the calculated power values (b,J) podology the low-frequency band,
and power values power_{est}(ib,J) of the assessment for power podology the high-frequency band, calculated on the basis of Equation (2) described above according to scheme 15 the power of podology the high-frequency band.

Equation 3

Here, in Equation (3), sb_{map}(ib) before�provide index podology original maps for the case
when popoloca ib is regarded as popoloca original maps, and is expressed by the following Equation 4.

Equation 4

In addition, in Equation (4), INT (a) is a function that discards the decimal point for the value.

Further, the circuit 16 signal calculates the high-frequency band signal podology x2 (ib,n) after adjusting the gain by multiplying G (ib,J) is the gain obtained by using Equation 3, the output of the bandpass filter 13, using the following Equation (5).

Equation 5

In addition, the circuit 16 signal calculates the high-frequency band signal X3 (ib,n) podology after adjusting the gain, which is the result of a cosine transform of the signal x2 (ib,n) podology after adjusting the gain, by performing a cosine transform in the frequency corresponding to the frequency of the upper end podology with the index of sb, from the frequency corresponding to the frequency of the lower end of podology having the index is sb-3, according to the following Equation (6).

Equation 6

In addition, in Equation (6), π is a constant circle. Equation (6) means that the signal x2 (ib,n) podology after adjusting the gain is shifted on the frequency of each 4-piece stripe on the sides of the high-frequency band.

Therefore, the circuit 16 forming the high-frequency band signal component calculates x_{high}(n) of the high-frequency band signal from the signal X3 (ib,n) podology after adjusting the gain with a shift to the side of the high-frequency band, in accordance with the following Equation 7.

Equation 7

Accordingly, a component of the high-frequency band signal shape circuit 16 forming the high-frequency band signal based on the 4 values of power podology the low-frequency band, obtained on the basis of 4 signals podology from the bandpass filter 13, and the value of the cardinality estimation podology the high-frequency band from the circuit 15 of the power podology the high-frequency band, and the resulting component of the high-frequency band signal fed into the filter 17 high frequency.

In accordance with the processing described above, since the power podology the low-frequency band, calculated from the set of signals podology, is set as the value of the characteristics relative to the input signal obtained after decoding encoded data using the encoding with the removal of the high-frequency band, the value of the power podology the high-frequency band is calculated based on the coefficient set soo�appropriate way for her and the component of the high-frequency band signal is formed from adaptive value cardinality estimation podology the low-frequency band and power podology the high-frequency band, as a result, it becomes possible to perform the estimation of power podology for bandwidth expansion frequency with high accuracy and to reproduce the music signal with the best sound quality.

As described above, the circuit 14 for calculating the magnitude characteristics illustrates an example in which calculated as the amount of features that only power podology the low-frequency band, calculated from the set of signals podology. However, in this case, power podology frequency band expansion cannot be estimated with high accuracy based on the input signal of this kind.

Here power evaluation podology frequency band extension in scheme 15 the power of podology the high-frequency band can be performed with high accuracy because the circuit 14 for calculating the magnitude features calculates the value of the characteristic having a strong correlation with output system power podology frequency band extension (shape of the power spectrum the high-frequency band).

Another example of the value characteristics as calculated by the scheme of calculating the magnitude characteristics

Fig.6 illustrates an example velicina.stanovi characteristics of the area of vocals, where most of the vocals are busy, and the power spectrum of the high-frequency band obtained by evaluating the power podology the high-frequency band, the result of the calculation only power podology the low-frequency band, as the magnitude of the characteristics.

As shown in Fig.6, in the frequency region of vocals, there are many cases where the estimation of the power spectrum the high-frequency band has a higher position than the power spectrum of the high-frequency band of the original signal. Because the sense of incompatibility of voices singing people easily perceived by the human ear, in the area of vocals, you must estimate power podology the high-frequency band with high accuracy.

In addition, as shown in Fig.6, the frequency characteristic of the area of vocals, there are many cases where a large deflection is from 4.9 kHz to 11,025 kHz.

Here, as described below, will be described an example that can apply a degree of deflection in the area from 4.9 kHz to 11,025 kHz in the frequency field, as the value of the characteristics used in estimating the power podology the high-frequency band area of vocals. Further, the amount of characteristics representing the degree of deflection is called the depth below.

An example of the calculation of the depth during the time frame J dip (J) will be described below.

Quick conversion�their Fourier transform (FFT) for 2048 points are performing against signals 2048 segments of the sample, included in the range of several frames before and after the temporary frame J input, and I expect the coefficients on the frequency axis. The power spectrum obtained by performing db transformation in terms of the absolute value of each of the calculated coefficients.

Fig.7 illustrates one example of a power spectrum obtained in the way described above. Here to remove the minor component of the power spectrum, for example, to delete a component 1.3 kHz or less, perform the upgrade process. If you do the upgrade process, it becomes possible to smooth out minor component peak of the spectrum by selecting each size of the power spectrum and performing the filtering process with the filter of low frequencies in accordance with a time sequence.

Fig.8 illustrates an example power spectrum of the input signal after lifting. In the spectrum of power after the restoration shown in Fig.8, the difference between the minimum value and the maximum value included in the range corresponding to 4.9 kHz 11,025 kHz, is set as the depth of the dip (J).

As described above, the value is calculated characteristics that have a strong correlation with the power podology for frequency band extension. In addition, an example of the calculation of the depth of the dip (J) is not limited as described above, and can be done�n another way.

Next will be described another example of the calculation of the amount of characteristics that have a strong correlation with the power podology for frequency band extension.

Another another example of the value characteristics as calculated by the scheme of calculating the magnitude characteristics

In terms of frequency characteristics in the area of the attack, which is an area that includes the audio signal of attack type in any input, there are often cases where the power spectrum of the high-frequency band is essentially flat, as described with reference to Fig.2. It is difficult for the method that calculates, as the magnitude characteristics, only the power podology the low-frequency band, to perform with high accuracy estimate of the capacity of podology virtually flat frequency band expansion, which can be seen in the area of the attack, to assess the capacity podology in the frequency band expansion size characteristics, indicating the variation of time with a specific input signal, which includes the area of the attack.

Here below will be described an example that uses a variation of time power podology the low-frequency band, as the amount of characteristics used to assess the capacity podology the high-frequency band in the area of the attack.

Vibration time power

Equation 8

In accordance with Equation 8, the variation in time of power_{d}(J) power podology the low-frequency band represents the ratio between the sum of the four values of the power podology the low-frequency band during the time frame J-1 and the sum of the four values of the power podology the low-frequency band during the time frame (J-1) one frame before the time frame J, and if this value becomes larger, the time variation of capacity between frames will be large, i.e., the signal included in the time frame J, is considered as having a stronger attack.

In addition, if the power spectrum illustrated in Fig.1, which is a statistically average, compared with the power spectrum of the field of attack (music signal type of attack), shown in Fig.2, the power spectrum in the area of attack increases towards the right in the middle lane. Between areas of attack are many cases that represent the frequency characteristics.

Accordingly, the following describes an example in which used slope in the middle lane, as the amount of characteristics is used to assess the capacity podology high�castetnau strip between areas of attack.

The slope of the slope (J) of the middle band in some time frames J, for example, is obtained from the following Equation (9).

Equation 9

In Equation (9) the factor w (ib) represents a weighting factor that adjusted for the possibility of its weighing capacity podology the high-frequency band. In accordance with Equation (9), the slope (J) is the ratio of the sum of the four values of the power podology the low-frequency band to the high band and the sum of the four values of the power podology the low-frequency band. For example, if the four values of power podology the low-frequency band will be set as a power in respect of podology middle band, the slope (J) is of great importance, when the power spectrum in the middle lane increases to the right, and the power spectrum has a smaller value when the power spectrum decreases to the right.

Since there are often cases where the average slope of the strip is substantially changed before and after footage of the attack, we can assume that the variation in time slope_{d}(J) for slope, expressed by the following Equation (10), represents the proportion of the characteristics used in estimating the power podology the high-frequency band in the area of the attack.

EQ�of 10

In addition, it can be assumed that the variation in time of the depth of the dip_{d}(J) described above, which is expressed by the following Equation (11), represents the proportion of the characteristics used in estimating the power podology the high-frequency band in the area of the attack.

Equation 11

*
F SIZE*

*In accordance with the method described above, since the value is calculated characteristics that have a strong correlation with the power podology for frequency band extension if it is used, the score for power podology frequency band extension in scheme 15 the power of podology the high-frequency band can be performed with high accuracy.*

*As described above, will be presented an example to calculate the magnitude characteristics that have a strong correlation with the power podology for frequency band extension. However, the example of estimating power podology the high-frequency band will be described below using the values of the features calculated by the method described above.*

*A description of the process schema evaluation capacity podology the high-frequency band*

*Here, with reference to Fig.8 will be described an example of estimating power podology the high-frequency band, using the depth and power podology the low-frequency band, ka�the degree of value specifications.*

Thus, in step S4 of the flowchart of the sequence of operations of Fig.4, the circuit 14 for calculating the magnitude features calculates the value characteristics, power podology the low-frequency band and depth, and delivers the calculated value of the power podology the low-frequency band and depth in scheme 15 the power of podology the high-frequency band for each podology of four signals podology from the bandpass filter 13.

Therefore, in step S5, the circuit 15 of the power podology the high-frequency band calculates the evaluation value of the power podology the high-frequency band based on the four power values podology the low-frequency band and the magnitude of the depth from the circuit 14 for calculating the magnitude characteristics.

Here, for power values podology and depth, since the ranges of obtained values (scales) differ from each other, the circuit 15 of the power podology the high-frequency band, for example, performs the following transform on the importance of depth.

Scheme 15 the power of podology calculates the high-frequency band power podology maximum bandwidth for four values of the power podology the low-frequency band and depth values in a predetermined large magnitude of the input signal and obtains the mean value and standard� deviation,
respectively. Here it is assumed that the mean power podology is a power_{ave}, RMS power podology is a power_{std}the average depth value represents a dip_{ave}and standard deviation of depth is a dip_{std}.

Scheme 15 the power of podology the high-frequency band converts the value of the depth of the dip (J), using this value in the following Equation (12) and receives a deep dip after conversion_{s}(J).

Equation 12

When performing the conversion described in Equation (12,
scheme 15 the power of podology the high-frequency band can statistically to convert the value of the depth of the dip (J) to equal a variable (depth) dip_{s}(J) for the mean and the variance of the power podology the low-frequency band and make the range of values obtained from the depth that is approximately equal to the range of values obtained from power podology.

In the frequency range extension, the importance of assessing power_{est}(ib,J) of capacity podology in which the index is ib, is expressed in accordance with Equation 13, by a linear combination of the four power values podology the low-frequency band power (ib,J) from the circuit 14 for calculating the magnitude of the characteristics and values of the depth of the dip_{s}(J) shown in Equation (12).

Equation 13

Here, in Equation (13), the coefficients C_{ib}(kb), D_{ib}E_{ib}are coefficients that have values that are different for each ib podology. The coefficients C_{ib}(kb), D_{ib}and E_{ib}are rates set accordingly, to obtain a favorable values with respect to different input signals. In addition, the coefficients C_{ib}(kb), D_{ib}and E_{ib}also changed to the optimum values to change podology sb. In addition, the output of the coefficients C_{ib}(kb), D_{ib}and E_{ib}will be described below.

In Equation (13), the value of the power podology the high-frequency band is calculated with the help of� linear combination, but are not limited to this. For example, the evaluation value may be calculated using a linear combination of values of multiple features of several frames before and after a temporary J frame, and can be calculated using a nonlinear function.

In accordance with the process described above, it is possible to reproduce a music signal having a better quality, due to the fact that the accuracy of cardinality estimation podology the high-frequency band in the vocal region is improved compared with the case where it is assumed that only power podology the low-frequency band represents the proportion of the characteristics of power podology the high-frequency band, using the values of specific vocal depth of field, as the amount of features, the power spectrum of the high frequency bands to form in his estimation, more than the power spectrum of the high-frequency band of the original signal and feelings of inadequacy can easily be perceived by the human ear, when using the installation method only podology the low-frequency band, as the magnitude of the characteristics.

Therefore, if the number of splits popolos is 16, because the frequency resolution is low, the depth, calculated as the value of characteristics but the way opisan�mu above (degree of concavity in terms of frequency characteristics in the area of vocals), the degree of concavity cannot be expressed only by the output podology the low-frequency band.

Here, the frequency resolution is improved, and it is possible to Express the degree of concavity only for power podology the low-frequency band, so that the number of units popolos increases (for example, 256 division 16 times), the number of divisions on the strip using a bandpass filter 13 increases (for example, 64 to 16 times), and the number of power values podology the low-frequency band calculated by the circuit 14 for calculating the magnitude characteristics increases (64 to 16 times).

If you use only the power podology the low-frequency band is assumed that it is possible to evaluate the power podology the high-frequency band with precision, essentially equal to the power rating podology the high-frequency band used as the value of the characteristics, and depth described above.

However, the amount of calculation increases with the number of divisions popolos, the number of splits lanes and the number of power values popolos the low-frequency band. If it is assumed that the power podology the high-frequency band can be estimated with an accuracy equal to any method, method, which estimates the power podology the high-frequency band, with required�of depth as the amount of characteristics, without increasing the number of splits popolos, consider how effective, from the point of view of computation.

As indicated above, there was described a method that allows to estimate the power podology the high-frequency band, using the depth and power podology the low-frequency band, but as the values of the characteristics used in estimating the power podology the high-frequency band, one or more values of characteristics described above (power podology the low-frequency band, depth, variation in time of power podology the low-frequency band, tilt, variation in time of the slope, and the variation in time of the depth), with no restrictions on the combinations. In this case it is possible to improve the accuracy in assessing the capacity podology the high-frequency band.

In addition, as described above, in the input signal it is possible to improve the accuracy of the evaluation phase, using the specific setting in which the assessment of capacity podology the high-frequency band is difficult to use, as the values of the characteristics used in estimating the power podology the high-frequency band. For example, change time power podology the low-frequency band, slope, change in time of the slope and the change over time of the depth represent specific pairs�Tr in the area of the attack, and can improve the accuracy of cardinality estimation podology the high-frequency band in the area of the attack, using its argument as the value of the characteristics.

Furthermore, even if power evaluation podology the high frequency bands will be performed using different values of the characteristics, except for the power podology the low-frequency band and the depth, i.e., change time power podology the low-frequency band, slope, change in time of the slope and the change in time but the depth, power evaluation podology the high-frequency band can be obtained as in the method described above.

In addition, each method of calculating the characteristics described in the description, is not limited to the method described above, and another method may be used.

The method of obtaining the coefficients C_{ib}(kb), D_{ib}E_{ib}

Next will be described the method to obtain the coefficients C_{ib}(kb), D_{ib}and E_{ib}in Equation (13) described above

Is used a method in which the coefficients are determined on the basis of the results of the study, which is performed by using the instruction signal having a predetermined broad band (below called the broadband signal descriptions) such as a way to obtain the coefficients C_{ib}(kb), D_{ib}and E_{ib}, the coefficients a _{
ib}(kb), D_{ib}and E_{ib}become the corresponding values with respect to different input signals in the evaluation of power podology for frequency band extension.

When performing a study of the coefficients C_{ib}(kb), D_{ib}and E_{ib}device learning coefficient, which includes the bandpass filter having the same bandwidth as bandpass filters 13-1-13-4 described with reference to Fig.5, is used for the high-frequency band that is higher than the band of the source of the. The unit of study factor performs the study, when injected broadband instruction.

An example of a functional configuration of the device for the study of the coefficient

Fig.9 illustrates an example of a functional configuration of the device the study of the factor executes the statement by the coefficients C_{ib}(kb), D_{ib}and E_{ib}.

Component of the low-frequency band signal that is lower than the band of the source of the expansion of the broadband signal instructions supplied to the device 20 of the learning coefficient in Fig.9 represents a signal that has been encoded using the same method as the method of coding performed when encoding an input signal having a limited band of input into the device 10 the extension of the frequency band in Fig.3.

The device 20 study ratio incl�em in the bandpass filter 21, scheme 22 computing power podology the high-frequency band, the circuit 23 for calculating the magnitude characteristics and circuit 24 coefficient estimates.

Bandpass filter 21 includes bandpass filters 21-1-21-(K+N) having a bandwidth that is different from each other. Bandpass filter 21-i (1≤i≤K+N) passes a signal with a given bandwidth of the input signal and delivers the missed signal to the circuit 22 of the computing power podology the high-frequency band or in a circuit 23 for calculating the magnitude characteristics, as one of many signals podology. In addition, band-pass filters 21-1-21k among bandpass filters 21-1-21-(K+N) skip the high-frequency band signal that is higher than the bandwidth of the expansion.

Scheme 22 computing power podology calculates the high-frequency band power podology the high-frequency band of each podology for each permanent of temporary frame in relation to a set of signals podology the high-frequency band of a bandpass filter 21, and delivers the calculated value of the power podology the high-frequency band in scheme 24 the coefficient estimates.

Circuit 23 for calculating the magnitude characteristics calculates the same value characteristics value characteristics calculated by the circuit 14 for calculating the magnitude characteristics of the device 10 the extension of the band of Fig.3 DL� the same respective time frames at regular time points in which I expect power podology the high-frequency band using the schema 22 computing power podology the high-frequency band. Thus, the circuit 23 for calculating the magnitude characteristics calculates one or more values of characteristics, using at least one of a plurality of signals podoley of bandpass filter 21, and the broadband signal instructions, and delivers the calculated value characteristics in scheme 24 the coefficient estimates.

Scheme 24 the coefficient estimates performs the assessment rate (data rate) used in scheme 15 the power of podology the high-frequency band of the device 10 the extension of the band of Fig.3 based on the capacity podology the high-frequency band from the circuit 22 computing power podology the high-frequency band and the magnitude characteristics of the circuit 23 for calculating the magnitude characteristics for each frame in constant time.

The process of learning coefficient in the unit of study coefficient

Next, with reference to the block diagram of the sequence of operations in Fig.10, will be described processing of the learning coefficient in the unit of study coefficient of Fig.9.

In step S11, the bandpass filter 21 divides the input signal (the instruction signal bandwidth extension) to (K+N) signals podology. Bandpass filters 21-1-21-k set si�nals podology the high-frequency band, which is higher than the bandwidth of the source extension in scheme 22 computing power podology the high-frequency band. In addition, the bandpass filters 21-(K+1)-21-(K+N) serves many signals podology for the low-frequency band that is lower than the bandwidth of the source extension in scheme 23 for calculating the magnitude characteristics.

In step S12, the circuit 22 computing power podology calculates the high-frequency band power power (ib,J) podology the high-frequency band of each podology for each frame of constant time in relation to a set of signals popolos the high-frequency band of the bandpass filters 21 (bandpass filter 21-1-21-k). Power power (ib,J) podology receive the high-frequency band by using the aforementioned Equation (1). Scheme 22 computing power podology the high-frequency band delivers the calculated value of the power podology the high-frequency band in scheme 24 the coefficient estimates.

In step S13, the circuit 23 for calculating the magnitude features calculates the value of the characteristics for each of the time frames, how to frame a constant time in which the power podology the high-frequency band is calculated using the schema 22 computing power podology the high-frequency band.

In addition, as described below, in the circuit 14 for calculating the magnitude characteristics of the device 10 is expanded�I band of Fig.3, it is assumed that the four values of power podology and depth of the low-frequency band was calculated as the amount of characteristics, and will be described that four values of power podology and depth of the low-frequency band is calculated in the circuit 23 for calculating the magnitude characteristics of the device 20 learning coefficient in a similar way.

Thus, the circuit 23 for calculating the magnitude characteristics calculates four values of power podology the low-frequency band, using four signal podology the same four signals corresponding popolos supplied to the circuit 14 for calculating the magnitude characteristics of the device 10 the extension of the bandwidth of bandpass filter 21 (bandpass filter 21-(K+1)-21-(K+4)). In addition, the circuit 23 for calculating the magnitude characteristics calculates the depth of the instruction signal bandwidth extension, and calculates the depth of the dip_{s}(J) based on Equation (12) described above. In addition, the circuit 23 for calculating the magnitude characteristics takes the four values of power podology the low-frequency band and the depth of the dip_{s}(J), as the magnitude characteristics in scheme 24 the coefficient estimates.

In step S14, the circuit 24 estimates of the coefficient evaluates the coefficients C_{ib}(kb), D_{ib}and E_{ib}on the basis of a plurality of combinations (eb-sb) capacity podology the high-frequency band, �Taweelah in the same time frames from the circuit 22 computing power podology the high-frequency band and the circuit 23 for calculating the magnitude characteristics and value characteristics (four values power low frequency podology
strip dip and depth_{s}(J)). For example, the circuit 24 estimates of the coefficient determines the coefficients C_{ib}(kb), D_{ib}and E_{ib}in Equation (13), making five characteristics values (four values of power podology the low-frequency band and the depth of the dip_{s}(J)) explanatory variable against one of podology high bands, and making power power (ib,J) podology the high-frequency band, poznavshaya variable and performing the regression analysis using method of least squares.

In addition, of course, the method for estimating the coefficients, C_{ib}(kb), D_{ib}and E_{ib}not limited to the above-mentioned method and you can apply various General methods of identification of the parameter.

In accordance with the processes described above, since the study of the coefficients used in estimating the power podology the high-frequency band is set to perform with the use of a predetermined instruction signal bandwidth expansion, it is possible to obtain the preferred outcome with respect to different input signal applied to the device 10 of the expansion of bandwidth and, thus, it becomes possible to reproduce a music signal having a better quality.

It's also possible to calculate the coefficients A_{ib}(kb) and B_{ib}in the above-mentioned Equations�AI (2) use the method of learning coefficient.

As described above, were described processes of learning coefficient, which assumes that each value cardinality estimation podology the high-frequency band is calculated by linear combination, such as four power values podology the low-frequency band and depth in scheme 15 the power of podology the high-frequency band of the device 10 the extension of the bandwidth.

However, the method for estimating power podology the high-frequency band in scheme 15 the power of podology the high-frequency band is not limited to the example described above. For example, as the circuit 14 for calculating the magnitude characteristics calculates one or more of other values, characteristics, except for the depth (variation time power podology the low-frequency band, the inclination, the time variation of the inclination and the time variation of depth), power podology the high-frequency band can be calculated, can be used a linear combination of a plurality of values of characteristics of a plurality of frames before and after time J frames, or can be used non-linear function. Thus, in the process of learning of the coefficient circuit 24 estimates of the coefficient can count (to study the coefficient under the same conditions as in relation to specifications, time frames and functions used in the case where the hop�the power podology the high-frequency band, using the evaluation scheme 15 power podology the high-frequency band of the device 10 the extension of the frequency range.

2. The second variant of implementation

In the second variant of implementation perform encoding processing and decoding processing in the encoding characteristics of the high-frequency band by using encoder and decoder.

An example of a functional configuration of the encoder

Fig.11 illustrates an example of a functional configuration of the encoder in which the present invention is applied.

The encoder 30 includes a filter 31 low-frequency circuit 32 encoding the low-frequency band, the circuit 33 of the separation popolos, the circuit 34 for calculating the magnitude characteristics, the circuit 35 for calculating pseudomodest podology the high-frequency band, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band, the encoding scheme 37 the high-frequency band, the circuit 38 multiplexing and circuit 39 decodes the low-frequency band.

The filter 31 low frequency filters an input signal using a predetermined cutoff frequency, and sends a signal to the low-frequency band that is lower than the cutoff frequency (below called the low-frequency band signal), the signal after filtering to the circuit 32 of encoding the low-frequency band, the circuit 33 of the separation popolos and circuit 34 for calculating the magnitude characteristics.

Scheme�and 32 encode the low-frequency band encodes the low frequency band signal from the filter 31 low frequency and delivers the encoded data of the low-frequency band, obtained from the result in scheme 38 multiplexing and circuit 39 decodes the low-frequency band.

Scheme 33 division popolos evenly divides the input signal and the low frequency band signal from the filter 31 low frequency on many signals popolos having a predetermined bandwidth, and delivers the separated signals in the circuit 34 for calculating the magnitude of a characteristic or in a circuit 36 for calculating the difference between pseudomodest podology the high-frequency band. In particular, scheme 33 division popolos delivers the many signals podology (below is called signal podology the low-frequency band), the resulting input signals in the low-frequency band, the circuit 34 for calculating the magnitude characteristics. In addition, scheme 33 division popolos signals podology (below is called signal podology the high-frequency band) of the high-frequency band that is higher than the cutoff frequency set by the filter 31 low frequency, among the many signals podology obtained by inputting the input signal to circuit 36 for calculating the difference between pseudomodest podology the high-frequency band.

Circuit 34 for calculating the magnitude characteristics calculates one or more values of characteristics, using any one of a plurality of signals podology signal podology the low-frequency band from the circuit 33 times�t popolos and the low-frequency band signal from the filter 31 low frequency, and delivers the calculated values of the characteristics in the circuit 35 for calculating pseudomodest podology the high-frequency band.

Scheme 35 calculate pseudomodest podology the high-frequency band forms pseudomodest podology the high-frequency band based on one or more values of characteristics from the circuit 34 for calculating the magnitude characteristics and takes the obtained value of pseudomodest podology the high-frequency band in circuit 36 for calculating the difference between pseudomodest podology the high-frequency band.

Circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the difference between pseudomodest podology the high-frequency band, described below, on the basis of the signal podology the high-frequency band from the circuit 33 of the separation popolos and pseudomodest podology the high-frequency band from the circuit 35 for calculating pseudomodest podology the high-frequency band, and supplies the calculated value of the difference between pseudomodest podology the high-frequency band in the encoding scheme 37 the high-frequency band.

Scheme 37 encoding the high-frequency band encodes the difference between pseudomodest podology the high-frequency band from the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band and delivers the encoded data of high-frequency bands obtained from this result�the one in scheme 38 multiplexing.

Scheme 38 multiplexing multiplexes the encoded data of the low-frequency band from the circuit 32 encoding the low-frequency band and high frequency encoded data strip from the circuit 37 encoding the high-frequency band, and outputs as an output code string.

Circuit 39 decodes the low-frequency band appropriately decodes the encoded data of the low-frequency band from the circuit 32 encoding the low-frequency band and delivers the decoded data obtained from this result, in scheme 33 division popolos in scheme 34 for calculating the magnitude characteristics.

Processing coding coder

Next, with reference to the block diagram of the sequence of operations in Fig.12 will be described processing of the encoding performed by the encoder 30 in Fig.11.

In step S111, the filter 31 low frequency filters an input signal using a predetermined cutoff frequency, and supplies this signal low-frequency band as the signal after filtering to the circuit 32 of encoding the low-frequency band, the circuit 33 of the separation popolos and circuit 34 for calculating the magnitude characteristics.

In step S112, the encoding scheme 32 encodes the low-frequency band signal of the low-frequency band of the filter 31 low frequency and delivers the encoded data of the low-frequency band, obtained from the result in figure 3, the multiplexing.

In addition, for encoding the low-frequency band signal in step S112, it is necessary to choose an appropriate encoding method, in accordance with the encoding efficiency, and the obtained value of the scheme, and the present invention does not depend on the encoding.

In step S113, the circuit 33 of the separation popolos equally splits the input signal and the low frequency band signal into many signals podology having a predetermined band width. Scheme 33 division popolos signals podology the low-frequency band obtained by inputting the low-frequency band signal to the circuit 34 for calculating the magnitude characteristics. In addition, scheme 33 division popolos signals podology the high-frequency band to the higher frequency band than the frequency limit of the band, which is set by the filter 31 low frequency among the plurality of signals podology obtained by inputting the input signal to circuit 36 for calculating the difference between pseudomodest podology the high-frequency band.

In step S114 circuit 34 for calculating the magnitude characteristics calculates one or more values of characteristics, using at least any one of a plurality of signals podology signal podology the low-frequency band from the circuit 33 of the separation popolos and the low-frequency band signal from the filter 31 low frequency, and delivers the Russ�Fannie magnitude characteristics in the circuit 35 for calculating pseudomodest podology the high-frequency band. In addition, the circuit 34 for calculating the magnitude characteristics of Fig.11 has basically the same configuration and function as the circuit 14 for calculating the magnitude characteristics of Fig.3. Since the process in step S114 is essentially identical performed in step S4 of the flowchart of the sequence of operations in Fig.4, a description is to be excluded.

In step S115, the circuit 35 for calculating pseudomodest podology the high-frequency band forms pseudomodest podology the high-frequency band based on one or more values of characteristics from the circuit 34 for calculating the magnitude characteristics and delivers the generated value of pseudomodest podology the high-frequency band in circuit 36 for calculating the difference between pseudomodest podology the high-frequency band. In addition, the circuit 35 for calculating pseudomodest podology the high-frequency band in Fig.11 has basically the same configuration and function as that of the scheme 15 the power of podology the high-frequency band in Fig.3. Therefore, since the process in step S115 is essentially identical performed in step S5 of the flowchart of the sequence of operations in Fig.4, a description is to be excluded.

In step S116, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the difference between pseudomodest podology the high-frequency band based on the signal Popolo�s the high-frequency band from the circuit 33 of the separation popolos and pseudomodest podology the high-frequency band from the circuit 35 for calculating pseudomodest podology the high-frequency band and supplies the calculated difference of pseudomodest podology the high-frequency band in the encoding scheme 37 the high-frequency band.

In particular, the circuit 36 for calculating the difference between pseudomodest podology calculates the high-frequency band power power (ib,J) podology (high frequency band) in the frames J continuous-time signal podology the high-frequency band from the circuit 33 of the separation popolos. In addition, in the embodiment of the present invention, all podology signal podology the low-frequency band and podology signal podology the high-frequency band are distinguished by an index ib. Method of calculating the power podology can be applied to the same method as in the first variant of implementation, i.e., the method used in Equation (1).

Further, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates a Delta value (difference of pseudomodest podology the high-frequency band) power_{diff}(ib,J) between power power (ib,J) podology the high-frequency band and pseudomodest power_{lh}(ib,J) podology the high-frequency band from the circuit 35 for calculating pseudomodest podology the high-frequency band in the time frame J. the power Difference_{diff}(ib,J) of Pseudomonas podology the high-frequency band obtained using the following Equation (14).

Equation 14

In Equation (14), the index is sb+1 represents the index podology the low-frequency band in the signal podology the high-frequency band. In addition, the index eb represents the index podology the high-frequency band that is encoded in the signal podology the high-frequency band.

As described above, the difference between pseudomodest podology the high-frequency band calculated by the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band, serves � scheme 37 encoding the high-frequency band.

In step S117, the circuit 37 encoding the high-frequency band encodes the difference between pseudomodest podology the high-frequency band from the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band and delivers the encoded data of high-frequency bands obtained by the result in scheme 38 multiplexing.

In particular, the circuit 37 encoding the high-frequency band determines the vector obtained by forming the difference between pseudomodest podology the high-frequency band from the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band (below called a vector difference of pseudomodest podology the high-frequency band), which owns the cluster among the set of clusters in the space of characteristics given difference podology pseudomodest the high-frequency band. Here, the vector difference of pseudomodest podology the high-frequency band in the time frame J has, as an element of a vector, the value of the difference between power_{diff}(ib,J) of Pseudomonas podology the high-frequency band for each index ib, and represents a vector of dimension (eb-sb). In addition, the space features minus pseudomodest podology the high-frequency band is set as a space of dimension (eb-sb), in the same way.

Therefore, the circuit 37 of the high-frequency encoding �tvoc measures the distance between the set of each representative vector from the set of defined clusters and the vector difference of pseudomodest podology the high-frequency band in the space of characteristics of the difference pseudomodest podology the high-frequency band, gets the index of the cluster having the shortest distance (below called id minus pseudomodest podology the high-frequency band), and delivers the received index, as encoded data of the high-frequency band, in scheme 38 multiplexing.

In step S118, the circuit 38 multiplexing multiplies the encoded data of the low-frequency band output from the circuit 32 encoding the low-frequency band, and the coded high-frequency band data that is output from the circuit 37 encoding the high-frequency band, and outputs the output code string.

Therefore, as the encoder in the encoding characteristics of the high-frequency band, in lined with the application for Japanese patent No. 2007-17908 discloses the technology of forming pseudosinella podology the high-frequency band from the signal podology the low-frequency band by comparing the pseudo signal podology the high-frequency band and signal strength podology the high-frequency band with each other for each podology, calculate the power gain for each podology so that it corresponds to pseudomodest signal podology the high-frequency band to the signal strength podology the high-frequency band, and ensure that the calculated gain in the line of code as the information about the product�the way of the high-frequency band.

In accordance with the processing described above, only Id the difference pseudomodest podology the high-frequency band can be switched to the output line of code, as information for estimating power podology the high-frequency band decoding. Thus, for example, if, given the number of clusters is equal to 64, as a recovery of the high-frequency band signal in the decoder, 6-bit information can be added to the code string for each time frame, and amount of information included in the code string can be reduced to improve the decoding efficiency compared to the method disclosed in application laid on the Japan patent No. 2007-17908, and it is possible to reproduce a music signal having a better quality sound.

In addition, in the processing described above, the circuit 39 decodes the low-frequency band may introduce a low-frequency band signal obtained by decoding encoded data of the low-frequency band from the circuit 32 encoding the low-frequency band in scheme 33 division popolos, and circuit 34 for calculating the magnitude characteristics, if there is a stock of largest characteristics. In the processing of the decoding performed by the decoder, the value is calculated characteristics of the low-frequency band signal by decoding �kodirovanie the low-frequency band data, and power podology appreciate the high-frequency band based on the magnitude characteristics. Therefore, the processing of coding, even if the difference between the id of pseudomodest podology the high-frequency band, calculated on the basis of the characteristics of the decoded signal of the low-frequency band, will be included in the string encoding, the processing of decoding by the decoder, can be obtained power evaluation podology the high-frequency band having the best accuracy. Therefore, it is possible to reproduce a music signal having a better quality sound.

An example of a functional configuration of the decoder

Next, referring to Fig.13, will be described an example of a functional configuration of the decoder corresponding to the encoder 30 of Fig.11.

The decoder 40 includes circuitry 41 demultiplexing, the circuit 42 decodes the low-frequency band, the circuit 43 division popolos, the circuit 44 for calculating the magnitude characteristics and circuit 45 decodes the high-frequency band, the circuit 46 computing the decoded power podology the high-frequency band, the circuit 47 of the formation of the decoded high-frequency band signal and scheme 48 synthesis.

Circuit 41 performs demultiplexing demultiplexing of the input code string, obtains encoded data of the high-frequency band and low frequency encoded data �tvoc, and delivers the encoded data of the low-frequency band in scheme 42 decoding the low-frequency band, and supplies the encoded data of the high-frequency band in scheme 45 decoding the high-frequency band.

Circuit 42 decodes the low-frequency band performs decoding of encoded data of the low-frequency band from the circuit 41 demultiplexing. Circuit 42 decodes the low-frequency band signals of low frequency bands obtained from the result of decoding (below called decoded low frequency band signal) in the circuit 43 division popolos, scheme 44 computing characteristics and scheme 48 synthesis.

Scheme 43 division popolos evenly divides the decoded low frequency band signal from the circuit 42 decodes the low-frequency band into many signals podology having a predetermined bandwidth, and sends a signal podology (decoded signal podology the low-frequency band) in the circuit 44 for calculating the magnitude and characteristics in scheme 47 the formation of the decoded high-frequency band signal.

Circuit 44 for calculating the magnitude characteristics calculates one or more values of characteristics, using any of a variety of signals podology from the decoded signals podology the low-frequency band from the circuit 43 division popolos, and the decoded signal�La the low-frequency band from the circuit 42 decodes the low-frequency band, and delivers the calculated values of the characteristics in scheme 46 computing the decoded power podology the high-frequency band.

Scheme 45 decoded high frequency band decodes encoded data of the high-frequency band from the circuit 41 demultiplexing and takes the ratio (lower is called the coefficient estimates of the decoded power podology the high-frequency band) to assess the capacity podology the high-frequency band, using the id of a difference of pseudomodest podology the high-frequency band received from the result, which is prepared for each set id (index) in scheme 46 computing the decoded power podology the high-frequency band.

Scheme 46 computing the decoded power podology calculates the high-frequency band decoded power podology the high-frequency band based on one or more values of characteristics from the circuit 44 for calculating the magnitude characteristics and rating of the decoded power podology the high-frequency band from the circuit 45 decodes the high-frequency band and delivers the calculated decoded power podology the high-frequency band in scheme 47 signal decoded high-frequency band signal.

Scheme 47 decoded signal generates the high-frequency band decoded�this high-frequency band signal based on the decoded signal podology the low-frequency band from the circuit 43 division popolos and the decoded power podology the high-frequency band from the circuit 46 computing decoded power podology the high-frequency band and delivers the generated signal and power in the scheme 48 synthesis.

Scheme 48 synthesis synthesizes the decoded low frequency band signal from the circuit 42 decodes the low-frequency band and the decoded high-frequency band signal from the circuit 47 of the formation of the decoded high-frequency band signal, and outputs the synthesized signal as output signal.

The decoding process of the decoder

Next, with reference to the block diagram of the sequence of operations shown in Fig.14, will be described processing of the decoding, using the decoder of Fig.13

In step S131, the circuit 41 demultiplexing demultiplexes the input code string encoded in the high-frequency band data and the encoded data of the low-frequency band, takes these encoded data of the low-frequency band in scheme 42 decoding the low-frequency band and delivers the encoded data of the high-frequency band in scheme 45 decoding the high-frequency band.

In step S132, the circuit 42 decodes the low-frequency band decodes encoded data of the low-frequency band from the circuit 41 demultiplexing and delivers the decoded signal of the low-frequency band, resulting in scheme 43 division popolos, the circuit 44 for calculating the magnitude brushless�and Ki in scheme 48 synthesis.

In step S133, the circuit 43 division popolos still divides the decoded low frequency band signal from the circuit 42 decodes the low-frequency band into many signals podology having a predetermined bandwidth, and delivers the received decoded signal podology the low-frequency band in circuit 44 for calculating the magnitude characteristics and circuit 47 of the formation of the decoded high-frequency band signal.

In step S134, the circuit 44 for calculating the magnitude characteristics calculates one or more values of characteristics from any one of the plurality of signals podology decoded signals podology the low-frequency band from the circuit 43 division popolos and the decoded low frequency band signal of circuit 42 decodes the low-frequency band, and supplies these signals to the circuit 46 computing the decoded power podology the high-frequency band. In addition, the circuit 44 for calculating the magnitude characteristics of Fig.13 basically has the same configuration and function as the circuit 14 for calculating the magnitude characteristics of Fig.3, and the processing in step S134 represents the same processing as in step S4 of the flowchart of the sequence of operations of Fig.4. Therefore, its description is eliminated.

In step s135 grade, the circuit 45 of the high-frequency band decoding decodes the encoded data of high-frequency �tvoc from the circuit 41 demultiplexing and feeds the coefficient estimates of the decoded power podology the high-frequency band, prepared for each set id (index), using the id of a difference of pseudomodest podology the high-frequency band, obtained from the result in scheme 46 computing the decoded power podology the high-frequency band.

In step S136, the circuit 46 computing the decoded power podology calculates the decoded power podology the high-frequency band based on one or more values of characteristics from the circuit 44 computing characteristics and rating of the decoded power podology the high-frequency band from the circuit 45 decodes the high-frequency band and delivers the power to the circuit 47 of the formation of the decoded signal of the high-frequency band. In addition, after decoding the high-frequency band, the circuit 46 computing the decoding podology the high-frequency band of Fig.13 has the same configuration and function as that of the circuit 15 of the power podology the high-frequency band of Fig.3, and the processing in step S136 is the same as the processing in step S5 in the flowchart of sequences of operations on Fig.4, a detailed description is eliminated.

In step S137, the circuit 47 of the formation of the decoded signal of the high-frequency band outputs the decoded high-frequency band signal based on the decoded signal podology the low-frequency band from the circuit 43 razdelnopoloe and the decoded power podology the high-frequency band from the circuit 46 computing the decoded power podology the high-frequency band. In addition, since the circuit 47 of the formation of the decoded high-frequency band signal of Fig.13 basically has the same configuration and function as in scheme 16 for the formation of the high-frequency band signal of Fig.3, and the processing in step S137 is identical with the processing in step S6 of the flowchart of the sequence of operations in Fig.4, detailed description thereof is excluded.

In step S138, the scheme 48 synthesis synthesizes the decoded low frequency band signal from the circuit 42 decodes the low-frequency band and the decoded high-frequency band signal from the circuit 47 of the formation of the decoded high-frequency band signal, and outputs the synthesized signal as output signal.

In accordance with the above-described processing, it becomes possible to improve the accuracy of cardinality estimation podology the high-frequency band, and thus, it becomes possible to reproduce music signals that have a good quality when decoding, using the coefficient estimates power podology the high-frequency band by decoding, in response to a characteristic difference between pseudomodest podology the high-frequency band, the pre-calculated during encoding, and the actual power podology the high-frequency band.

In addition, in accordance with the processing, since the information for the formation of signal�and the high-frequency band, included in the code string that has only id the difference pseudomodest podology the high-frequency band, it is possible to effectively perform decoding processing.

As described above, although the encoding processing and decoding processing in accordance with the present invention, described below, hereinafter will be described a method of calculation of each representative vector of the plurality of clusters in a particular space given difference of pseudomodest podology the high-frequency band in the encoding scheme 37 the high-frequency band, the encoder 30 in Fig.11 and the coefficient estimates decoded pseudomodest podology the high-frequency band output from the circuit 45 decodes the high-frequency band decoder 40 in Fig.13.

The calculating method for calculating a representative vector of a plurality of clusters in a particular space minus pseudomodest podology the high-frequency band and decoding the coefficient estimates power podology the high-frequency band corresponding to each cluster

As a method of obtaining a representative vector of a plurality of clusters and the coefficient estimates of the decoded power podology the high-frequency band of each cluster, you must prepare the factor for estimating power podology the high-frequency band with a high accuracy at decoder�provided in response to the calculated vector difference of pseudomodest podology the high-frequency band encoding. Therefore, examine in advance with the help of a broadband signal, the instructions and method of determining the study applied on the basis of the result of the study.

An example of a functional configuration of the device for the study of the coefficient

Fig.15 illustrates an example of a functional configuration of the device for the study of the coefficient of performing a study of representative vectors of the plurality of clusters, and the coefficient estimates of the decoded power podology the high-frequency band of each cluster.

Preferably, the component signal to a wideband signal instructions entered into the device 50 study of the coefficient of Fig.15 and at the cutoff frequency or less set by the filter 31 low frequency encoder 30, consisted of the decoded low frequency band signal, which is input to the encoder 30 passes through the filter 31 low frequency, which is coded by a coding scheme 32 the low-frequency band and which is decoded by the circuit 42 decodes the low-frequency band decoder 40.

The device 50 study of the coefficient includes a filter 51 low frequency, the circuit 52 of the separation popolos, the circuit 53 for calculating the magnitude characteristics, the circuit 54 for calculating pseudomodest podology the high-frequency band, the circuit 55 for calculating the difference between pseudomodest podology high frequency half�Sy scheme 56 separation of clusters minus pseudomodest podology the high-frequency band and scheme 57 the coefficient estimates.

In addition, since each of the filters 51 low-frequency circuits 52 division popolos, circuit 53 for calculating the magnitude characteristics and circuit 54 for calculating pseudomodest podology the high-frequency band in the device 50 of studying the coefficient of Fig.15 basically has the same configuration and function as each of the filters 31 low frequency, scheme 33 division popolos, circuit 34 for calculating the magnitude of the number and characteristics of the circuit 35 for calculating pseudomodest podology the high-frequency band in the encoder 30 of Fig.11, description thereof is appropriately excluded.

In other words, although the scheme 55 for calculating the difference between pseudomodest podology the high-frequency band provides the same configuration and function as the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band in Fig.11, the calculated difference of pseudomodest podology the high-frequency band is supplied to the circuit 56 of the separation of clusters minus pseudomodest podology the high-frequency band, and power podology the high-frequency band, calculated in the calculation of the difference of pseudomodest podology the high-frequency band, is fed into a circuit 57 of the coefficient estimates.

Scheme 56 separation of the clusters�s the difference pseudomodest podology the high-frequency band is divided into clusters, the vector difference of pseudomodest podology the high-frequency band, obtained from the difference of pseudomodest podology the high-frequency band from the circuit 55 for calculating the difference between pseudomodest podology the high-frequency band, and calculates a representative vector in each cluster.

Circuit 57 calculates the coefficient estimates coefficient estimates power podology the high-frequency band for each cluster divided by the cluster circuit 56 of the separation of clusters minus pseudomodest podology the high-frequency band based on the power podology the high-frequency band from the circuit 55 for calculating the difference between pseudomodest podology the high-frequency band and one or more values of characteristics from the circuit 53 for calculating the magnitude characteristics.

The process of learning coefficient of the device for the study of the coefficient

Below, with reference to the block diagram of the sequence of operations shown in Fig.16, will be described processing of the learning coefficient device 50 study of the coefficient of Fig.15

In addition, the process in step S151-S155 of the flowchart of the sequence of operations in Fig.16 are identical are presented on the steps S111, S113-S116 of the flowchart of the sequence of operations in Fig.12, except that the signal entering the device 50 learning coefficient, is a broadband signal to the instruction and, thus, their description is excluded.

In this way�m, in step S156, the circuit 56 of the separation of clusters minus pseudomodest podology the high-frequency band is divided into many clusters of vectors difference of pseudomodest podology the high-frequency band (a large number of time frames) obtained from the difference of pseudomodest podology the high-frequency band from the circuit 55 for calculating the difference between pseudomodest podology the high-frequency band 64 cluster and calculates a representative vector for each cluster. As an example of the method of separation of clusters, for example, apply the method of separation of clusters by k States. Scheme 56 separation of clusters minus pseudomodest podology the high-frequency band set in the center vector of each cluster obtained from the execution result of the division into clusters using the k state representative vector of each cluster. Furthermore, the method of separation of clusters or the number of clusters is not limited to this but can be applied other ways.

In addition, the circuit 56 of the separation of clusters minus pseudomodest podology the high-frequency band, measures a distance between a 64 representative vectors and the vector difference of pseudomodest podology the high-frequency band obtained from the difference of pseudomodest podology highly�attotney strip from the circuit 55 for calculating the difference between pseudomodest podology the high-frequency band during the time frame J, and determines the index CID (J) of the cluster included in the representative vector that has the shortest distance. In addition, the index CID (J) takes the integer value 1 for the number of clusters (e.g., 64). Therefore, the circuit 56 of the separation of clusters minus pseudomodest podology outputs the high-frequency band representative vector and delivers the index CID (J) in the scheme of assessment 57 coefficient.

In step S157, the circuit 57 displays coefficient estimates coefficient estimates of the decoded power podology the high-frequency band in each cluster for each set that have the same index CID (J) (included in the same cluster) among the many combinations of quantities (eb-sb) power values podology the high-frequency band and amount of characteristics supplied in the same time frames from the circuit 55 for calculating the difference between pseudomodest podology the high-frequency band and circuits 53 evaluation value characteristics. Method for coefficient calculation circuit 57, the coefficient estimates are identical to the method performed by the circuit 24 estimates of the coefficient of the device 20 the study of the coefficient of Fig.9. However, it can be used another way.

In accordance with the above-described processing by using a predetermined wideband signal instructions because the study of each representative vector of a VA�TWA clusters in a particular space minus pseudomodest podology the high-frequency band, pre-defined in the schema 37 encoding the high-frequency band, the encoder 30 in Fig.11, and is studying for the decoded coefficient estimating power podology output circuit 45 decodes the high-frequency band decoder 40 in Fig.13, it becomes possible to obtain a desired output result for different input signal applied to the encoder 30, and various input code strings supplied to the decoder 40, and it is possible to reproduce a music signal with high quality.

In addition, with regard to encoding and decoding of the signal, coefficient data for calculating power podology the high-frequency band in the circuit 35 for calculating pseudomodest podology the high-frequency band, the encoder 30 in the circuit 46 computing the decoded power podology the high-frequency band of the decoder 40 can be handled as follows. Thus, it is possible to write the ratio in the forward position of a code string using a different data of the coefficient, depending on the type of the input signal.

For example, it is possible to improve coding efficiency by changing the coefficient data on a signal, such as speech and jazz.

Fig.17 illustrates a code string, obtained from the above method.

In code line A in Fig.17 coded speech, and optimal�data and coefficient when speech is recorded in the header.

In contrast, as in the code string In Fig.17 encoded jazz, optimal data β coefficient for jazz recorded in the header.

The coefficient data set described above, can be easily studied in advance by the music signal of the same type, and encoder 30 may select the data of the coefficient from information of the genre written in the header of the input signal. In addition, the genre is defined by performing the analysis of waveforms, and can be selected coefficient data. Thus, the method of analysis of the genre of the signal is not restricted in particular.

When time allows calculation, the encoder 30 is equipped with a device study, described above, and thus, the processing performed using the ratio that is specialized for signal and, as provided in the code string in Fig.17, in the end, it is also possible to write a ratio in the header.

The advantage of using the method will be described below.

Form power podology the high-frequency band includes many similar provisions in a single input signal. When using characteristics of multiple input signals, and by performing a study of the coefficient for estimating power podology the high-frequency band of each input separately, reduced redundancy, similar�of agenia power podology the high-frequency band, resulting in improved coding efficiency. In addition, it becomes possible to perform the estimation of power podology the high-frequency band with a higher accuracy than in the study of the coefficient for estimating power podology the high-frequency band, statistically using a variety of signals.

In addition, as described above, coefficient data, study from the input signal by decoding, can take the form of a single insertion in each of the several frames.

3. The third variant of implementation

An example of a functional configuration of the encoder

In addition, although it has been described that the id minus pseudomodest podology the high-frequency band output from the encoder 30 and the decoder 40, as coded data of the high-frequency band, the index of the coefficient to obtain the decoded coefficient estimating power podology the high-frequency band can be set to a coded high-frequency band data.

In this case, the encoder 30, for example, as shown in Fig.18. In addition, in Fig.18 parts corresponding to parts in Fig.11 have the same numbers of reference positions, and their description is accordingly excluded.

The encoder 30 in Fig.18 is the same, except that the encoder 30 in Fig.11 and the circuit 39 decodes the low-frequency band is not provided, and in the East�flax are the same.

In the encoder 30 of Fig.18, the circuit 34 for calculating the magnitude characteristics counting capacity podology the low-frequency band as the amount of characteristic using the signal podology the low-frequency band supplied from the circuit 33 of the separation popolos, and feeds the circuit 35 for calculating pseudomodest podology the high-frequency band.

In addition, in the circuit 35 for calculating pseudomodest podology the high-frequency band multiple valuation ratios decoded power podology the high-frequency band resulting from a given regression analysis, compare with the index of the coefficient that sets the coefficient estimates of the decoded power podology the high-frequency band intended to record.

In particular, the sets of coefficients A_{ib}(kb) and coefficient (B_{ib}for each podology used in the performance of Equation (2) described above, are prepared beforehand, as the coefficient estimates of the decoded power podology the high-frequency band. For example, the coefficient A_{ib}(kb) and coefficient (B_{ib}estimated using the regression analysis using method of least squares, by installing power podology the low-frequency band for poznavshaya variable and power podology the high-frequency band for the variable, poznavshaya dev�.
In regression analysis, the input signal comprising a signal podology the low-frequency band and the signal podology the high-frequency band is used as a broadband signal instructions.

Scheme 35 calculate pseudomodest podology the high-frequency band computes pseudomodest podology the high-frequency band for each podology on the side of the high-frequency band, using the coefficient estimates of the decoded power podology the high-frequency band and the magnitude characteristics of the circuit 34 for calculating the magnitude characteristics recorded for each of the coefficient estimates of the decoded power podology the high-frequency band and delivers the power podology in circuit 36 for calculating the difference between pseudomodest podology the high-frequency band.

Circuit 36 for calculating the difference between pseudomodest podology compares the high-frequency band power podology the high-frequency band received from the signal podology the high-frequency band supplied from the circuit 33 of the separation popolos, pseudomodest podology the high-frequency band from the circuit 35 for calculating pseudomodest podology the high-frequency band.

In addition, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band takes the index of the coefficient for the coefficient estimates of the decoded power CCTP�band high-frequency band, where pseudomodest podology the high-frequency band close to most pseudomodest podology the high-frequency band, among the results of the comparison and the many valuation ratios decoded power podology the high-frequency band, in scheme 37 encoding the high-frequency band. Thus, get the index of the coefficients for the coefficient estimates of the decoded power podology the high-frequency band from which the high-frequency band signal input intended for playback by decoding, which is a decoded high-frequency band signal that is closest to the true value.

The encoding process of the encoder

Next, with reference to the block diagram of the sequence of operations in Fig.19, will be described processing of the encoding performed by the encoder 30 in Fig.18. In addition, the processing in step S181 - step S183 shown in the step S111 to the step S113 in Fig.12. Therefore, the description here is excluded.

In step S184, the circuit 34 for calculating the magnitude features calculates the value characteristics using signal podology the low-frequency band from the circuit 33 of the separation popolos and takes the value characteristics in the circuit 35 for calculating pseudomodest podology the high-frequency band.

In particular, the circuit 34 calculations led�ranks features calculates as the magnitude characteristics, power power (ib,J) podology the low-frequency band for frames J (where, 0≤J) in respect of each podology ib (where sb-3≤ib≤sb) on the side of the low-frequency band by performing an operation in accordance with Equation (1) described above. Thus, power power (ib,J) podology the low-frequency band is calculated, by digitizing RMS value for the sample value of each sample of the signal podology the low-frequency band gap of frames J.

In step S185, the circuit 35 for calculating pseudomodest podology the high-frequency band computes pseudomodest podology the high-frequency band based on the magnitude characteristics supplied from the circuit 34 for calculating the magnitude characteristics, and delivers pseudomodest podology the high-frequency band in circuit 36 for calculating the difference between pseudomodest podology the high-frequency band.

For example, a circuit 35 for calculating pseudomodest podology the high-frequency band in advance, calculates pseudomodest power_{est}(ib,J), podology the high-frequency band, in accordance with the above-mentioned Equation (2) using the coefficient A_{ib}(kb) and coefficient (B_{ib}recorded as the decoded power coefficient podology the high-frequency band, and evaluation of Pseudomonas power_{est}(ib,J) podology the high-frequency band, which�testvol operation of the above-mentioned Equation (2),
using power power (kb,J) podology the low-frequency band (where, sb-s≤kb≤sb).

Thus, the coefficient, A_{ib}(kb) for each podology multiplies power power (kb,J) podology the low-frequency band (kb,J) of each podology on the side of the low-frequency band, supplied as the value of the characteristics, and the coefficient B_{ib}add to the amount of power podology the low-frequency band, resulting in the ratio and then multiply it becomes pseudomodest power_{est}(ib,J) podology the high-frequency band. This pseudomodest podology the high-frequency band is calculated for each podology on the side of the high-frequency band in which the index is sb+1 to eb.

In addition, the circuit 35 for calculating pseudomodest podology the high-frequency band calculating pseudomodest podology the high-frequency band for each coefficient estimates the decoded power podology the high-frequency band, prerecorded. For example, it is assumed that the index of the coefficient allows to prepare in advance from 1 to K (where, 2≤K) decodings of the coefficient estimates podology the high-frequency band. In this case, pseudomodest podology the high-frequency band of each podology calculated for each of the coefficients To assess the decoded power podology the high-frequency band.

NATPE S186, circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the difference between pseudomodest podology the high-frequency band based on the signal podology the high-frequency band from the circuit 33 of the separation popolos, and pseudomodest podology the high-frequency band from the circuit 35 for calculating pseudomodest podology the high-frequency band.

In particular, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band does not perform the same operation that the Equation (1) described above, and calculates the power power (ib,J) podology the high-frequency band in frames J in the signal podology the high-frequency band from the circuit 33 of the separation popolos. In addition, in a variant implementation, all podporou signal podology the low-frequency band and signal podology distinguish the high-frequency band, using the index ib.

Further, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band performs the same operation as in Equation (14) described above and calculates the difference between the power power (ib,J) podology the high-frequency band in frames J and pseudomodest power_{est}(ib,J) podology the high-frequency band. In this case, the power difference_{diff}(ib,J) of Pseudomonas podology the high-frequency band gain for each of the coefficient estimates of dekodieren�th power podology the high-frequency band in respect of each podology on the side of the high-frequency band,
the index is sb+1 to eb.

In step S187, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the following Equation (15) for each coefficient estimates the decoded power podology the high-frequency band and calculates the sum of squares of the difference of pseudomodest podology the high-frequency band.

Equation 15

In addition, in Equation (15) the sum of squares for the difference E (J,id) is prepared in respect of the coefficient estimates of the decoded power podology the high-frequency band in which the index coefficient is id and frames J. in addition, in Equation (15), power_{diff}(ib,J,id) is prepared in respect of the coefficient estimates of the decoded power podology the high-frequency band in which the index coef�of icient represents the id of the decoded power podology the high-frequency band and represents the difference between pseudomodest podology the high-frequency band (power_{
diff}(ib,J)) for the difference of power_{diff}(ib,J) of Pseudomonas podology the high-frequency band frame J podology, the index is ib. The sum of squares of the difference E (J,id) is calculated against the amount For each coefficient estimates the decoded power podology the high-frequency band.

The sum of squares for the difference E (J,id), obtained above, is a similar degree of power podology the high-frequency band, calculated from the actual high-frequency band signal and pseudomodest podology the high-frequency band, calculated using the coefficient estimates of the decoded power podology the high-frequency band, and the index of the coefficient represents the id.

Thus, the error values of the estimates are shown in relation to the true value of power podology the high-frequency band. Therefore, the smaller the sum of squares for the difference E (J,id), the more the decoded signal of the high-frequency band, closed the actual high-frequency band signal obtained as a result of the operation, using the coefficient estimates of the decoded power podology the high-frequency band. Thus, the coefficient estimates of the decoded power podology the high-frequency band in which the sum of squares for the difference E (J,id) is minimal, is a co�efficient evaluation most suitable for processing the extension of the frequency band that is performed by decoding the output code string.

Circuit 36 for calculating the difference between pseudomodest podology the high-frequency band selects the sum of squares for the difference with the minimum value among the sums of squares for the difference E (J,id), and sends the index of the coefficient representing the coefficient estimates of the decoded power podology the high-frequency band corresponding to the sum of squares for the difference in scheme 37 encoding the high-frequency band.

In step S188, the circuit 37 encoding the high-frequency band encodes the index of the coefficient supplied from the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band, and supplies the obtained encoded data of the high-frequency band in scheme 38 multiplexing.

For example, in step S188, perform entropy encoding, etc. in relation to the index of the coefficient. Therefore, the amount of coded information for the high-frequency band data that is output to the decoder 40 may be compressed. In addition, if the encoded data of the high-frequency band consists of information about what you get optimal coefficient estimates of the decoded power podology the high-frequency band, any information is preferable; for example, the index can before�add up to a high frequency encoded data strip, in the form as it is.

In step S189, the circuit 38 multiplexing multiplexes the encoded data of the low-frequency band supplied from the circuit 32 encoding the low-frequency band, and the coded data of the high-frequency band supplied from the circuit 37 encoding the high-frequency band, and outputs the output code string, and the encoding process ends.

As described above, the coefficient estimates of the decoded power podology high frequency bands, the most suitable for processing, can be obtained by outputting encoded data of high-frequency bands obtained by encoding the index of the coefficient as the output code string, the decoder 40 that takes input as a string output code, together with coded data of the low frequency. Therefore, it becomes possible to obtain a signal having a higher quality.

An example of a functional configuration of the decoder

In addition, the output code string output from the encoder 30 of Fig.18, is introduced as input a line of code, and, for example, a decoder 40 for decoding has a configuration shown in Fig.20. In addition, in Fig.20, parts corresponding to the case shown in Fig.13, are denoted by the same symbols, and description thereof is excluded.

The decoder 40 in Fig.20 is identical to the decoder 40 in Fig.13 in that the circuit 41 for demultiplexing �Hemi 48 synthesis configured but differs from decoder 40 in Fig.13 the fact that the decoded low frequency band signal from the circuit 42 decodes the low-frequency band is supplied to the circuit 44 for calculating the magnitude characteristics.

In the decoder 40 of Fig.20, the circuit 45 decodes the high-frequency band writes the coefficient estimates of the decoded power podology the high-frequency band, identical to the coefficient estimates of the decoded power podology the high-frequency band in which a prerecorded circuit 35 for calculating pseudomodest podology the high-frequency band of Fig.18. Thus, the set of coefficient A_{ib}(kb) and coefficient (B_{ib}as the coefficient estimates of the decoded power podology the high-frequency band, as a result of the regression analysis, record, in accordance with the index of the coefficient.

Scheme 45 the high-frequency band decoding decodes the encoded data of the high-frequency band supplied from the circuit 41 demultiplexing, and supplies the coefficient estimates of the decoded power podology the high-frequency band indicated by the index of the coefficient obtained from the result in scheme 46 computing the decoded power podology the high-frequency band.

The decoding process of the decoder

Then, the processing of decoding performed by the decoder 40 of Fig.20, will be described� with reference to the block diagram of the sequence of operations in Fig.21.

The decoding processing is started when the output code string output from the encoder 30, which serves as the input code string, the decoder 40. In addition, since the processing in step S211 - S213 identical presents in step S131 - step S133 in Fig.14, a description is excluded.

In step S214, the circuit 44 for calculating the magnitude features calculates the value characteristics using the decoded signal podology the low-frequency band, from the circuit 43 division popolos, and feeds the circuit 46 computing the decoded power podology the high-frequency band. More circuit 44 for calculating the magnitude features calculates the value of the features for power power (ib,J) podology the low-frequency band for frames J (no,0≤J), when performing the operation of Equation (1) described above in respect of each ib podology on the side of the low-frequency band.

In step S215, the circuit 45 of the high-frequency band decoding performs decoding of encoded data of the high-frequency band supplied from the circuit 41 demultiplexing, and supplies the coefficient estimates of the decoded power podology the high-frequency band indicated by the index of the coefficient obtained by the result in scheme 46 computing the decoded power podology the high-frequency band. Thus, deduce the coefficient estimates decoded�Oh power podology the high-frequency band, which is designated with the index of the coefficient obtained by decoding in a variety of valuation ratios decoded power podology the high-frequency band, recorded in advance in the circuit 45 decodes the high-frequency band.

In step S216, the circuit 46 computing the decoded power podology calculates the high-frequency band decoded power podology the high-frequency band based on the magnitude characteristics supplied from the circuit 44 for calculating the magnitude characteristics, and decodes the coefficient estimates of the decoded power podology the high-frequency band supplied from the circuit 45 decodes the high-frequency band, and supplies it to the circuit 47 of the formation of the decoded high-frequency band signal.

Thus, the circuit 46 computing the decoded power podology the high-frequency band performs an operation according to the Equation (2) described above, using the coefficient A_{ib}(kb), as the coefficient estimates of the decoded power podology the high-frequency band and power power (kb,J) podology the low-frequency band, and the coefficient B_{ib}(where, sb-3≤kb≤sb), as the value of the characteristic, and calculates a decoded power podology the high-frequency band. Therefore, the decoded power podology the high-frequency band receive in respect of each�Oh podology on the side of the high-frequency band,
the index is sb+1 to eb.

In step S217, the circuit 47 of the formation of the decoded signal generates the high-frequency band decoded high-frequency band signal based on the decoded signal podology the low-frequency band supplied from the circuit 43 division popolos, and the decoded power podology the high-frequency band supplied from the circuit 46 computing the decoded power podology the high-frequency band.

In more detail, the circuit 47 of the formation of the decoded high-frequency band signal performs an operation in accordance with the above-described Equation (1) using the decoded signal podology the low-frequency band, and calculates the power podology the low-frequency band in respect of each podology on the side of the low-frequency band. In addition, the circuit 47 of the formation of the decoded high-frequency band signal calculates the ratio G (ib,J) of amplification for each podology on the side of the high-frequency band by performing an operation in accordance with Equation (3) described above, using the power podology low frequency band decoded and received power podology the high-frequency band.

In addition, the circuit 47 of the formation of the decoded signal generates the high-frequency band signal X3 (ib,n) podology the high-frequency band, performed by�I operations on Equations (5) and (6), as described above, using the coefficient G (ib,J) of the gain and the decoded signal podology the low-frequency band in respect of each podology side of the high-frequency band.

Thus, the circuit 47 of the formation of the decoded high-frequency band signal performs amplitude modulation of the decoded signal x (ib,n) podology the high-frequency band in response to the power ratio of podology low frequency band decoded for power podology the high-frequency band and, thus, performs frequency modulation of the received decoded signal x2 (ib,n) podology the low-frequency band. Therefore, the signal frequency component of podology on the side of the low-frequency band is converted to a frequency signal component podology on the side of the high-frequency band, and receives a signal X3 (ib,n) podology the high-frequency band.

As described above, the processing for receiving the signal podology the high-frequency band for each podology is a process described in more detail below.

Four podology, forming lines in the frequency field, is called a block of bands, and the frequency band is divided so that one block of the strip (hereafter called block low-frequency band) was composed of four popolos in which the index is present on the low side, is from sb to sb-3. � this case, for example, a streak that includes podporou in which the index side of the high-frequency band includes from sb+1 to sb+4, represents one block of the strip. In addition, the side of the high-frequency band, i.e., the power strip comprising podporou in which the index is sb+1 or more, in particular, is called a block of the high-frequency band.

In addition, attention is drawn to one podporou, a component unit of the high frequency bands and generate a signal podology the high-frequency band for this podology (below called podoloski attention). Initially, the circuit 47 of the formation of the decoding of the high-frequency band signal sets popolos block the low-frequency band, which has the same ratio of provisions to the provisions podology attention to block the high-frequency band.

For example, if the index podology attention is a sb+1, podporou block the low-frequency band, having the same relationship provisions, which has popoloca attention, establish how popolos, the index is sb-3, as popoloca attention is a band of frequency which is the lowest in blocks the high-frequency band.

As described above, popoloca if popoloca for podology block the low-frequency band, having the same relationship provisions, h�about and popoloca attention is a specific use power podology the low-frequency band and the decoded signal podology the low-frequency band and the decoded power podology the high frequency bands and generate a signal podology the high-frequency band for podology attention.

Thus, the decoded power podology the high-frequency band and power podology the low-frequency band substituted into Equation (3), so that the calculated gain factor in accordance with its degree of power. In addition, the calculated gain is multiplied by the decoded signal podology low frequency band decoded signal podology the low-frequency band multiplied by the gain, is set as frequency modulation, through the operation of Equation (6), which is set as a signal podology the high-frequency band for podology attention.

When processing a receive signal podology the high-frequency band of each podology on the side of the high-frequency band. In addition, the circuit 47 of the formation of the decoded high-frequency band signal performs the Equation (7) described above, to obtain the amount of each of the signals podology the high-frequency band and to form a decoded signal of high frequency band. SEMA 47 forming decode�separate the high-frequency band signal delivers the received decoded signal of the high-frequency band in scheme 48 synthesis, and the processing goes from step S217 to step S218, and then the decoding processing ends.

In step S218, the scheme 48 synthesis synthesizes the decoded low frequency band signal from the circuit 42 decodes the low-frequency band and the decoded high-frequency band signal from the circuit 47 of the formation of the decoded high-frequency band signal, and outputs, as output signal.

As described above, since the decoder 40 has an index coefficient encoded data from the high-frequency band obtained by demultiplexing the input lines of code, and calculates the decoded power podology the high-frequency band with the use of the coefficient estimates of the decoded power podology the high-frequency band, by use of the coefficient estimates of the decoded power podology the high-frequency band indicated by the index ratio, it becomes possible to improve the accuracy of cardinality estimation podology the high-frequency band. Therefore, it is possible to generate a music signal, with high quality.

4. The fourth variant of implementation

The encoding process of the encoder

First, as described above, will be described a case in which only the index factor included in encoded data of the high-frequency band. However, it may be included and other�formation.

For example, if the index coefficient is included in encoded data of the high-frequency band, the coefficient estimates of the decoded power podology the high-frequency band, which is decoded power podology the high-frequency band that is closest to the power podology actual high-frequency band signal the high-frequency band, is passed, as a notification on the side of the decoder 40.

Therefore, the actual capacity podology the high-frequency band (true value) and the decoded power podology the high-frequency band (evaluation value) obtained from the decoder 40, form the difference, essentially, is the difference of power_{diff}(ib,J) of Pseudomonas podology the high-frequency band, calculated in the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band.

Here, if the index ratio and the difference of pseudomodest podology the high-frequency band in popoloca will be included in the coded data of the high-frequency band, the decoded error power podology the high-frequency band in relation to the actual power podology the high-frequency band becomes approximately known on the side of the decoder 40. If so, it becomes possible to improve the accuracy of cardinality estimation podology the high-frequency band, using this difference.

Milling�TKA encoding and decoding processing, in the case where the difference between pseudomodest podology the high-frequency band included in the coded data of the high-frequency band will be described with reference to the block diagram of the sequence of operations of Fig.22 and 23.

First, the processing of the encoding performed by the encoder 30 in Fig.18, will be described with reference to the block diagram of the sequence of operations in Fig.22. In addition, the processing in step S241 - S246 identical to the processing of step S181 - step S186 in Fig.19. Therefore, its description is eliminated.

In step S247, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band performs the operation of Equation (15) described above to calculate the sum E (J,id) of the squares of the difference for each coefficient estimates the decoded power podology the high-frequency band.

In addition, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band selects the sum of squares for the difference, where the sum of squares for the difference is set as a minimum in the sum of squares for the difference among the sum E (J,id) squares for the difference and delivers the index of the coefficient denoting the coefficient estimates of the decoded power podology the high-frequency band corresponding to the sum of the square of the difference in scheme 37 encoding the high-frequency band.

In addition, the circuit 36 for calculating the difference between pseudomodest Podo�wasps delivers the high-frequency band power difference_{
diff}(ib,J) of Pseudomonas podology the high-frequency band in each podporou received in respect of the coefficient estimates decoded power podology the high-frequency band corresponding to the selected sum of the squares of the residual errors in scheme 37 encoding the high-frequency band.

In step S248, the circuit 37 encoding the high-frequency band encodes the index of the coefficient and the difference between pseudomodest podology the high-frequency band supplied from the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band, and supplies the encoded data of high-frequency bands obtained from the result in scheme 38 multiplexing.

Therefore, the difference between pseudomodest podology the high-frequency band for each power podology on the side of the high-frequency band where the index is sb+1 to eb, that is, the difference in cardinality estimation podology the high-frequency band serves as encoded data of the high-frequency band in the decoder 40.

If the coded data of the high-frequency band is received, then perform encoding processing in step S249, to stop processing of the encoding. However, the processing in step S249 is identical to the processing in step S189 of Fig.19. Therefore, its description is eliminated.

As described above, when the difference of pseudomodest podology by high frequency�wasps will be included in the coded data of the high-frequency band, it becomes possible to improve the accuracy of cardinality estimation podology the high-frequency band, and to obtain a signal with good quality in the decoder 40.

Processing decoding decoder

Then, the processing of decoding performed by the decoder 40 of Fig.20, will be described with reference to the block diagram of the sequence of operations of Fig.23. In addition, the processing in step S271 - S274 identical to the processing performed in step S211 to the step S214 in Fig.21. Therefore, its description is eliminated.

In step S275, the circuit 45 of the high-frequency band decoding performs decoding of encoded data of the high-frequency band supplied from the circuit 41 demultiplexing. In addition, the circuit 45 decodes the high-frequency band coefficient takes the assessment decoded power podology the high-frequency band indicated by the index of the coefficient obtained by decoding, and the difference between pseudomodest podology the high-frequency band for each podology obtained by decoding, in scheme 46 computing the decoded power podology the high-frequency band.

In step S276 scheme 46 computing the decoded power podology calculates the high-frequency band decoded power podology the high-frequency band based on the magnitude characteristics supplied from the circuit 44 you�of Alenia size characteristics, and the coefficient 216 evaluation decoded power podology the high-frequency band supplied from the circuit 45 decodes the high-frequency band. In addition, in step S276 performs the same processing as that in step S216 in Fig.21.

In step S277, the circuit 46 computing the decoded power podology the high-frequency band adds the difference between pseudomodest podology high frequency bands, referred to the scheme 45 the high-frequency band decoding to the decoded power podology the high-frequency band, and supplies the result of the summation as the final decoded power podology the high-frequency band, in scheme 47 the formation of the decoded high-frequency band signal.

Thus, the difference between pseudomodest podology the high-frequency band in the same popoloca summarize the decoded power podology the high-frequency band of each settlement podology.

In addition, after that carry out the processing in step S278 and step S279, the decoding processing ends. However, this treatment is identical to the step S217 and step S218 of Fig.21. Therefore, its description is eliminated.

Performing described above, the decoder 40 receives the index of the coefficient and pseudomodest podology the high-frequency band encoded data from the high-frequency band, the resulting demultiple�funding line of the input code. In addition, the decoder 40 calculates the decoded power podology the high-frequency band, using the coefficient estimates of the decoded power podology the high-frequency band indicated by the index ratio, and the difference between pseudomodest podology the high-frequency band. Therefore, it becomes possible to improve the precision power podology the high-frequency band, and to reproduce the audio signal with high sound quality.

In addition, the difference value evaluation capacity podology the high-frequency band is formed between the encoder 30 and the decoder 40, that is, the difference (below called the estimation of the difference between the device) between pseudomodest podology the high-frequency band and the decoded power podology the high-frequency band can be considered.

In this case, for example, the difference between pseudomodest podology the high-frequency band used as the encoded data of high-frequency bands are adjusted by the valuation difference between the devices, and the evaluation of the difference between the devices is included in encoded data of the high-frequency band, while the difference between pseudomodest podology the high-frequency band correcting the difference of the evaluation between the devices on the side of the decoder 40. In addition, the difference in estimates between the device can be pre-recorded n� the side of the decoder 40, and decoder 40 may perform the correction by summing the difference of the evaluation between devices with the difference between pseudomodest podology the high-frequency band. Therefore, it becomes possible to obtain the decoded high-frequency band signal, closed in the actual high-frequency band signal.

5. The fifth variant of the implementation

In addition, in the encoder 30 in Fig.18 described that the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band selects the optimal index from the set of the indices of the coefficients, using the square of the sum E (J,id) for the difference. However, the scheme may choose the index of the index, using the index that differs from the square of the sum for the difference.

For example, as an index, when the index is selected coefficient, you can use the RMS value, maximum value and average value of the residual error power podology the high-frequency band and pseudomodest podology the high-frequency band. In this case, the encoder 30 in Fig.18 performs encoding processing illustrated in the flowchart of sequences of operations on Fig.24.

The process of encoding, using the encoder 30 will be described with reference to the block diagram of the sequence of operations in Fig.24. In addition, the processes in step S301 - step S305 are identical are presented in step S181 - S185 the f�G. 19. Therefore, their description will be omitted. If you do the process in step S301 - step S305, pseudomodest podology the high-frequency band of each podology calculated for each number of coefficients To assess the decoded power podology the high-frequency band.

In step S306, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the evaluation value Res (id,J) using the current frame J to be processed for each quantity To the coefficient estimates of the decoded power podology the high-frequency band.

In more detail, the circuit 36 for calculating the difference between pseudomodest podology calculates the high-frequency band power power (ib,J) podology the high-frequency band in frames J, by performing the same operation as that in Equation (1) described above, using the signal podology the high-frequency band of each podology supplied from the circuit 33 of the separation popolos. In addition, in the embodiment of the present invention, it is possible to distinguish all podology signal podology the low-frequency band and podology the high-frequency band, using the index ib.

If the power power (ib,J) podology the high-frequency band is received, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the following Equation (16) and calculates engine residual�nd the root-mean-square values Res_{
std}(id,J).

Equation 16

Thus, the difference between power power (ib,J) podology the high-frequency band (ib,J) and pseudomodest power_{est}(ib,id,J) podology the high-frequency band gain for each podology on the side of the high-frequency band where the index is sb+1 through eb, and the square of the sum for this dierence becomes residual RMS Res_{std}(id,J). In addition, pseudomodest power_{rest}(ibh,id,J) podology the high-frequency band indicates pseudomodest podology vyskocil�a combined strip for frames J podology,
where the index is ib, which is obtained in respect of the coefficient estimates of the decoded power podology the high-frequency band where the index is equal to ib.

And again, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the following Equation (17) and calculates the residual maximum value Res_{max}(id,J).

Equation 17

In addition, in Equation (17),denotes the maximum value among the absolute values of the difference between power power (ib,J) podology the high-frequency band of each podology, where index is in the range from sb+1 to eb and the pseudo�ewnost power_{
est}(ib,id,J) podology the high-frequency band. Therefore, the maximum value of the absolute value of the difference between power power (ib,J) podology the high-frequency band in frames J and pseudomodest power_{est}(ib,id,J) podology the high-frequency band is set as the residual maximum value Res_{max}(id,J) of the difference.

In addition, the circuit 36 minus pseudomodest podology the high-frequency band calculates the following Equation (18) and calculates the average residual value Res_{ave}(id,J).

Equation 18

Thus, for each podology on the side of the high-frequency band in which the index is sb+1 to eb, get the difference between power power (ib,J) podology the high-frequency band of the frames J and pseudomodest power_{est}(ib,id,J) podology the high-frequency band, and receive the amount of this difference. In addition, the absolute value of the value obtained by dividing the amount of the difference in the number popolos (eb-sb) on the side of the high-frequency band is set as the average residual value Res_{ave}(id,J). The residual mean value Res_{ave}(id,J) denote the average value of the estimation error for each podology symbol, which is taken into account.

In addition, if the residual RMS Res_{std}(id,J), residual maximum value Res_{max}(id,J) of the difference and the average residual value Res_{ave}(id,J) are received, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the following Equation (19) and calculates the final evaluation value Res (id,J).

Equation 19

Thus, the residual RMS Res_{std}(id,J), residual maximum value Res_{max}(id,J) and residual mean value Res_{ave}(id,J) is summed with the weight and set the final value Res (id,J) evaluations. In addition, in Equation (19), W_{max}and W_{ave}are required weight and, for example, W_{max}=0.5, W_{ave}=0.5.

Circuit 36 for calculating the difference between pseudomodest podology the high-frequency band performs the processing described above and calculates the value Res (id,J) of the assessment for each of the number of coefficients To assess the decoded power podology the high-frequency band, i.e., the number To the index id of the index.

In step S307 arc, the circuit 36 vychisleny� difference of pseudomodest podology the high-frequency band selects the index id of the coefficient based on the value of Res assessment for each received index id of the index (id,J).

The value Res (id,J) estimates obtained by the processing described above, indicates the degree of similarity between the power podology the high-frequency band, calculated from the actual high-frequency band signal, and pseudomodest podology the high-frequency band, calculated using the coefficient estimates of the decoded power podology the high-frequency band, which is the index id of the index. Thus is indicated the size of the estimation error for the high-frequency band component.

Accordingly, as the assessment Res (id,J) becomes low, the decoded high frequency signal strip located closer to the actual high-frequency band signal obtained by the operation using the coefficient estimates of the decoded power podology the high-frequency band. Therefore, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band selects the data rate that is set as the minimum value among the number of values To evaluate Res (id,J) and takes the index of the coefficient denoting the coefficient estimates of the decoded power podology the high-frequency band corresponding to the value of evaluation in scheme 37 encoding the high-frequency band.

If the index coefficient is output to the encoding scheme 37 high-frequency� stripes after that, the processing in step S308 and step S309, the encoding processing is then terminated. However, since identical processing in step S188 of Fig.19 and in step S189, a description will be excluded.

As described above, in the encoder 30 uses the value Res (id,J) estimates calculated by using the residual mean square values Res_{std}(id,J), residual maximum value Res_{max}(id,J) and residual mean value Res_{ave}(id,J), and choose the optimal index of the coefficient estimates of the decoded power podology the high-frequency band.

If you use the value Res (id,J), since the estimation accuracy of the power podology the high-frequency band can be estimated using a larger number of standards assessments in comparison with the case of using the square of the sums of the difference, it becomes possible to select a more appropriate factor evaluation decoded power podology the high-frequency band. Hence, the decoder 40, the receiving input of the output lines of code, can obtain the coefficient estimates of the decoded power podology the high-frequency band, which is most suitable for the process of expanding the frequency band and the signal with higher sound quality.

Example 1 modification

In addition, if the process �of tiravanija, as described above, is performed for each frame of the input signal, there is a case in which the index of the coefficient, wherein in each subsequent frame in the stationary region, where small variation in time of power podology the high-frequency band of each podology side of the high-frequency band of the input signal.

Thus, since the power podology the high-frequency band of each frame is practically identical values in consecutive frames that comprise a standard area of the input signal, the same index coefficient is a need to continuously choose their frame. However, the index coefficient is selected for each frame on a plot of successive frames is changed and, thus, the high-frequency band component of the voice played on the side of the decoder 40 can no longer be stationary. If so, there is a mismatch in audio playing.

Accordingly, if the index coefficient is chosen in the encoder 30, the result of the evaluation component of the high-frequency band in the previous frame time can be taken into account. In this case, the encoder 30 in Fig.18 performs the encoding process shown in the flowchart of sequences of operations on Fig.25.

As described below, the encoding process performed by encoder 30 will be described with reference to b�OK-the flow chart in Fig.25. In addition, the processing in step S331 - S336 identical presents in step S301 - step S306 of Fig.24. Therefore, its description will be eliminated.

Circuit 36 for calculating the difference between pseudomodest podology the high-frequency band computes the value ResP (id,J) estimates, using the last frame and the current frame in step S337.

In particular, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band writes pseudomodest podology the high-frequency band for each podology obtained by the coefficient estimates of the decoded power podology the high-frequency band index of the coefficient selected in the end, in relation to the frames J-1 earlier than the J frame, which is processed one at a time. Here is chosen, ultimately, the index ratio is called the index of the coefficient that is output to the decoder 40 as a result of encoding using the encoding scheme 37 the high-frequency band.

As described below, in particular, the id index of the coefficient selected in a frame (J-1), set as id_{selected}(J-1). In addition, pseudomodest podology the high-frequency band for podology, the index which is obtained using the coefficient estimates of the decoded power podology the high-frequency band of index id_{selected}(J-1) coefficient equal to ib (where sb+1≤ib≤eb), constantly poyasnee�Xia,
as power_{est}(ib,id_{selected}(J-1),J-1).

Circuit 36 for calculating the difference between pseudomodest podology the high-frequency band first calculates the following Equation (20), and then receives a score residual RMS ResP_{std}(id,J).

Equation 20

Thus, the difference between pseudomodest power_{est}(ib,id_{selected}(J-1),J-1) podology the high-frequency band frame J-1 and the dog�domonatia -
power_{est}(ib,id,J) podology the high-frequency band frame J receive in respect of each podology side of the high-frequency band where the index is sb+1 to eb. In addition, the sum of squares for the difference between sets, as the difference of the estimation error for RMS ResP_{std}(id,J). In addition, pseudomodest podology the high-frequency band - power_{est}(ib,id,J) indicates pseudomodest podology the high-frequency band frames (J) podology, the index is ib, which is obtained in respect of the decoded coefficient estimating power podology the high-frequency band, where the index of the coefficient is equal to id.

Since this estimate of the residual value of the square ResP_{std}(id,J) is the sum of the squares of the difference of pseudomodest podology the high-frequency band between frames that are continuous in time, the smaller the estimation of residual RMS ResP_{std}(id,J), the less variation in the time value component of the high-frequency band.

Further, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the following Equation (21) and calculates the estimation residual maximum value ResP_{max}(id,J).

Equation 21

In addition, in Equation (21),denotes the maximum absolute value of the difference between pseudomodest power_{est}(ib,id_{selected}(J-l),J-1) podology the high-frequency band of each podology in which the index is sb+1 to eb and pseudomodest power_{est}(ib,id,J) podology the high-frequency band. Therefore, the maximum value of absolute values of differences between frames that are continuous in time, is set as the estimation residual maximum value ResP_{max}(idp,J) RA�of the error.

The smaller the maximum value of the ResP_{max}(id,J) for estimating the residual error, the closer the result of the evaluation of the high-frequency component between consecutive frames.

If you receive an assessment of residual maximum value ResP_{max}(id,J), then, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the following Equation (22), and calculates the estimation residual mean value ResP_{ave}(id,J).

Equation 22

Thus, the difference between pseudomodest power_{est}(ib,id_{selected}(J-1),J-1) podology the high-frequency band frame (J-1) and pseudomodest power_{est}(ib,id,J) podology the high-frequency band frame J receive in respect of each podology on the side of the high-frequency band, when the index is in the range from sb+1 to eb. In addition, the absolute value of the value obtained by dividing the sum of the difference of each podology on the number popolos (eb-sb) on the side of the high-frequency band, is set as the estimation residual mean value ResP_{ave}(id,J). The average value of the ResP_{ave}(id,J) is the residual error estimate is the size of the average value of the difference value evaluation podology between frames, consider where the symbol.

In addition, if RMS ResP_{std}(id,J) residual evaluation, the maximum value of the ResP_{max}(id,J) is the residual error estimates and the average value of the ResP_{ave}(id,J) residual of the evaluation will be obtained, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the following Equation (23) and calculates the average value ResP (id,J).

Equation 23

<>Thus, the average value of the ResP_{std}(id,J) residual evaluation, the maximum value of the ResP_{max}(id,J) is the residual error estimates and the average value of the ResP_{ave}(id,J) is the average residual evaluation summarize with the weight and set that as the value ResP (id,J) evaluations. In addition, in Equation (23), W_{max}and W_{ave}represent a given weight, for example, W_{max}=0,5, W_{ave}=0,5.

Therefore, if the count value ResP (id,J) estimates using the previous frame and the current value, the processing goes to step S337 at step S338.

In step S338, the circuit 3 for calculating the difference between pseudomodest podology the high-frequency band calculates the Equation (24) and calculates the final value of estimation ReS_{
all}(id,J).

Equation 24

Thus, the obtained value Res (id,J) assess and value ResP (id,J) summarize the evaluation with weight. In addition, in Equation (24), W_{p}(J), for example, represents a weight determined by the following Equation (25).

Equation 25

In addition, power_{r}(J) in Equation (25) represents the value defined by the following Equation (26).

Equation 26

This value is power_{r}(J) represents the average of the difference between the power values podology wysokosc�frequency band frames (J-1) and frames J.
In addition, in accordance with Equation (25) when power_{r}(J) represents the value of a predetermined range near 0, the smaller the value of power_{r}(J), the closer W_{p}(J) to 1 and when power_{r}(J) is greater than the value of the specified range, it is set as 0.

Here, when power_{r}(J) represents the value of a predetermined range near 0, the average value of the difference of power podology the high-frequency band between successive frames becomes small to a certain extent. Thus, the variation in time of the high-frequency band component of the input signal becomes small, and the current frame of the input signal become established area.

Since the high-frequency band component of the input signal is stable, the weight W_{p}(J) becomes a value close to 1, whereas, in the case where the high-frequency band component is not stable, the weight (W_{p}(J) becomes a value close to 0. Therefore, the value Res_{all}(id,J) of the evaluation, shown in Equation (24), since the variation in time of the high-frequency band component of the input signal becomes small, the coefficient of determination of the evaluation value ResP (id,J) by considering the result of the comparison and the evaluation result the high-frequency band component, as standards for evaluation in the previous frame�x,
becomes large.

Therefore, in the steady-state region of the input signal, selects the coefficient estimates of the decoded power podology the high-frequency band obtained in close proximity to the result of evaluating the high-frequency band component in the previous frame, and the decoder 40 becomes possible to more naturally reproduce sound with high quality. At that time, instead of the transient region of the input signal, the member value ResP (id,J) of the evaluation value Res_{all}(id,J) evaluation set equal to 0, and get the decoded signal of the high-frequency band close to the actual high-frequency band signal.

Circuit 36 for calculating the difference between pseudomodest podology the high-frequency band computes the value Res_{all}(id,J) of the assessment for each of the number of coefficients To assess the decoded power podology the high-frequency band, by performing the above-mentioned processing.

In step S339, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band selects the index id of the coefficient based on the evaluation value Res_{all}(id,J) obtained for each coefficient estimates the decoded power podology the high-frequency band.

Value Res_{all}(id,J) is the score obtained in the process described above, linearly combines the value of Re (id,J) assess and value ResP (id,J) evaluation
using weight. As described above, the smaller the value Res (id,J) the score is, the closer the decoded high-frequency band signal to the actual signal the high-frequency band can be obtained. In addition, the smaller the value ResP (id,J) evaluation of the decoded high-frequency band signal, the closer the decoded high-frequency band signal of the previous frame can be obtained.

Therefore, the smaller the value Res_{all}(id,J) evaluations, especially corresponding to the decoded high-frequency band signal get. Therefore, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band selects the evaluation value having the minimum value among To estimates Res_{all}(id,J), and transmits the index of the coefficient denoting the coefficient estimates of the decoded power podology the high-frequency band corresponding to this value of evaluation in scheme 37 encoding the high-frequency band.

If the index coefficient is chosen, then perform the processing in step S340 and step S341 to the end of the encoding process. However, since this processing is the same as the processing in step S308 and step S309 of Fig.24, the description here is excluded.

As described above, in the encoder 30 uses the value Res_{all}(id,J) is the score obtained by the linear combination �values Res (id,J) evaluation
and value ResP (id,J) of assessment used in such a way that selects the index of the optimal coefficient of the coefficient estimates of the decoded power podology the high-frequency band.

If the value Res_{all}(id,J) estimation is used, as in the case when the value Res (id,J) of the evaluation, it becomes possible to select a more appropriate factor evaluation decoded power podology the high-frequency band using a much larger number of evaluation standards. However, if the value Res_{all}(id,J) estimation is used, it becomes possible to control the variation in time established in the area the high-frequency band component of the signal intended for playback in the decoder 40, and it is possible to obtain a signal with high quality.

Example 2 modification

Incidentally, in the processing of expanding the bandwidth, if you want to get the sound with high quality, popoloca on the side of the lower band is also important, in the sense of listening. Thus, among popolos on the side of the high-frequency band, as the accuracy of the estimate podology close to the side of the low-frequency band becomes greater, it becomes possible to reproduce sound with high quality.

Here when calculating the evaluation value for each friction coefficient close to�the evaluation decoded power podology the high-frequency band, the weight can be set for podology on the side of the lower band. In this case, the encoder 30 in Fig.18 performs the encoding process shown in the flowchart of sequences of operations on Fig.26.

Below, the processing of the encoding performed by the encoder 30 will be described with reference to the block diagram of the sequence of operations in Fig.26. In addition, the processing in step S371 - step S375 identical presents in step S331 - S335 in Fig.25. Therefore, its description is eliminated.

In step S376, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the estimated value ResW_{band}(id,J) using the current frame J, intended for processing, for each of the number of coefficients To assess the decoded power podology the high-frequency band.

In particular, the circuit 36 for calculating the difference between pseudomodest podology calculates the high-frequency band power power (ib,J) podology the high-frequency band in the J frames, performs the same operation as that described above in Equation (1) using the signal podology the high-frequency band for each podology supplied from the circuit 33 of the separation popolos.

If the power power (ib,J) podology the high-frequency band is received, a circuit for calculating 36 minus pseudomodest podology the high-frequency band calculates the following Equation 27 and rasschityva�t residual RMS Res_{
std}W_{band}(id,J).

Equation 27

Thus, the difference between power power (ib,J) podology the high-frequency band frames (J) and pseudomodest podology the high-frequency band (power_{est}(ib,id,J), and the difference is multiplied by a weight W_{band}(ib) for each podology, for each podology on the side of the high-frequency band, DG� the index is sb+1 to eb.
In addition, the sum of squares for the difference, which multiply the weight W_{band}(ib), is set as the residual mean square error value Res_{std}W_{band}(id,J).

Here the weight W_{band}(ib) (where, sb+1≤ib≤eb is determined by the following Equation 28. For example, the weights W_{band}(ib) becomes so large as popoloca on the side of the lower band.

Equation 28

Further, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the residual maximum value Res_{max}W_{band}(id,J). In particular, the maximum value of the absolute values, multiplying the difference between the power power (ib,J) podology the high-frequency band of each podology, where index is in the range from sb+1 to eb, and pseudomodest power_{est}(ib,id,J) podology the high-frequency band to the weight W_{band}(ib), is set as the maximum value Res_{max}W_{band}(id,J) of the difference residual error�.

In addition, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band computes the average value Res_{ave}W_{band}(id,J) is the residual error.

In particular, every popoloca, where the index is sb+1 to eb, difference between power power (ib,J) podology the high-frequency band and pseudomodest power_{est}(ib,id,J) podology receive the high-frequency band and, thus, the weight W_{band}(ib) is multiplied so that you receive the total amount of the difference, which multiply the weight W_{band}(ib). In addition, the absolute value of the value obtained by dividing the total amount received difference on podporou room (eb-sb) on the side of the high-frequency band, is set as the average value Res_{ave}W_{band}(id,J) is the residual error.

In addition, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band computes the value ResW_{band}(id,J) evaluations. Thus, the amount of the residual RMS values Res_{std}W_{band}(id,J), the maximum value Res_{max}W_{band}(id,J), the residual error, which multiply the weight (W_{max}), and the average value Res_{ave}W_{band}(id,J) is the residual error, which multiply the weight (W_{ave}), set as the average value ResW_{band}(id,J).

In step S377, the circuit 36 for calculating the difference between pseudomodest podology in�pass band calculates the average value ResPW_{
band}(id,J) using the past frame and the current frame.

In particular, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band writes pseudomodest podology the high-frequency band for each podology obtained by using the coefficient estimates of the decoded power podology the high-frequency band for the coefficient selected in the end in relation to the frames J-1 one frame before the frame (J) intended for processing time.

Circuit 36 for calculating the difference between pseudomodest podology the high-frequency band initially calculates the estimate of the average value ResP_{std}W_{band}(id,J) is the residual error. Thus, for each podology on the side of the high-frequency band in which the index is sb+1 to eb, the weight W_{band}(ib) is multiplied, by obtaining the difference between pseudomodest power_{est}(ib, id_{selected}(J-1),J-1) podology the high-frequency band and pseudomodest power_{est}(ib,id,J), podology the high-frequency band. In addition, the sum of squares of the difference from which is calculated the weights W_{band}(ib), is set as the estimate of the average value ResP_{std}W_{band}(id,J) of the difference of the error.

Circuit 36 for calculating the difference between pseudomodest podology the high-frequency band continuously calculates the maximum value ResP_{mx}
W_{band}(id,J) is the residual error. In particular, the maximum value of the absolute values obtained by multiplying the difference between pseudomodest power_{est}(ib,id_{selected}(J-1),J-1) podology the high-frequency band for each podology in which the index is sb+1 to eb and pseudomodest - power_{est}(ib,id,J) podology the high-frequency band to the weight W_{band}(ib), is set as the maximum value ResP_{max}W_{band}error (id,J) is the residual error.

Further, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band computes the average value ResP_{ave}W_{band}(id,J) is the residual error. In particular, multiply the difference between pseudomodest power_{est}(ib,id_{selected}(J-1), J-1) podology the high-frequency band and pseudomodest power_{est}(ib, ID, J) podology the high-frequency band gain for each podology, where the index is sb+1 to eb and the weight W_{band}(ib). In addition, the total amount of the difference, which multiply the weight W_{band}(ib), represents the absolute value of values obtained by dividing the number (eb-sb) popolos on the side of the high-frequency band. However, as an estimate of the average value of the residual error set ResP_{ave}W_{band}(id,J).

In addition, the circuit 36 for calculating the difference between pseudomona�ti podology receives the high-frequency band the amount of RMS R_{
es}P_{std}W_{band}(id,J) is the residual error for the estimation of maximum values ResP_{max}W_{band}(id,J), the residual error, which multiply the weight W_{max}and the estimate of the mean value ResP_{ave}W_{band}(id,J) is the residual error, which multiply W_{ave}and the amount is set as a value of evaluation ResPW_{band}(id,J).

In step S378, the circuit 36 for calculating the difference between pseudomodest podology summarizes the high-frequency band value ResW_{band}(id,J) of the evaluation value ResPW_{band}(id,J) estimates that multiply the weight W_{p}(J) by Equation (25), to calculate the final values Res_{all}W_{band}(id,J) assess the value Res_{all}W_{band}(id,J) estimates calculated for each of the quantities on the coefficient estimates of the decoded power podology the high-frequency band.

In addition, after that, executes the processing in step S379 - step S381, the end of the processing of the encoding. However, since these processes are identical are presented with reference to step S339 - step S341 in Fig.25, a description of them are excluded here. In addition, the value Res_{all}W_{band}(id,J) evaluation choose the minimum among the number of indexes To factor in step S379.

As described above, to place the weight on podporou on the side of the low-frequency band, it is possible to get a sound with more high quality side decode�and 40, providing a weight for each podology.

In addition, as described above, the choice of the number of valuation ratios decoded power podology the high-frequency band has been described as being performed based on the value Res_{all}W_{band}(id,J) evaluations. However, an estimate of the decoded power podology the high-frequency band can be selected based on the value ResW_{band}(id,J) of the evaluation.

Example 3 modification

In addition, since the properties of the human ear is such that people correctly perceive a greater frequency range amplitude (power), the value of assessment in respect of each decoded coefficient estimating power podology the high-frequency band can be calculated so that the weight can be set for podology having more power.

In this case, the encoder 30 in Fig.18 performs the encoding process represented in the flowchart of sequences of operations on Fig.27. Processing of the encoding performed by the encoder 30 will be described below with reference to the flowchart of sequences of operations on Fig.27. In addition, since the processing in step S401 - step S405 identical to the steps S331 to step S335 of Fig.25, the description here will be omitted.

In step S406, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band computes the value ResW_{power}(id,J) of�sister,
using the current frame J, intended for processing, for the number To the coefficient estimates of the decoded power podology the high-frequency band.

In particular, the circuit 36 for calculating the difference between pseudomodest podology calculates the high-frequency band power power (ib,J) podology the high-frequency band in the frame J by performing the same operation as that in Equation (1) described above, using the signal podology the high-frequency band of each podology supplied from the circuit 33 of the separation popolos.

If the power power (ib,J) podology the high-frequency band is received, the circuit 36 for calculating the difference between pseudomodest podology the high-frequency band calculates the following Equation (29) and calculates the RMS value Res_{std}W_{power}(id,J) is the residual error.

Equation 29