# Method for reverse filtration, method for synthesizing filtration, device for reverse filtration, device for synthesizing filtration and tools containing such devices

FIELD: systems/methods for filtering signals.

SUBSTANCE: in accordance to invention, filtration of input signal is performed for generation of first filtered signal; first filtered signal is combined with aforementioned input signal for production of difference signal, while stage of filtering of input signal for producing first filtered signal contains: stage of production of at least one delayed, amplified and filtered signal, and production stage contains: storage of signal, related to aforementioned input signal in a buffer; extraction of delayed signal from buffer, filtration of signal for forming at least one second filtered signal, while filtration is stable and causative; amplification of at least one signal by amplification coefficient, while method also contains production of aforementioned first filtered signal, basing on at least one aforementioned delayed, amplified and filtered signal.

EFFECT: development of method for filtering signal with delay cycle.

10 cl, 10 dwg

The invention relates to a method of inverse filtering. In addition, the invention relates to a method of synthesizing filtering. The invention relates to a device for inverse filtering, synthesizing filter and to devices containing such devices for filtering. The invention also relates to a computer program for implementing the stages of the method in accordance with the invention.

A device for filtering is known from the article Anima "Implementation of frequency-warped recursive filters", Signal Processing 80 (2000) 543-548. This article describes the encoder with linear prediction in the frequency distortion (WLP-encoder and decoder with linear prediction in the frequency distortion (WLP-decoder). WLP-encoder contains the standard filter with finite impulse response (FIR), in which a single delay replaced seastate filters of the first order.

The disadvantage of the encoder is known from the article, is that without any additional devices WLP-decoder would contain cycles without delay. This article describes two solutions to this problem. First, WLP-decoder can be selected in such a way as to avoid cycles without delay. Secondly, the calculation of the output signal of the decoder and correction of internal States of the filter can be separated. In both cases, the solutions WLP-decoder differs from WLP-coder. In addition, because RA is communication between the encoder and decoder parameters WLP-encoder, such as prediction coefficients must be transformed for WLP-decoder, which requires more processing and is associated with numerical tasks.

Thus, the object of the present invention is the creation of the encoder and decoder, which can have a similar design. Therefore, the invention provides a method of inverse filtering in accordance with claim 1 of the claims.

Thus, the synthesizing filter does not contain cycles without delay, because it creates a delay. Therefore, inverse filtering and synthesizing filtering can be essentially the same.

In addition, the invention provides a method of synthesizing filtering in accordance with paragraph 17 of the claims. In addition, the invention discloses a device for inverse filtering in accordance with paragraph 18 of the formulas of the invention, a device for synthesizing filtering in accordance with paragraph 19 of the claims and devices containing such filtering devices. The invention also provides a computer program for implementing the stages of the method in accordance with the invention.

Specific embodiments of the invention are disclosed in dependent claims. Further details, aspects and embodiments of the invention will be described with reference to accompanying with the giving of the drawings.

Figure 1 is a block diagram of the first variant of implementation of the device for inverse filtering in accordance with the invention.

Figure 2 is a block diagram of the first variant of implementation of the device for synthesizing filtering in accordance with the invention.

Figure 3 is a block diagram of the first variant of the method of inverse filtering in accordance with the invention.

Figure 4 is a block diagram of a first variant implementation of the method of synthesizing filtering in accordance with the invention.

Figure 5 is a diagram of a device for transmission of data created on the basis of the encoder with the prediction, in accordance with the invention.

6 is a diagram of a device for storing data created on the basis of the encoder with the prediction, in accordance with the invention.

Fig.7 is a diagram of a device for processing data, based on the decoder prediction, in accordance with the invention.

Fig is a diagram of an audiovisual device, based on the decoder prediction, in accordance with the invention.

Fig.9 is a diagram of an audiovisual recording device, based on the decoder prediction, in accordance with the invention.

Phi is .10 is a diagram of a device for storing data, based encoding method with prediction, in accordance with the invention.

In this description the following terms are used. "Sample x(n)is an example of the signal at some point. The segment represents the number of consecutive samples, for example, x(n)x(n+1),..., x(n+j-1), x(n+j). When using either of the terms "signal", "sample" or "segment", and the other of these options can be considered. "The transfer function H(z)" represents the ratio between the input and the output signal of the filter, taken in the definition area z. (For z=exp^{iθ} _{}i is the square root of -1, H(z) gives characteristics in determining the frequency). "Impulse response filter is a response filter to a pulse signal, i.e. a signal having the value "1" for n=0 and a value of 0 for n not equal to 0, and n marks the point in time. In this description, the term "device for filtering" refers not to a device having only a delay device or multiple devices delay, although in a very limited sense, the delay device is a device for filtering. It is understood that the device includes at least one filter device and one or more delay devices, is a device for filtering. At least the,
assume that the filter is causal if the output signal does not depend on any "future" of the input signals, i.e. the output signal of the filter depends only on the current signal and/or the previous signals. They say that the filter is stable if the filter is limited by the amplitude of the output signal for any limited amplitude input signal presented at the input of the filter.

Figure 1 shows the block diagram of the first variant example of implementation of the device 1 for inverse filtering in accordance with the invention. Shows an example of a device 1 for backward filtering device or encoder includes port 11 input, which can be represented by the input signal x. The input port is connected to the system 13 of the filter, which is able to filter the received input signal x and output the first filtered signalPort 11 input and the system 13 of the filter are connected together to the first combining device 12, which is capable of combining the first filtered signaland the input signal x, resulting in a differential signal r.

System 13 of the filter contains a buffer or device 131 memory connected to port 11 input, and many of the second filtering device 132 connected to in the course of device 131. In the shown example, the second filter device 132 form a filter device 130 "one input - many outputs(SIMO). The second filtering device 132 is also connected to the amplifier 133, which are further connected to the second combining unit 134. A combining device 134 is connected with the output of the first combining device 12.

Buffer or device 131 memory, which in this description referred to as a delay device, preserves the input sample x(n) and outputs the sample u(n). Sample u(n) is the previous sample x(n-j) input signal, where j is the delay device and j is greater than 0. Thus, the sample u(n) is the previous input signal u is equal to the sample x(n-j) input signal x, where j is the delay device 131 delays and j greater than or equal to 0. The second filtering device 132 generates a second filtered signals y_{1}, y_{2,.}.., y_{k}on the basis of the signal u. The second filtration device is stable and causal. Thus, the filter device 130 SIMO also stable and causal. In a variant implementation of the filtering device 130 SIMO contains only the second filtering device 132. However, the device SIMO may also contain one or more devices delay or even direct excitation in parallel with the second filtering device 132.

Wuxi is Italy 133 amplify or multiply each of the second filtered signal y_{
1}, y_{2,.}.., y_{k}with the gain or multiplier α_{1}that α_{2},..., α_{k}. From this point of view, to gain α_{1}that α_{2},..., α_{k}referred to as the prediction coefficients α_{1}that α_{2},..., α_{k}where prediction coefficients vary with time or depend on the signal. Thus, the second filtered signals are combined as a weighted sum using a second combining device 134.

The output signal of the second combining device 134 is first filtered signalwhere every sample(n)thus, based on the previous sample x(n-j) input signal x, where j is greater than 0. The second combining unit 134 generates a first filtered signaland transmits the first filtered signalto the first combining device 12. The first combining unit 12 combines the input signal x with the first filtered signaland generates a differential signal r.

Because device 131 delays in the system 13 filter no cycles without delay. Thus, as the inverse filter and the synthesis filter can be one and the e design,
i.e., the filters can be made complementary. For example, an example of a reverse filter in accordance with figure 1 and an example of the synthesizing filter in accordance with figure 2 are complementary. Also time-frequency resolution of the filtration system can be configured in advance through the appropriate choice of transfer function H_{k}the second filter as second filters can be stable and causal filters of any suitable type, for example, by choosing the parameters (such as increment, poles and zeros) of the transfer function H_{k}thus, the filter is configured on an area with a specific frequency.

Device delays and filters and/or amplifiers can be swapped, i.e. the filter and/or amplifiers can be placed before the delay device. In this case, the delay device will save the first filtered signaland output a first filtered signal with the prediction, which is then combined with the input signal x to obtain the difference signal r. Mathematically speaking, the device 131 delay and filter and/or amplifier commutative. However, regardless of the relative positioning of the device delay, filter and/or amplifier, filter communicative connected to the delay device and the first comb is neraudia device.

In addition, the parameters used in the inverse filter can be used in the corresponding synthesis filter, for example, in the example shown in figure 2. Thus, the synthesizing filter can be implemented without devices for the conversion coefficient prediction, and therefore, the synthesizing filter can be cheaper. Installation reverse filter can then be transferred to the synthesizing filter, for example, through the dedicated data channel, or combined with the signal r.

Figure 2 shows a device 2 for synthesizing filter or decoder, which is essentially opposite to the device for inverse filtering figure 1. The device 2 for synthesizing filter has an input port 21 connected to the first combining device 22. A combining device 22, also connected to the system 23 of the filter and the output 24 of the device 2 for synthesizing filtering. At the inlet 21 may be represented by the input signal r. Then the input signal r is received first combining device 22 and is combined with the first filtered signal from the system 23 of the filter, from which is obtained an output signal x. If the input signal r is a differential signal from the device 1 for the inverse filter 1, the output signal x is essentially the same as the input signal is x device for inverse filtering.

The system 23 of the filter device 231 delay (which is also referred to as a buffer device or a memory device)connected to the output 24, and many of the second filtering device 232. The second filtration device 232 are connected to the amplifiers 233 which are connected to the second combining device 234. Second combining device 234 is connected with the output of the first combining device 12.

Device 231 delay saves the output sample x(n) and outputs the original output sample x(n-j), where j is greater than 0. The second filtering unit 232 generates a second filtered signal based on the original output signal. Amplifiers 233 multiply each of the second filtered signal by a factor predictions α_{1}that α_{2},..., α_{k}. Thus, the second filtered signals are combined as a weighted sum using a second combining device 234. The output signal of the second combining device 234 is first filtered signalwhere every sample(n)thus, based on the previous sample x(n-j) output signal x when j is greater than 0. The second combining unit 234 outputs a first filtered signal
and transmits the first filtered signalthe first combining device 1. The first combining unit 22 combines the input signal r with the first filtered signaland gives the output signal x.

Because of the delay device in the system 23 of the filter no cycles without delay. Thus, the synthesizing filter can be a simple way is made so that it is complementary to the inverse filter. Delay and filter and/or amplifiers may be interchanged, i.e. the filter and/or amplifiers can be located in front of the delay device. Mathematically speaking, the delay device and the filter and/or amplifiers are commutative.

In examples 1 and 2, the second filtering devices are connected in parallel to the delay device or buffer device. Thus, each sample of each of the second filtered signal based on the previous samples of the input signal relative to the delay device or buffer device. The second filtering devices can also be connected in a cascading manner. In this case, the k-th second filtered signal y_{k}based on (k-1)-th second filtered signal y_{k-1.}

In the device corresponding to the invention, the device delay and can have any desired delay. Preferably, the delay is such that the preceding signal immediately precedes the signal received at the buffer, i.e. the delay is a unit delay.

Figure 3 is a block diagram of the method of inverse filtering in accordance with the invention. At stages I-VI accepted the input sample, and generates a first filtered sample(n). After stage VI the first filtered sample(n) and the input sample x(n) are combined, resulting in a differential sample r(n) on the stage of the first combining VII. In the shown example, the combining stage VII is a subtraction method, but it is also possible to carry out a different operation until the desired differential signal which is a measure of similarity between the input signal and the filtered signal. It is therefore assumed in the following sample, and stage I-VII are carried out again.

Generating a first filtered signal(n) at stages I-VI begins with the stage of conservation I. under conservation I accepted the input sample x(n) and stored in the buffer. At the second stage, the previous input sample u(n) is extracted from the buffer. In the example preceding input sample u(n) is the immediately preceding input samples is Scam.
It is also possible to use one or more prior samples. Use only the immediately preceding sample allows the buffer to be as small as possible. In phase III, the counter value k is chosen so that the following value was k+1. After stage III is the stage of the second filter IV. At the stage of the second filter filtering method is implemented on the previous input sample u(n), the result is the second filtered sample y_{k}(n). At stage V, the value of the counter k is compared with some pre-specified value of K, and K indicates the total number of stages of the second filter, which are carried out. If the value of k is not similar to the predetermined value K, stage II-V are carried out again. If the value of k is similar to the predetermined value K, the second filtered signals y_{1}(n), y_{2}(n),..., y_{k}(n) combined with a coefficient α_{k}at the second stage combining VI, resulting in the first filtered sample(n).

Figure 4 shows a block diagram of an example of a method of synthesizing filtering in accordance with the invention. The method of synthesizing filtering represented by the block diagram of figure 4, may, for example, the implementation is taken using the device for synthesizing filtering figure 2.

At the stage II sample u(n) is extracted from the buffer. Sample u(n) is the previous output sample x(n-1). In phase III, the counter value k is chosen so that the following would be the value k+1. After stage III is the stage of the second filter IV. At the second stage filter method filter with transfer function H_{k}(z) is implemented in the sample u(n), the result is the second filtered sample y_{k}(n). At stage V, the value of the counter k is compared with some pre-specified value K indicating the total number of stages of the second filter, which are carried out. If the value of k is not similar to the predetermined value K, stage II-V are carried out again. If the value of k is similar to the predetermined value K, the second filtered signals y_{1}(n), y_{2}(n),..., y_{k}(n) combined with a coefficient α_{k}at the second stage combining VI, resulting in the first filtered sample(n). On the stage of the first combining VIII input sample r(n) is combined with the first filtered sample(n), resulting in an output sample x(n). After that, the output sample x(n) is stored in the buffer, and the procedure is repeated.

In the method or device in the accordance with the invention stage of the second filter or the second filter device can be of any type, suitable for a particular implementation, because they are stable and causal. Furthermore, the method or the device according to the invention can, in addition, at least one filter, to include one or more delay devices or direct excitation.

The second stage filter or filtration device can, for example, be stages recursive filtering or filtering devices or stages of filtration and filtration devices with unlimited impulse response (IIR). In the IIR method is also used weighted samples of the output signal and/or samples with a delay to obtain the output signal. In addition, at least one of the second filtering devices may be non-linear filtering device.

On the device for the second filter or the second filter device can be affected psychoacoustic; i.e. it may have a frequency-time resolution comparable to the system of human hearing. For example, the second filter or generate the at least one second filtered signals can be seastate filter with transfer function:

in this equation (1) z^{-1}represents the delay device, k is the number of stages of the secondary filter, it is positively the m integer between 1 and K,
K represents the total number of secondary filters or filtration stages and λ is a constant having an absolute value between zero and one. Parameter λ may, for example, be selected so that the filter has a frequency-time resolution comparable to the system of human hearing.

Also the filter when the psychoacoustic effects may be filtering Legarra with transfer function H_{k}(z), as described using a mathematical algorithm:

In this equation (2) k represents the number of recursive filtration stages, z^{-1}is the delay and λ is a parameter that has an absolute value between zero and one.

It is also possible to perform the second filtering as filtering Cauca with transfer function H_{k}(z), as described using a mathematical algorithm:

In equation (3) k is the number of recursive filtration stages, z^{-1}represents the operation of the delay and λ_{m}is a parameter that has an absolute value between zero and one and λ_{m} ^{*}- complex conjugated value.

The second filter may also be gamma-tone filter or the gamma-tone filter is in,
as, for example, it is known from the article by T. Irino et. al., "A time domain, level-dependent auditory filter", J. Acoust. Soc. Am., 101: 412 - 419, 1997. In General, gamma-tone filters are stationary filters with impulse response h_{k}defined by the formula:

where the parameters are chosen in accordance with the relevant psychoacoustic data. In this equation, a member of the t^{γ-1}e^{σt} _{}is a statistical gamma distribution, ω_{k}represents the frequency or tone of a member of the cos function, t is time and f_{k}phase.

After the second filtration may be some additional processing, such as a matrix operation. The combined transfer function of the filter and matrix operations can then be represented using a mathematical algorithm:

where the algorithm H_{k}(z) represents the combined transfer function of the second filter and the matrix k represents the number of filtration stages, with_{kn}represents the value of the matrix element in position k,n in the matrix, G_{n}(z) represents the transfer function of the second n filter. In equation (5) filters G_{n}(z) can, for example, be the filters Legarra, as determined using equation (2), or filters Cauca as opredelaetsa the equation (3).

For example, the second filtered signals y_{1}, y_{2,.}.., y_{k}can be multiplied with the Fourier matrix. In this case, the matrix values of c_{kn}equation (5) can be chosen equal to:

In this equation (6) w is a weight function, i represents the square root of -1, K represents the number of parts of the second filter.

Filter device and method of filtering in accordance with the invention can be used for data compression, such as coding with linear prediction. For example, in the coding system that contains the encoder and decoder, communicative attached to the encoder, the encoder may include a device for inverse filtering in accordance with the invention, the decoder may include a device for synthesizing filtering in accordance with the invention.

In the filter with prediction or the encoder or the decoder prediction prediction coefficients α_{1}that α_{2},..., α_{k}can be obtained using the following procedure. In the shown example, the prediction coefficients depend on the signals present in the filter. For example, the coefficient of the prediction can be based on some optimization procedure of the samples or signals, such as minimizing significant is Bagrationi errors.

To determine α_{k}at time n is selected portion of the input signal x around n, for example, the segment x(t), where t = (n-M_{1}n-M_{1}+1,..., n+M_{2}), where M_{1}, M_{2}>K. Next segment x(t) is processed by the window (for example, using Windows Janiga) to get processed by the method of open segment s.

Processed by the method of open segment s can then be adapted for the new segment s. For example, the signal may be padded with zeros, the signal may be added a small amount of noise in order to prevent numerical problems when the inverse of the matrix (which is done at a later stage), or the segment of the signal s can be transformed into another segment. This can be done, for example, to obtain a suitable psychoacoustic signal. In this case, the hidden threshold can be calculated using the segment s and the inverse Fourier transform can be used on a hidden threshold to obtain the corresponding time signal.

The signal s'is not necessarily adapted or modified, then processed using the method of filtration or filtration device in accordance with the invention, and obtained the second filtered signals y_{k}. Then you define the prediction coefficients α_{1}that α_{2},..., α_{k}by solving the equation:

In this equation (7) α is a vector containing the coefficients of the prediction: α=[α_{1}that α_{2},..., α_{k}]^{t}and Q is a matrix and P - vector whose components are defined as follows:

In this equation (8) k and l equal to or greater than one, but less than or equal to K, * identifies complex conjugation. To prevent numerical problems associated with the conversion matrix, which is required to determine αcan be used a known method of regularization, such as the addition of a small matrix offset εI of the matrix Q before treatment, and ε is a small number, I is the unit matrix. Determination of the coefficients of the prediction can be performed at any point in time n. However, in practice, the coefficients can be determined at regular time intervals. Using procedures interpolation prediction coefficients can then be determined for other points in time.

In addition, the filtering method in accordance with the invention can be used in the method adaptive differential pulse code modulation (ADPCM). Also a filter device in accordance with the invention can be used in the device is daptive differential pulse code modulation,
as is known in the art, for example from K.Sayood "Introduction to Data compression", 2^{nd}ed. Morgan Kaufmann 2000, chapter 10.5.

Also a filter device or a filtering method in accordance with the invention can be used in coding a speech or audio or filtering.

The filtering device in accordance with the invention can be used in various devices, such as device 20 to transmit data, such as a radio transmitter or router computer network that includes a device 21 for receiving the input signal and the transmitting device 22, such as an antenna, for transmitting the encoded signal, they can be provided by the encoding device 1 with the prediction in accordance with the invention, which is connected to the device 21 for receiving the input signal and the transmitting device 22, as shown in figure 5. Such a device can transmit large amounts of data using a small bandwidth, since the encoding process compresses the data.

Equally it is possible to use the encoder 1 with the prediction, in accordance with the invention, the device 30 for storing data, such as a programmer EPROM super audio CD (SACD), the programmer EPROM DVD, or recordable mini discs for data storage in the device 31 for storing d is the R, such as SACD, DVD, CD or computer hard drive. Such a device 30 contains a holding device 32 for device 31 for storing data, the recording device 33 for recording data in the device 31 for storing data, the device 34 for receiving the input signal, such as a microphone, and the encoder 1 with the prediction, in accordance with the invention, which is connected to the device 34 for receiving the input signal and the recording device 33, as shown in figure 6. This device 30 for data storage capable of storing more data in the device 31 for storing data, at the same time avoids the disadvantages of the known devices for data storage.

It is also possible to create a device 40 for processing data containing the device 41 for receiving the input signal, such as a DVD-rom player, and a device 42 for processing data using a decoder 11 for coded signals with the prediction in accordance with the invention, as shown in Fig.7. Such a device 40 for processing data may be a computer or a computer console to the TV.

It is also possible to create the audio device 50, such as a home stereo or multi-player containing the input device 51 for data, such as audio CD player, and audio output device 52, such as a loudspeaker, with p the power of the decoder 11 for coded signals prediction in accordance with the invention, as shown in Fig. Similarly, recording the audio device 60, such as shown in Fig.9, containing an audio input device 61, such as a microphone, and the output device 62 to the data can be created using the encoder 11 predicting that allows you to record more data when using the same amount of space to store data.

In addition, the invention can be applied to data that is stored in the data storage device like a floppy disk 70, shown in figure 10, the data storage device may, for example, also be a digital versatile disk, or by super-audio CD (SACD), or the first original, or a matrix for producing such a DVD or SACD.

The invention is not limited to the implementation described examples of devices, but can also be used in other devices. In particular, the invention is not limited to physical devices, but can also be used in the logical devices of a more abstract type or software performing the functions of the device. In addition, the device can be physically distributed across a number of devices, while logically it is considered as a separate device. Also devices that are logically considered as a separate device, can the be combined into a single physical device. For example, a buffer or delay device can be physically combined in the second filtration device, although logically can be considered as a separate device, for example, by implementing each of the second filtering device 132 in the delay device 1. Also, a device to reverse or synthesizing filter itself can be implemented as a separate integrated circuit.

The invention may also be implemented in a computer program for calculations in a computer system, including at least the code part for the implementation stages of the method in accordance with the invention, or in order to give an opportunity to widely used computer systems to perform the functions of the filtering device in accordance with the invention. Such a computer program may be created on a data carrier such as CD-rom or diskette, can be stored with data loadable in a memory of the computer system, and data represent a computer program. The data carrier can also be a connecting device for data, such as telephone cable, or wireless connection device, transmitting signals representing the computer program, in accordance with the invention.

In the above description the invention has been described with reference is as specific embodiments of the invention. However, it should be obvious that can be done various modifications and changes without departure from the broader essence and scope of the invention defined by the attached claims. Descriptions and drawings should be regarded, respectively, as illustrative and not as limiting the scope of the invention.

1. The method of inverse filtering of the signal containing at least filtering (I-VI) of the input signal x for the formation of the first filtered signalcombining (VII) of the said first filtered signalwith the above-mentioned input signal x to obtain a difference signal, and the filter phase (I-VI) of the input signal x to produce (I-VI) of the first filtered signalincludes a step of obtaining at least one detainee, amplified and filtered signal, and the step of obtaining includes the preservation of (I) the signal x(n), related to the mentioned input signal x in the buffer (131); extract (II) from the buffer (131) of the delayed signal x(n-j)referred to the delayed signal x(n-j) was stored in the buffer (131) before mentioned signal x(n) associated with said input signal x; filtering the signal u(n) for the formation of (III-V), at least one second filtered the wow signal y_{
k}(n), which filter is stable, the second filtered signal_{k}(n) is limited in amplitude for any arbitrary bounded by the amplitude of the signal u(n), and cause - and-second filtered signal_{k}(n) is dependent only on the signal u(n) and/or previous signals u(n-j); the strengthening of the at least one signal gain α_{k}where the gain α_{k}is at least time-varying or dependent on the input signal; the method also includes receiving (VI) of the said first filtered signal^{}based at least on said at least one delayed, enhanced and filtered signal α_{k}the_{k}.

2. The method of inverse filtering of the signal according to claim 1, in which the preservation of the first signal x(n)related to the input signal x in the buffer (131), and removing from the buffer of the delayed signal x(n-j) is carried out before the said filtered signal u(n) to form at least one second filtered signal y_{k}(n), and the above-mentioned first signal x(n) is the input signal x referred to the delayed signal x(n-j) filtered for forming the aforementioned at least one second filtered the wow signal y_{
k}(n).

3. The method of inverse filtering of the signal according to claim 1, in which the preservation of the first signal x(n)related to the input signal x in the buffer (131), and removing from the buffer of the delayed signal x(n-j) is performed after the filtering of the signal u(n) to form at least one second filtered signal y_{k}(n), and the above-mentioned first signal is referred to the first filtered signalreferred to the input signal x filtered for forming the aforementioned at least one second filtered signal y_{k}(n).

4. The method of inverse filtering of the signal according to claim 1, which referred to the delayed signal x(n-j) immediately precedes said signal x(n), related to the mentioned input signal X.

5. The method of inverse filtering of the signal of claim 1, wherein filtering the input signal x for the formation of the first filtered signal^{}contains at least one stage of nonlinear filtering.

6. The method of inverse filtering of the signal according to claim 1, in which the filtering of the signal u(n) to form at least one second filtered signal y_{k}(n) contains at least one stage of the recursive filter.

7. The method of inverse filtering of the signal according to claim 1, which has h the con-temporary permit, comparable to the system of human hearing.

8. The method of inverse filtering of the signal according to claim 6, in which the filtering of the signal u(n) to form at least one second filtered signal y_{k}(n) includes at least one stage seastate filtering.

9. The method of inverse filtering of the signal according to claim 1, in which the filtering of the signal u(n) to form at least one second filtered signal y_{k}(n) includes at least one phase of the Laguerre filter.

10. The method of inverse filtering of the signal according to claim 1, in which the filtering of the signal u(n) to form at least one second filtered signal y_{k}(n) includes at least one stage filtering Cauca.

11. The method of inverse filtering of the signal according to claim 1, in which the filtering of the signal u(n) to form at least one second filtered signal y_{k}(n) includes a stage gamma-tone filter.

12. The method of inverse filtering of the signal according to claim 1, which also perform a matrix operation with at least one of the second filtered signal y_{k}(n), with the combined transfer function mentioned filtering of the signal u(n) to form a second filtered signal y_{k}(n) and matrix operation is carried out is about the mathematical algorithm

where N_{to}(z) is the combined transfer function of the second filter and the matrix; k is the number of filtration stages;

c_{kn}represents the value of the matrix element in position k, n in the matrix;

G_{n}(z) represents the transfer function of the second filter n for the formation of the second filtered signal y_{k}(n).

13. The method of inverse filtering of the signal according to claim 1, in which the amplification of at least one signal includes multiplying at least one of the second filtered signal y_{k}(n) using the coefficient prediction α_{k}where the coefficient prediction α_{k}obtained in accordance with the filtering method with prediction.

14. The method of inverse filtering of the signal according to claim 1, in which the amplification of at least one signal includes multiplying at least one of the second filtered signal y_{k}(n) using the coefficient prediction α_{k}where the coefficient prediction α_{k}obtained in accordance with the method of adaptive differential pulse code modulation.

15. The method of synthesizing filtering of the signal containing at least the combination of (VIII) of the first filtered signal
the input signal r to determine the output signal x; filter mentioned output signal for forming the (I-VI) of the first filtered signalat this stage of filtering includes a step of obtaining at least one detainee, amplified and filtered signal, and the step of obtaining includes the preservation of (I) the first signal x(n), related to the mentioned input signal r in the buffer (131); extract(II) from the buffer (131) of the delayed signal x(n-j)referred to the delayed signal x(n-j) was stored in the buffer (131) before mentioned signal x(n) associated with said input signal r; filtering the signal u(n) to form (III-V), at least one second filtered signal y_{k}(n), and the generation is stable, the second filtered signal y_{k}(n) is limited in amplitude for any arbitrary bounded by the amplitude of the signal u(n), and cause - and-second filtered signal y_{k}(n) is dependent only on the signal u(n) and/or previous signals u(n-j); the strengthening of the at least one signal gain α_{k}where the gain is at least time-varying or dependent on the input signal; the method also includes receiving (VI) mentioned Otti trojanova signal x,
based at least on said at least one delayed, enhanced and filtered signal α_{k}y_{k}.

16. A device for inverse filtering of the signal containing at least the port (11) input for receiving the input signal x, the first combining device (12)connected to the input port for calculating a differential signal r by combining the first filtered signalwith input signal x; system (13) filter connected to the input port and the first combining device (12) for filtering mentioned input signal x for the formation of the first filtered signaland the transfer of the aforementioned first filtered signalthe first combining device (12); receiving at least one detainee, amplified and filtered signal device for filtering of the signal also contains a port (14) output connected to the first combining device (12) to derive a differential signal r, and the filtration system (13) contains a buffer unit (131) for storing the first signal x(n), related to the mentioned input signal x, and removing the delayed signal x(n-j)referred to the delayed signal x(n-j) has been maintained in is the buffer (131)
as mentioned signal x(n) associated with said input signal x; at least one second filter device (130, 132), communicative connected to the buffer device (131) and the first combining device (12)for filtering of the signal u(n) to form at least one second filtered signal y_{k}(n); and a filtering device (130, 132) is stable, the second filtered signal y_{k}(n) is limited in amplitude for any arbitrary bounded by the amplitude of the signal u(n), and cause - and-second filtered signal y_{k}(n) is dependent only on the signal u(n) and/or previous signals u(n-j); at least one amplifier (133), which has a gain of α_{k}and the gain α_{k}is at least time-varying or dependent on the input signal; however, the filtration system also includes a second combining device (134)connected at least to one of the amplifiers (133)mentioned buffer device and said second filtration device, for receiving the first filtered signalof these, at least one detainee, amplified and filtered signal α_{k}y_{k}.

17. Eliminate the STV for synthesizing filtering of the signal,
containing at least port (21) input for receiving the input signal r; a first combining device (22) for combining the mentioned input signal r with the first filtered signalin the result of which receive the output signal x; system (23) filter connected to the port (21) input and the first combining device (22) for filtering the output signal x to the formation of the first filtered signaland transmission of the first filtered signalthe first combining device (21); a device for filtering also contains a port (22) output connected to the first combining device (21) for outputting the output signal x, to obtain at least one detainee, amplified and filtered signal, the filtering system also includes a buffer device (231) for storing the first signal x(n) and extract the delayed signal x(n-j)referred to the delayed signal x(n-j) was stored in the buffer (131) before mentioned signal x(n) associated with said input signal r; at least one second filter device (230, 232), communicative connected to the buffer device (231) and the first combining device (22)for filtering of the signal u(n) to build the project,
at least one second filtered signal y_{k}(n), and the second filter device is stable, the second filtered signal y_{k}(n) is limited in amplitude for any arbitrary bounded by the amplitude of the signal u(n), and cause - and-second filtered signal_{k}(n) is dependent only on the signal u(n) and/or previous signals u(n-j); at least one amplifier (233), which has a gain of α_{k}and the gain α_{k}is at least time-varying or dependent on the input signal; however, the filtration system also includes a second combining device (234)mentioned buffer unit and said second filter device, for receiving the first filtered signalof these, at least one detainee, amplified and filtered signal α_{k}the_{k}.

18. The data device containing a device for receiving an input signal, a transmitter for transmitting the filtered signal and a device for inverse filtering of the signal according to clause 16, is connected to the receiver input and the transmitter.

19. The data storage device in the storage device containing the retaining Pris is osobine for storage devices a recording device for recording data in the storage device, the receiver input signal and a device for inverse filtering of the signal according to clause 16, is connected to the receiver input and recording device.

20. A device for processing data containing the receiver input for receiving input data, a device for synthesizing filtering signal 17, communicative connected to the receiver input signal to filter mentioned input data and outputting the filtered input data, and means for processing the aforementioned filtered input data.

21. Audiovisual device containing the input device data, the device for synthesizing filtering signal 17, is connected by its input to the input device for the data, and audio-visual output device, connected by its input to output device for synthesizing filtering of the signal.

22. Audiovisual recording device that contains audio-visual input device, a device for inverse filtering of the signal according to clause 16, is connected by its input to output audiovisual input device, and output device for data connected with its input to output device for inverse filtering of the signal, in fact the output device to the data made with the possibility of connection with the storage device, for data storage.

23. The coding system that contains the encoder; a decoder, communicative connected to the encoder, and the encoder includes at least one device for inverse filtering of the signal by article 16, and the decoder includes at least one device for synthesizing filtering of the signal by 17.

**Same patents:**

FIELD: radio engineering, possible use in coherent-impulse radio location stations for detecting and controlling aerial traffic for selecting signals from moving targets on background of passive interference of unknown correlation properties.

SUBSTANCE: rejection of passive interference by means of a parallel rejecter filter is based on m-channeled (with m=4) decomposition of vector H of rejecter filter of n-numbered order (with n=7) with integer-valued weight coefficients on channel vectors H_{1},H_{2},...,H_{m}, to projections of which h_{i}(m) integer-value condition is applied. To achieve aforementioned result, in conjunction with first channel, including seven memory blocks and first weight adder, additional channels are utilized, wherein accumulated rejection remainders are processed.

EFFECT: increased efficiency of rejection of passive interferences.

2 cl, 2 dwg, 4 tbl

FIELD: radio engineering, possible use in systems for digital processing of speech and images in real time scale.

SUBSTANCE: composition of digital filter in system of residual classes includes (dwg.1) generator of harmonic signal 3, transformer of residual classes system code to code of positional notation 5 and N modulus calculators 4m each one of which contains (dwg.2) transformers of positional code to modular discontinuous-logarithmic form 7, shift registers of digital signal counts 8.1 and shift registers of digital counts of impulse filter characteristic 8.2, modulus adders 9, transformer of code from discontinuous-logarithmic to modular form 10, controllable phase shifters 11 and phase shift meter 12.

EFFECT: increased speed of modulus arithmetic operations during computation of differential equation of digital filter.

2 dwg

FIELD: radio electronics; linear permanent-parameter difference circuits.

SUBSTANCE: proposed signal convolution device that can be used, for instance, to construct electronic analog filters of desired phase-frequency response characteristic close to ideal one has elementary phase circuits, proportional sections, and adders; elementary phase circuits are integrated into three groups incorporating equal number of elementary phase circuits. Picked off output of first group is input signal delayed copy. Two-input adders are inserted between elementary phase circuits of two other series-connected groups. Each adder is connected through one of its inputs to output of respective proportional section. Inputs of all proportional sections whose gain depends on synthesized difference circuit characteristic turned over from left to right are integrated and connected to input of first group of elementary phase circuits to form input signal bus. Output signal is picked off last adder.

EFFECT: enhanced speed and ability of implementing real-time convolution operation.

1 cl, 1 dwg

FIELD: radio engineering; real-time digital speech and image processing systems.

SUBSTANCE: proposed digital filter in residual-class system has harmonic signal generator, residual-class-to-positional-number code converter, and N modulo calculators, each incorporating positional-to modular digital-logarithmic code converter, digital signal reading shift registers, filter pulse characteristic digital reading shift registers, modulo adders, digital/arithmetic-to-modular code converters, controlled phase shifters, and phase shift meter.

EFFECT: enhanced speed of modulo arithmetic operation execution in calculating difference equation of digital filter.

1 cl, 2 dwg

FIELD: automatics and analog computer engineering, possible use for building functional nodes of analog computers, means of automatic adjustment and control, analog processors, etc.

SUBSTANCE: rank filter contains seven relators P 1_{1}-1_{7} (dwg.1), each one of which contains comparator 2, connected by output to first input of XOR element 3, second input of which is input for controlling adder frequency, while output is connected to controlling input of closing and opening keys 4_{1},4_{2}, outputs of which are combined and form output of relator, first and second inputs of which are, respectively, inputs K 2, connected appropriately to inputs of keys 4_{1},4_{2}. Rank filter (dwg.1) by means of two controlling signals (f_{1},f_{2}) realizes selection of minimal, super-minimal, sub-maximal or maximal one of four input analog signals (x_{1}-x_{4}) and has higher speed of operation compared to prototype.

EFFECT: increased speed of operation with preservation of functional capabilities of prototype.

2 dwg

FIELD: automatics and analog computer engineering, possible use for building functional nodes of analog computers, means of automatic adjustment and control, analog processors, etc.

SUBSTANCE: rank filter contains five relators P 1_{1}-1_{5} (dwg.1), third-fifth of which contain comparator 2, connected by output to first input of XOR element 4, second input of which is control input P (f), and output is connected to controlling input of keys (3_{1}, 3_{2}), outputs of which are combined and form output P, first and second inputs of which are inputs K 2, connected respectively to inputs of keys 3_{1}, 3_{2}. First and second relators contain inputs K 2 and additional keys 3_{3},3_{4}, outputs of which are additional output of relator P. Rank filter (dwg.1) by means of two controlling signals f_{1},f_{2} realizes selection of minimal, super-minimal, sub-maximal or maximal one of four input analog signals x_{1}-x_{4} and has structure, simpler compared to that of the prototype, due to decreased number of controlling inputs.

EFFECT: simplified structure due to decreased amount of controlling inputs.

3 dwg

FIELD: automatics and analog computer engineering, possible use for building functional nodes of analog computers, means of automatic adjustment and control, analog processors, etc.

SUBSTANCE: rank filter contains eight relators P 1_{1}-1_{8} (dwg.1), each one of which contains comparator 2, connected by its output to first input of XOR element 3, second input of which is relator control input f, while output is connected to controlling input of closing and opening keys 4_{1},4_{2}, outputs of which are combined and form the output of relator together, first and second inputs of which are inputs K 2, connected respectively to inputs of keys 4_{1},4_{2}. Rank filter (dwg.1) by means of two control signals f_{1},f_{2} realizes selection of minimal, super-minimal, sub-maximal or maximal one of four input analog signals x_{1}-x_{4} and has decreased hardware resource costs when compared to the prototype.

EFFECT: decreased hardware resource costs with preservation of functional capabilities of the prototype.

2 dwg

FIELD: computer science.

SUBSTANCE: device has sum forming blocks, matrix functions forming block, difference forming block, delay lines, apriori data output block.

EFFECT: higher precision.

6 dwg

**FIELD: radio engineering; construction of radio communication, radio navigation, and control systems using broadband signals.**

**SUBSTANCE: proposed device depends for its operation on comparison of read-out signal with two thresholds, probability of exceeding these thresholds being enhanced during search interval with the result that search is continued. This broadband signal search device has linear part 1, matched filter 2, clock generator 19, channel selection control unit 13, inverter 12, fourth adder 15, two detectors 8, 17, two threshold comparison units 9, 18, NOT gates 16, as well as AND gate 14. Matched filter has pre-filter 3, delay line 4, n attenuators, n phase shifters, and three adders 7, 10, 11.**

**EFFECT: enhanced noise immunity under structural noise impact.**

**1 cl, 3 dwg**

FIELD: analysis and synthesis of speech information outputted from computer, possible use in synthesizer-informers in mass transit means, communications, measuring and technological complexes and during foreign language studies.

SUBSTANCE: method includes: analog-digital conversion of speech signal; segmentation of transformed signal onto elementary speech fragments; determining of vocalization of each fragment; determining, for each vocalized elementary speech segment, of main tone frequency and spectrum parameters; analysis and changing of spectrum parameters; and synthesis of speech sequence. Technical result is achieved because before synthesis, in vocalized segments periods of main tone of each such segment are adapted to zero starting phase by means of transferring digitization start moment in each period of main tone beyond the point of intersection of contouring line with zero amplitude, distortions appearing at joining lines of main tone periods are smoothed out and, during transformation of additional count in the end of modified period of main tone, re-digitization of such period is performed while preserving its original length.

EFFECT: improved quality of produced modulated signal, allowing more trustworthy reproduction of sounds during synthesis of speech signal.

2 cl, 8 dwg

FIELD: digital speech encoding.

SUBSTANCE: speech compression system provides encoding of speech signal into bits flow for later decoding for generation of synthesized speech, which contains full speed codec, half speed codec, one quarter speed codec and one eighth speed codec, which are selectively activated on basis of speed selection. Also, codecs of full and half speed are selectively activated on basis of type classification. Each codec is activated selectively for encoding and decoding speech signal for various speeds of transfer in bits, to accent different aspects of speech signal to increase total quality of synthesized speech signal.

EFFECT: optimized width of band, required for bits flow, by balancing between preferred average speed of transfer in bits and perception quality of restored speech.

11 cl, 12 dwg, 9 tbl