Method for improved detection of speed errors in receivers with varying speed and device for realization of said method

FIELD: radio engineering, possible use in wireless communication systems.

SUBSTANCE: in the method, frames are found for which data transfer speeds have been determined incorrectly, and processing is performed, providing for erasing of frames and/or clearing of memory to prevent expansion of distortion among multiple frames. Frames with erratically determined speeds of data transfer are found by checking unsanctioned transitions from one speed to another and reserve bits, confirming correctness of unused combinations of bits for filter and analyzing connections between transfer coefficients of fixed encoding table and transfer coefficients for linear prediction.

EFFECT: decreased appearance of sound artifacts, such as creaks or background signals, caused by errors in speed determining algorithm.

3 cl, 12 dwg

The technical field to which the invention relates.

Disclosed here relate to wireless communication. In particular, the disclosed variants refer to a new and improved method and apparatus for detection in the receiver of a communication system with variable speed transmission errors in determining the speed at which were encoded data for transmission.

The level of technology

In Fig. 1 shows a block diagram of a system 10 transmission multiple access, code-division multiplexing (mdcr, CDMA) with variable speed described in the standard radio interface industry Association telecommunications TIA/EIA : Interium Standard 95 (interim standard) and its modifications, such as, for example, IS-V (referred to here collectively as the is-95). This transmission system can be provided, for example, in the base station transmission system of cellular communication for use in transmitting signals to the mobile subscriber units within the boundaries of cell cells surrounding the base station. This system can also be provided in the mobile telephone subscriber units for use in transmitting signals to the base station.

The microphone 11 perceives the speech signal, which is then sampled and digitized by analog-to-digital Converter (not shown). The data source 12 variable speed before the Chi receives digitized samples of the speech signal and encodes the signal, creating packages coded speech, consisting of frames of equal length. The data source 12 variable speed transmission may, for example, to carry out the conversion of the digitized samples of the input speech signal in the digitized speech parameters representing the input voice signal using the encoding method of linear prediction (LP, LPC). In the example embodiment, the data source variable speed transmission is the vocoder variable speed, described in detail in U.S. patent No. 5414796, the rights to which are owned by the assignee of the present invention and the content of which is included here as a reference. The data source 12 variable speed transmission generates data packets with a variable speed transmission having four possible values of the speed (frequency) frame rate: 9600 bits per second (bps), 4800 bps, 2400 bps and 1200 bps, called here at full speed, half speed, quarter speed, and rate 1/8 (full speed). Packets encoded with full speed, contain 172 information bits; packets encoded at half speed, contain 80 information bits; packets encoded with a quarter speed, contain 40 information bits, and packets encoded with the speed of 1/8, contain 16 information Bito is. The packet formats shown in Fig. 2A-2D. All packages regardless of size have a duration of one frame, that is 20 MS. Here, the terms "frame" and "packet" can be used interchangeably. The packets are encoded and transmitted at different speeds to compress the data contained in them on the basis of, in particular, the complexity or the amount of information presented in the frame. For example, if the input speech signal contains minor changes or no changes, perhaps due to silence the speaker, the information bits corresponding package can be compressed and encoded with a speed of 1/8. This compression leads to loss of resolution of the relevant part of the speech signal, but if the relevant portion of the speech signal contains little information or information in General is not present, then the lower resolution signal is usually insignificant. Alternatively, if the corresponding input speech signal package contains a lot of information, perhaps due to the fact that the speaker is actively emits different sounds, the packet is encoded at full speed and compression of the input speech signal is reduced to provide a higher quality of speech.

This method of compression and encoding is used to limit the average amount of information transmitted on any in the temporal interval, that allows more efficient use of the entire bandwidth of the transmission system, revealing, for example, the ability to handle a larger number of telephone calls at any time interval.

Packets with variable speed created by the data source 12, are fed into the device 13 packaging that selectively adds the bits of the control performed using cyclic redundancy code (CEC, CRC and tail bits. As shown in Fig. 2A, when the frame is encoded by the data source 12 variable speed with full speed, device packaging 13 creates and adds twelve bits of the CEC and eight tail bits. Similarly, as shown in Fig. 2B, when the frame is encoded by the data source 12 variable speed to half speed, the device 13 packaging creates and adds eight bits of the CEC and eight tail bits. As shown in Fig. 2C, when the frame is encoded by the data source 12 variable speed with a quarter speed, the device 13 packaging creates and adds eight tail bits. As shown in Fig. 2D, when the frame is encoded by the data source 12 variable speed with speed of 1/8, the device 13 packaging creates and adds eight tail bits.

Then packets with variable speed from device 13 packaging is served in the encoder 14, which is coding for the t bits of packets with variable speed for detection and correction of errors. In the example embodiment, the encoder 14 is a convolutional encoder with rate 1/3. Then the characters, past convolutional coding, served in the extender mdcr 16, which is described in detail in U.S. patent No. 5103459 and No. 4901307. The extender mdcr 16 converts the eight coded symbols in a 64-bit Walsh symbol, and then expands the characters Walsh in accordance with a pseudorandom noise (PN, PN) code.

The generator 17 reps adopts advanced packages. For packages with a velocity that is less than full speed, the generator 17 repetitions creates copies of the symbols in packages creating packages with a constant data rate. When a packet with variable speed is half the speed, the generator 17 reps introduces a redundancy factor equal to two, i.e. each wide character is repeated twice in the output package. When a packet with variable speed is a quarter the speed, the generator 17 reps introduces a redundancy factor equal to four. When a packet with variable speed is the speed of 1/8, the generator 17 reps introduces a redundancy factor equal to eight.

The generator 17 reps above provides redundancy by splitting package extended data into smaller subpacket, here called "group power control". In given the nom in an example embodiment, each power control contains 6 PSH extended characters Walsh. Frame with constant velocity is created by consistent repetition of each group power control repeatedly to fill the frame, as described above.

Then extended packages are randomization 18 packets of data, which eliminates the redundancy of extended packet in accordance with a pseudorandom process described in U.S. patent No. 5535239, the rights to which are owned by the assignee of the present invention. Randomization 18 data packets selects one of the advanced groups power control for transmission in accordance with a pseudo-random process of selection and does not let other redundant copies of the power control.

From randomization 18 packets data packets are served in filter 20 with a finite impulse response (FIR, FIR), an example of which is described in U.S. patent No. 5659569, the rights to which are owned by the assignee of the present invention. Then the filtered signal is supplied to d / a Converter 22 and is converted into an analog signal. After this analog signal is fed to the transmitter 24, which converts it with increasing frequency and increases for transmission via the antenna 26.

In Fig. 3 shows the components of the base station related to the invention. In another embodiment, the device according to figure 3 can be in mobile is Mr. the telephone 28 or another mobile station, receiving a transmitted signal. The signal is received by the antenna 30, is converted with decreasing frequency and amplified, if necessary, by receiver 32. Then the signal goes into the block 33 determine the frame rate, which divides the signal into packets and determines the appropriate frame rate for each package. The frame rate can be defined, depending on the implementation, by determining the duration of individual bits in the frame. Then package and a signal identifying found the frame rate for this package, proceed to the block 34 by the CEC to perform inspections using a cyclic redundancy code or associated checks to detect errors, to try to ensure that there are no transmission errors or errors of determining the frame rate. The error in determining the frame rate leads to a discretization of the package with the wrong frequency that generates the sequence of bits that are effectively random. The transmission error usually leads to only one or two erroneous bits. Usually, if there is a transmission error or the detection error rate frames, the block of the CEC identifies this error. Defective frames that have not been monitoring the CEC, erased or otherwise discarded unit 36 erasing frames. "Good" frames, the past is dsie control CEC, sent to the decoder 40 with variable speed to convert back to digitized voice signals. The digitized speech signals are converted into analog signals by the d / a Converter (not shown) for final output through the speaker 42 of the mobile phone.

Depending on the specific implementation of the separate unit 36 erasing frames may not be required. Preferably the block 34 CEC to configure in such a way that he just did not take the defective frames in the decoder 40 with variable speed. However, the presence of block erase frame facilitates the formation of signals erase personnel for their direction back to the base station to notify the base station error due to the Erasure frame. The base station can use the information on erasing frames to modulate the power level used for transmitting signals, possibly as part of a feedback system (closed system)designed to minimize the transmitted power along with minimizing errors in frames.

As noted above, the speed change of frames in the packet to compress the information contained within the entire bandwidth of the system is used more efficiently, usually without any noticeable effect on the transmitted signal. However, from time to time problems arise is s, which have a significant impact on the quality of the signal. One such problem arises when the frame, despite the presence of errors determine the speed of frames or transmission errors, however, passes control of the CEC. In these cases, the defective block is not erased and is processed together with other "good" frames. The error may be noticeable or negligible. For example, if a transmission error when the wrong is only one or two bits of encoded speech, this error can have only a very weak and likely negligible effect on the output speech signal. However, if this is an error in determining the frame rate, because of the presence of the entire package will be processed using an incorrect frame rate, which will cause the input to the decoder bits with effectively random values, and this, in all likelihood, will lead to noticeable artifacts in the output speech signal. Significant artifacts caused by errors of determining the frame rate are unacceptable distortion, as squeaks or buzzer signals. It was found that in some systems, the cases of incorrect definition of speed frames appear with a probability of roughly 0.005%, resulting in an incorrectly received packet and the corresponding artifact in the output speech signal CA is RNO every 16 minutes for talk time. Although described problems are related to the system mdcr using the protocols IS-95, with the same problems you might encounter in almost any transmission system, which uses variable speed transmission, as well as related systems. Due to the influence on the reception parameters such channel condition as noise and fading due to multipath propagation of the signal, the use of algorithms to determine the speed (AOC RDA)used in blocks 33 determine the frame rate, does not guarantee the accuracy of determining the rate of received frames. Given that this fact is a limitation of the AOC, it is desirable to ensure that these errors EPA did not cause significant sound distortion, such as squeaks or buzzer signals. If the received frame is not suitable for accurate determination of speed due to poor conditions in the channel, the algorithm AOC or determines that the frame should be wiped off, or he assigns to the package the wrong speed. Usually speech decoder is processing engine, providing erasing frames that perceptual smooths lost frames using past frames, and reproduces it, that is not annoying companion. However, if instead of erasing the frame algorithm ACO assigns the wrong frame rate, the decoder 40 with variable soon the TEW served random bits. If the random bits are not detected, they can create a very loud unpleasant artifacts such as squeaks and buzzer signals. In General it can be argued that the Erasure rate is not so strongly degrades the speech quality, as a frame with an incorrect speed.

It is desirable to process these frames with incorrect speed without creating audible artifacts. It is desirable to determine the frame with the wrong speed and process, providing the erasing frame, and/or to clear the memory in the decoder 40 with variable speed so that the consequences of failing to determine the frame rate does not spread to a large number of frames.

Thus, it can be argued that there is an urgent need for the method, which detects errors of determining the frame rate in the wireless communication system and eliminates audible artifacts caused by these errors.

The invention

Open options aimed at creating a system and method for detecting errors generated by the algorithm for determining the frame rate, the receiver of a communication system with variable speed (Fig. 4). Accordingly describes how error detection rate in the receiver variable speed, namely, that accept the encoded speech signal, performs an algorithm for determining the speed of said the second signal to provide a coded speed and detect errors in the provided rate.

In another embodiment describes a system for error detection speed, contains a receiver for receiving the encoded speech signal, an element for determining the speed of executing an algorithm for determining the speed on the speech signal to provide a coded speed, and detector error rate for detecting errors in the provided rate.

Brief description of drawings

The features, objectives and advantages of the disclosed variants will become more apparent from the following detailed description together with the drawings in which the same reference position identify corresponding elements in all the drawings and where:

Fig. 1 is a block diagram of a known transmitter of the base station of a digital cellular telephone system;

Fig. 2A-2D - known formats of frames used in the system of figure 1;

Fig. 3 is a block diagram of a known reception of the cellular phone configured disclosed without options, for receiving signals transmitted by the system of figure 1;

Fig. 4 is a block diagram of the receiving part of the mobile subscriber unit, configured in accordance with the disclosed variants detector error rate, for receiving signals transmitted by the system of Fig. 1;

Fig. 5 is a block diagram of a method for error detection rate in frames, identified as frames at full speed;/p>

Fig. 6 is a block diagram of a method for error detection rate in frames, identified as frames at half speed;

Fig. 7 is a block diagram of a method of error detection rate in frames, identified as frames with a quarter speed;

Fig. 8 is a block diagram of a method of error detection rate in frames, identified as frames with a speed of 1/8;

Fig. 9 is a graph illustrating an example of the dependence of the gain of the fixed codebook (code tables) in the form of a threshold curve for encoding the linear prediction method (CLP).

A detailed description of the preferred variants of the invention

Given as an example version of the advanced error detection speed receivers with variable speed is implemented in the vocoder selectable mode (WWR, SMV). WWR is a vocoder variable rate, which is a candidate for use of the system mdcr third generation IS2000. In the algorithm vocoder VVR is used by many different parameters, such as speed, controlled by source, type of frame, the transfer coefficients of the linear prediction (LP), adaptive and fixed parameters table encoding. Speech encode reviewed the contained amount of perceptual information. T is some analysis allows us to classify it into different types, such as background noise, stationary newcaledonia it (whisper), stationary vocalized speech and non-speech (initial consonant syllables, sharp transitions, and so on). Miraculou background noise is encoded using a rate 1/8. Stationary newcaledonia speech is encoded using a linear prediction scheme, excited by the noise (LPVS, NELP), with a quarter speed. Stationary vocalized speech is encoded using a linear prediction scheme with a code excitation (LPCV, CELP) type 1 with full or half speed. Non-stationary speech is encoded using the scheme LPCB type 0 with full or half speed. Information about the type regulates several aspects of the coding frame, such as the size of Subhadra, the parameters used for speech representation and coding scheme for these parameters. Frames of type 0 are "non-periodic" footage, in which the model parameters, such as correlation and lag of the fundamental tone (the speech signal), can change quickly. Thus, in circuits LPCB type 0 delay pitch is encoded and transmitted more frequently (that is, for each subcode). Frames of type 1 are "periodic" frames, which have a high frequency and is represented by the smooth track pitch with good the perception is. In the scheme LPCB type 1 lag of the fundamental tone is encoded in each frame and this lag receive the interpolated track pitch. Due to the high frequency and smooth track pitch transfer coefficients of the primary colors have a very high stability and quanthouse collectively. To specify the schema type LPCV use one bit of each frame stationary vocalized and non-stationary speech.

Specialists in the art it is obvious that the vocoder VVR can be implemented using gate arrays, programmable by the user during operation (FPGA), programmable logic devices (PLD), digital signal processors (DSP), one or more microprocessors, a specialized integrated circuit (ASIC) or any other device capable of performing the above functions vocoder VVR.

Open the options described in the context of phones mdcr. However, it should be understood that disclosed here is applicable to communication systems and methods for modulation of other types, such as personal communication system (PCS), wireless local communication line (WLL), phone system for private use (PBX) or other known systems. In addition, the disclosed variants can be used in systems that employ other Izv the local modulation scheme, for example, in systems multiple access with time division multiplexing (MDR TDMA) systems, multiple access, frequency division multiple access (FDMA equipment, FDMA), and other systems with spread spectrum.

According to one variant of figure 4 shows the corresponding components of the mobile subscriber unit 28 or another mobile station, receiving the signal generated by the transmission system base stations, such as the system of figure 1, in which the signal is transmitted in the form of packets with variable speed. Frame rate include full speed, half speed, quarter speed, and 1/8 as shown in figa-2D. Packages include the parameters of the encoded speech signal representing a compressed speech signal. In addition, each package contains bits CEC and/or the tail bits of the encoder. Additional details related to the content of the packages presented above in connection with figure 1 and the above-mentioned U.S. patent No. 5414796.

Shown in figure 4, components similar to the components in figure 3 and further details, only the relevant differences. The transmitted signal is received by the antenna 30, and then converted with decreasing frequency and amplified by the receiver 32. After that, the signal is supplied to the block 33 determine the frame rate, which seeks to identify suitable the second frame rate for a given package, using the algorithm of definition of speed (AOC). Then, the packet is served in the block 34, the CEC to perform control frames received signal using a cyclic redundancy code when you try to verify the absence of errors determine the speed of frames or transmission errors. Frames that have not been monitoring the CEC, that is defective frames are erased block 36 erasing frames. As noted above, the presence of individual block erase frames is not required. It is preferable that the frames with detected errors CEC simply not output from block 34 of the CEC. In any case frames that have passed the CEC, that is potentially "good" frames are sent to the detector 38 error rate. Depending on the implementation of a separate block 38 bug detector speed is not required. It is preferable that the block 38 detector error rate was implemented in the vocoder VVR or was integrated with other components of the receiver.

The detector 38 error rate, in addition, performs analysis of frames, to ensure that the frame rate defined by the algorithm AOS unit 33 determine the frame rate. In addition, frames are checked by the detector 38 error rate using methods validation for frames with full, half, quarter speed and speed 1/8, described in detail below with reference to Fig.-8. Frames are not authenticated can be erased by block 36 erasing frames. Frames are not authenticated can also be treated in this way to clear the memory in the decoder 40 with variable speed, so that the distortion was not extended to many frames. Information about the speed, management, and frames output from detector 38 error rate at the decoder 40 with variable speed for processing, provide clearance. Frames that have been tested in the detector error rate, directly to the decoder 40 with variable speed.

The decoder 40 variable speed processes the frames by decoding contained therein parameters of speech signals for the inverse transform to the digitized speech signals. In the end, the digitized speech signals are converted into analog signals by the d / a Converter (not shown) for output through the speaker 42 to the listener, if the receiver is a mobile subscriber unit. If the receiver is a base station, the digital signal may be distributed later in the wireless system.

Figure 5-8 are described in detail how to verify that the frame rate in accordance with variations for frames with full, half, quarter speed and the speed of 1/8, which are implemented by the detector error rate (see figure 4, item 38). In ways which checks for new unauthorized use according to the classification of transitions for frames with certain speeds and types, check back bits, check for unauthorized type of filter and the analysis of the curves of the transmission factor of the fixed code tables (FTC, FCB) from the threshold gear ratio for CLP. Additionally, in the disclosed embodiments, for 5-8 new used treatment, providing erasing frames, and processing memory States to smooth the effects of errors detected frame rate.

In disclosed embodiments, it is recommended to use the new structure transitions from one state to transition from one speed to another consecutive frames based on the information on the classification of speech and phonetic features of speech. Transitions from one speed to another, violates this structure are unauthorized and are used for error detection frame rate. These unauthorized transitions from one speed to another is determined in the following list:

the frame rate 1/8 behind the scenes at full speed;

the frame rate 1/8 behind the scenes type 1 with full speed;

the frame rate 1/8 behind the scenes type 1 at half speed;

frame with full speed type 1 followed by a frame with quarter speed;

frame with half speed type 1 followed by a frame with quarter speed;

frame full / min net is using type 1 followed by a frame with a speed of 1/8;

frame with half speed type 1 followed by a frame with a speed of 1/8;

the frame rate 1/8 behind the scenes with quarter speed, which follows the frame rate 1/8;

the frame rate 1/8 behind the scenes at half speed, which follows the frame rate 1/8;

the frame rate 1/8 behind the scenes with full speed, which follows the frame with the speed of 1/8.

Based on the speeds and types of the current and previous frames by the presence of unauthorized crossings define the error AOC or in the current frame or the previous frame.

In disclosed embodiments, on the new used spare bits transmitted at full speed and quarter speed for error detection AOC. Package with full speed has 171 information bits per 20-millisecond frame in which 1 bit is reserved. The encoder can be set to a fixed value reserve bits equal to either zero or one. Backup bit is checked by the error detector speed (see figure 4, item 38)to determine whether the backup bit expected fixed encoded value. If you receive spare bits have unexpected values, this indicates that the current frame with full speed there is a bug AOC. The quarter package with speed is 40-bit to 20-millisecond frame in which the DL is schema LPVS used 39 bits, and one bit is not used. The encoder can also set a fixed value unused bits equal to either zero or one. Unused bits are checked by the detector error rate (see figure 4, item 38)to determine whether the adopted unused bits of the expected fixed encoded value. If the values obtained unused bits do not match the expected, this indicates the presence of errors AOC in the current frame with quarter speed.

In disclosed embodiments, on the new validation is used unauthorized type of filter to frame schemes LPVS with quarter speed, detecting the error rate. When encoding scheme LPVS used spectral formation of pseudo-random excitation using one of 3 different shaping filters. Two bits are used for transmission of the index of the selected filter. Three of the two bit combinations are used to identify the selected shaping filter, leaving the fourth 2-bit combination of unused or prohibited. The presence of unused or illegal combination indicates an error AOC in the current frame schemes LPVS with a quarter speed.

In disclosed embodiments, on the new using the encoded parameters for error detection speed. Research who is Astia errors AOC on vocoders showed what cause audible artifacts, such as squeaks and buzzer signals are generally very high values of the transfer coefficients of the FTC, accompanied by high values of the transmission coefficients for predicting when CLP. Natural speech in its analysis by the encoder, to encode the parameters, generates transfer coefficients of the FTC and the transmission ratios to predict when CLP, which are connected to each other by a reverse dependence. In other words, when the gear ratio when CLP large, the transmission coefficient FTC in General small, and when the gear ratio when KLP small, the transmission coefficient FTC usually large.

An inverse relationship between the transfer coefficients of the FTC and transmission coefficients at CLP in natural speech is characterized by a curve on the above graph, the dependence of the transfer coefficient from the FTC gear ratio when CLP where good natural speech is not presented. The transmission coefficient of the FTC, and therefore, the curve on the graph, may be a function of the level of the input speech signal. Frames received with levels that lie above this curve, where good natural speech is not presented indicate that an error rate in this frame. A new way to eliminate variability due to changes in the level of the input signal when the error rate OBN is razivaetsya graphically is the normalization factor FTC based on average energy calculated on the basis of previous frames. Figure 9 shows the scattering curve showing the relationship between the normalized transmission coefficient FTC and the transfer coefficient for predicting PL. Circles under the solid curve was obtained for the case of clean speech, and the stars above the solid curve correspond to unacceptable squeaks caused by errors AOC. This solid curve represents the threshold curve, which separates the region of a good speech from areas of unacceptable squeaks and other distortions. This threshold curve can easily be represented in a parametric form and enter in the detector error rate (see figure 4, item 38). After the received packet is set to the transmission factor of the FTC and the transmission ratios when CLP, you can perform a check to determine lies whether the frame is below the threshold curve. If the frame does not lie below the threshold curve, this indicates the presence of scratches due to errors AOC.

In disclosed embodiments, on a new used transmission coefficient for excitation at a speed of 1/8 for error detection speed. Because the coding scheme with rate 1/8 is used only for periods of speech with background noise, the excitation energy, which is quantized with used is an option gain has an upper limit value. When the received packets received transfer coefficient for excitation, it is possible to perform a check to determine lies whether the transmission coefficient for excitation below the upper limit for this parameter. If the transmission coefficient for excitation does not lie below the upper limit value, it indicates the presence of errors AOC.

When any of the disclosed here mechanisms of error detection indicates an error in the speed of the frames in the current frame or in the immediately preceding frames, in disclosed embodiments may use one or several new schemes in the decoding process to eliminate the resulting distortion and/or prevent the spread of this distortion on many frames. These schemes include processing, providing erased frames, the reduction gear ratio FTC and reset the memory.

Vocoders typically have built-in capability to implement the erase process frames. The process of erasing frames can be used in the disclosed here available for any frame in which the error detector speed (see Fig. 4, item 38) encountered an error AOC. The process of erasing frames synthesizes speech without using any information from the current frame and receives the CE decoder parameters from the previous memory state, to create speech, perceptual smoothed with respect to previous frames. If an error is detected speed as a result of Riem frame with a level lying above the curve showing the dependence of the transfer coefficient from the FTC gear ratio when CLP where good natural speech, the decoder (Fig. 4, item 40) can forcibly reduce the transmission coefficient FTC to smaller values that will prevent squeaking or buzzer signal with high energy at the output of the decoder (Fig. 4, item 40).

The vocoder algorithms usually recover it using the past state memory. These state of the memory include memory status quantizer vector moving average (CWSS, MAVQ) to gain FTC, memory, arousal, memory synthesis at CLP and memory of the synthesis filter. Undetected error AOC may lead to putting these tools memory faulty values. The impact of faulty values can persist for many frames in the future, even if all subsequent frames are "healthy". To prevent damage in future frames, errors EPA, found in the current or immediately preceding frames, call record in memory CWSS to gain FTC, memory, arousal, memory, filter synthesis at CLP and memory f is ltr for the synthesis of the output filter of the prescribed values, which will not allow you to create crackles with the energy of a high level. In one embodiment, the memory values correspond to zero values. In another embodiment, the memory values are overwritten by the corresponding values set to their initialization.

In Fig. 5 shows a method according to one variant of the invention for error detection rate in frames, the algorithm identifies the AOC as frames with full speed. Specialists in the art will understand that the order of steps shown in Fig. 5, is not a limitation. The way you can easily edit by omitting or reorder the stages shown, without departing from the scope of the disclosed here options.

Phase detector 502 error rate enters the data frame specified by the algorithm AOC as a frame with full speed. Control then passes to step 504.

At step 504 is checked back a bit or bits ready to determine is whether the obtained value is a fixed value set by the encoder. If this bit is not equal to a fixed value set by the encoder that indicates the error rate of the frame, then control passes to step 506. Otherwise, control passes to step 510.

At step 506 executes processing that provides the erased frame, and/or treatment, provide the General restoration of the original state memory. Control passes to step 508, which continues decoding.

At step 510 checks the frame to determine the frame type. If this frame is 0, then control passes to step 512. If the frame type 1, then control passes to step 520.

At step 512 for frame type 0 are the transfer coefficients of the FTC and CLP for the frame and checks that determines lies whether this frame is below the threshold curve. If the frame is below the threshold curve, then control passes to step 514, which continues decoding. If the frame does not lie below the threshold curve, then control passes to step 516.

At step 516 can be processed, providing the erased frame, and/or the reduction gear ratio, and/or treatment, which restores the original state of the memory. Control passes to step 518, which will continue decoding.

At step 520 for frame type 1 is checked previous frame to determine whether this frame by frame with a speed of 1/8 or frame with quarter speed. If the previous frame is not a frame with a speed of 1/8 or quarter speed, which indicates that sanctioned the transition from one speed to another, the control proceeds to step 526, which continues decoding at full speed. If the previous is th frame is a frame with a speed of 1/8 or quarter speed, that indicates unauthorized switch from one speed to another, then control passes to step 522.

At step 522 can be processed, providing the erased frame, and/or treatment, which restores the original state of the memory. Control passes to step 524, which continues decoding.

In Fig. 6 shows a method according to one variant of the invention for error detection rate in frames, the algorithm identifies the AOC as frames at half speed. Specialists in the art will understand that the order of steps shown in Fig. 6, is not a limitation. The way you can easily edit by omitting or reorder the stages shown, without departing from the scope of the disclosed here options.

At step 602 detector error rate enters the data frame specified by the algorithm AOC as a frame with half speed. Control then passes to step 604.

At step 604 checks the frame type. If this frame is 0, the control proceeds to step 606. If the frame type 1, then control passes to step 614.

At step 606 for frame type 0 are the transfer coefficients of the FTC and CLP for the frame and checks that determines lies whether this frame is below the threshold curve. If the frame is below the threshold to the willow, then control passes to step 610, which continues decoding. If the frame does not lie below the threshold curve, then control passes to step 608.

At step 608 can be processed, providing the erased frame, and/or reducing the transfer coefficients of the FTC, and/or treatment, which restores the original state of the memory. Control passes to step 612, which continues decoding.

At step 614 to frame 1 of the previous frame is checked to determine whether this frame by frame with a speed of 1/8 or frame with quarter speed. If the previous frame is not a frame with a speed of 1/8 or quarter speed, which indicates that sanctioned the transition from one speed to another, the control proceeds to step 620, which continues decoding at half speed. If the previous frame is a frame with a speed of 1/8 or quarter speed, which indicates that an unauthorized switch from one speed to another, then control passes to step 616.

At step 616 can be processed, providing the erased frame, and/or treatment, which restores the original state of the memory. Control passes to step 618, which continues decoding.

In Fig. 7 shows a method according to one variant from which retene for error detection rate in frames, the algorithm identifies the AOC as frames with a quarter speed. Specialists in the art will understand that the order of steps shown in Fig. 7, is not a limitation. The way you can easily edit by omitting or reorder the stages shown, without departing from the scope of the disclosed here options.

At step 702, the error detector speed enters the data frame specified by the algorithm AOC as a frame with a quarter speed. Control then passes to step 704.

At step 704 checks back a bit or bits ready to determine is whether the obtained value is a fixed value set by the encoder. If this bit is not equal to a fixed value set by the encoder, which indicates the presence of errors, frame rate, then control passes to step 706. Otherwise, control passes to step 710.

At step 706 processed, providing the erasing frame, and/or treatment, which restores the original state of the memory. Control passes to step 708, which continues decoding.

At stage 710 is checked case of double-bit combination used to identify the selected shaping filter. If case of double-bit combination is valid, then control passes to step 716, which continues decode the Finance with a quarter speed. If the case of double-bit combination is not valid, then control passes to step 712.

At step 712 is performed processing, providing the erasing frame, and/or treatment, which restores the original state of the memory. Control passes to step 714, which continues decoding.

In Fig. 8 shows a method according to one variant of the invention for error detection rate in frames, the algorithm identifies the AOC as frames with a speed of 1/8. Specialists in the art will understand that the order of steps shown in Fig. 8, is not a limitation. The way you can easily edit by omitting or reorder the stages shown, without departing from the scope of the disclosed here options.

At step 802 detector error rate enters the data frame specified by the algorithm AOC as a frame with a speed of 1/8. Control then passes to step 804.

At step 804 previous frame is checked to determine whether this frame by frame with full speed. If the previous frame is not a frame with full speed, which indicates that sanctioned the transition from one speed to another, the control proceeds to step 810. If the previous frame is a frame with full speed, which indicates that an unauthorized switch from one speed to another, the control lane who goes to step 806.

At step 806 can be processed, providing the erased frame, and/or treatment, which restores the original state of the memory. Control passes to step 808, which continues decoding.

At step 810 checks the previous frame to determine whether this frame by frame with quarter speed, half speed or full speed. If the previous frame is a frame with quarter speed, half speed or full speed, indicating possible unauthorized switch from one speed to another, the control proceeds to step 820. If the previous frame is not a frame with quarter, half, or full speed, which indicates that authorized the transfer speed 1/8, then control passes to step 812.

At step 812, the transfer coefficient for excitation at a speed of 1/8 compared with the maximum threshold value. If the transmission coefficient for excitation at a speed of 1/8 less than the threshold value, then control passes to step 818, which continues the decoding speed of 1/8. If the transmission coefficient for excitation at a speed of 1/8 is greater than a threshold that indicates an error rate, the control proceeds to step 814.

At step 814 can be processed, providing abrasion to the DRA, and/or treatment, which restores the original state of the memory. Control passes to step 816, which continues decoding.

At stage 820 is checked preduprediti frame to determine whether this frame by frame with the speed of 1/8. If preduprediti frame is not a frame with a speed of 1/8, which indicates that sanctioned the transition from one speed to another, the control proceeds to step 826. If preduprediti frame is a frame with a speed of 1/8, which indicates that an unauthorized switch from one speed to another, the control proceeds to step 822.

At step 822 can be processed, providing the erased frame, and/or treatment, which restores the original state of the memory. Control proceeds to step 824, which continues decoding.

At step 826 checks the previous frame to determine whether this frame by frame type 1 at half speed or frame type 1 with full speed. If the previous frame is not a frame type 1 at half speed or frame type 1 with full speed, which indicates that sanctioned the transition from one speed to another, the control proceeds to step 832. If the previous frame is a frame of type 1 at half speed or frame type 1 with full speed, that is it shows on unauthorized switch from one speed to another, then control passes to step 828.

At step 828 may be processed, providing the erased frame, and/or treatment, which restores the original state of the memory. Control passes to step 830, which continues decoding.

At step 832, the transfer coefficient for excitation at a speed of 1/8 compared with the maximum threshold value. If the transmission coefficient for excitation at a speed of 1/8 less than the threshold value, then control passes to step 838, which continues the decoding speed of 1/8. If the transmission coefficient for excitation at a speed of 1/8 is greater than a threshold that indicates an error rate, the control proceeds to step 834.

At step 834 can be processed, providing the erased frame, and/or treatment, which restores the original state of the memory. Control continues to step 836, which continues decoding.

In Fig. 9 shows the scattering curve showing the relationship between the normalized transmission coefficient FTC and the transfer coefficient for predicting PL. Circles under the solid curve was obtained for the case of clean speech, and the stars above the solid curve correspond to unacceptable squeaks caused by errors AOC. The solid curve represents the threshold curve, which is th separates the region of a good speech from areas of unacceptable squeaks. This threshold curve can easily be represented in a parametric form and enter in the bug detector speed.

Thus, there has been described a new and improved method and apparatus for error detection speed receivers with variable speed. Specialists in the art will understand that the various logical blocks, modules, circuits, and algorithm steps described in connection with open options here, can be implemented as electronic hardware, computer software, or combinations thereof. Here the various components, blocks, modules, circuits, and steps are described in General with respect to their functional purpose. How implemented these functions in hardware or using software depends on the particular application and design constraints imposed on the entire system. Specialists in the art it is obvious that in these circumstances the hardware and software are interchangeable, and it is clear how best to implement the functions described here for each specific application. For example, here are the various logical blocks, modules, circuits, and algorithm steps described in connection with the disclosures provided in the description of the options can be implemented or performed with: processor digital to the output signals (DSP); specialized integrated circuit (ASIC); gate arrays, programmable by the user during operation (FPGA) or other programmable logic device, discrete gate or transistor logic circuits; discrete hardware components, such as registers and the FIFO memory (FIFO); processor executing a set of software and hardware commands or a known programmable software module and a processor, or any combination thereof. The processor may be a microprocessor, but alternatively it may be any known processor, controller, microcontroller, or state machine. A software module may reside in random access memory (RAM), flash memory, permanent memory (ROM), registers, hard disk, removable disk, ROM, CD (CD-ROM), or any other media known in the art. Specialists in this field of technology, it is also clear that the data, commands, information, signals, bits, symbols and signal elements, which are mentioned throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.

The preceding description of predpochtitel the different options are for to enable specialists in the art to make or use the disclosed here. Specialists in the art of the obvious various modifications of these options, as well as the fact that certain here the basic principles can be applied to other variants without the use of inventive skills. Thus, it is not intended that the disclosed variants are reduced to shown here are examples, and must meet the widest scope consistent with the disclosed here, the principles and new features.

1. Method of detecting errors when determining the coding rate of the data in the receiver, and the encoding speed is a variable that contains the stages, which are encoded speech signal, determine the speed of the encoding frame using the algorithm for determining the speed and verify the correctness of the above mentioned definitions of the coding rate, frame by using the information about the coding rate, related to the previous frame.

2. The method according to claim 1, characterized in that when detecting errors in the provided rate to create a set of unauthorized transitions from one speed to another for successive frames of speech on the basis of information on the classification of speech and phonetic features of speech and define p is the phenomenon of unauthorized transitions from one speed to another.

3. The method according to claim 2, characterized in that the set of unauthorized transitions from one speed to another includes a frame with full speed, followed by a frame with a speed of 1/8.

4. The method according to claim 2, characterized in that the set of unauthorized transitions from one speed to another includes a frame with full speed, type 1, followed by a frame with a speed of 1/8.

5. The method according to claim 2, characterized in that the set of unauthorized transitions from one speed to another includes a frame with half speed type 1, followed by a frame with a speed of 1/8.

6. The method according to claim 2, characterized in that the set of unauthorized transitions from one speed to another includes a frame with a quarter speed, followed by a frame with full speed type 1.

7. The method according to claim 2, characterized in that the set of unauthorized transitions from one speed to another includes a frame with a quarter speed, followed by a frame with half speed type 1.

8. The method according to claim 2, characterized in that the set of unauthorized transitions from one speed to another includes a frame with a speed of 1/8, followed by a frame with full speed type 1.

9. The method according to claim 2, characterized in that the set of unauthorized transitions from one speed to another includes a frame with a speed of 1/8, C is followed by a frame with half speed type 1.

10. The method according to claim 2, characterized in that the set of unauthorized transitions from one speed to another includes a frame with a speed of 1/8, followed by a frame with quarter speed, followed by the frame with the speed of 1/8.

11. The method according to claim 2, characterized in that the set of unauthorized transitions from one speed to another includes a frame with a speed of 1/8, followed by a frame with half speed, followed by the frame with the speed of 1/8.

12. The method according to claim 2, characterized in that the set of unauthorized transitions from one speed to another includes a frame with a speed of 1/8, followed by a frame with full speed, followed by the frame with the speed of 1/8.

13. The method according to claim 1, characterized in that when detecting errors in the provided rate encode at least one redundant bit of a frame of speech with assignment of the setpoint and confirm compliance with the specified value of the received backup bits encoded backup bat.

14. The method according to claim 1, characterized in that when detecting errors in the provided rate encode unused ID filter type for the frame of speech with assignment of the setpoint and confirm compliance with the specified value unused type identifier encoded filter ID filter type.

15. the manual on 14 characterized in that the identifier of the type of filter consists of two bits.

16. The method according to item 15, wherein three of the four case of double-bit combinations identify three types of filter, and one case of double-bit combination for type ID filter is not used.

17. The method according to claim 1, characterized in that when detecting errors in the provided rate analyze the relationship between the transfer coefficients of the fixed code tables and the transfer coefficients of the linear prediction to generate the threshold curve and validate the received speech on the threshold curve.

18. The method according to claim 1, characterized in that it further carry out perceptual aliasing effects detected error rate by performing processing for the restoration of the original state of the memory.

19. The method according to p, characterized in that the treatment, which restores the original state of the memory, carry out overwriting memory values to zero.

20. The method according to p, characterized in that the treatment, which restores the original state of the memory, carry out overwriting memory values on initialization values.

21. The method according to p, characterized in that the treatment, which restores the original state of the pam is ti, perform memory quantitate vector moving average for gear ratio : fixed code tables.

22. The method according to p, characterized in that the processing for the restoration of the original state memory, perform in the memory of the excitement.

23. The method according to p, characterized in that the processing for the restoration of the original state memory, perform in memory synthesis for encoding the linear prediction method (CLP).

24. The method according to p, characterized in that the treatment, which restores the original state of the memory, perform the memory-synthesis post-filter.

25. The receiver system to detect errors when determining the coding rate of the data, and the encoding rate of the data is a variable containing means for receiving the encoded speech signal, means for determining the coding rate, frame by using the algorithm of definition of speed, and means for checking the correctness of the above-mentioned definition of the coding rate, frame by using the information about the coding rate, related to the previous frame.

26. System for error detection speed A.25, characterized in that the means for receiving the encoded speech signal is a mobile subscriber unit.

27. System for obnaruzheniya speed A.25, characterized in that the means for receiving the encoded speech signal is a base transceiver station.

28. System for error detection speed A.25, characterized in that the means for detecting errors in the provided rate includes means for creating an unauthorized set of transitions from one speed to another for successive frames of speech on the basis of information on the classification of speech and phonetic features of speech and the means for determining the occurrence of unauthorized transitions from one speed to another.

29. System for error detection speed p, characterized in that the means for creating an unauthorized set of transitions from one speed to another includes means for creating an unauthorized switch from one speed to another frame with full speed, followed by a frame with a speed of 1/8.

30. System for error detection speed p, characterized in that the means for creating an unauthorized set of transitions from one speed to another includes means for creating an unauthorized switch from one speed to another frame with full speed, type 1, followed by a frame with a speed of 1/8.

31. System for error detection speed p, from ecaudata fact, that means to create a set of unauthorized transitions from one speed to another includes means for creating an unauthorized switch from one speed to another frame at half speed type 1, followed by a frame with a speed of 1/8.

32. System for error detection speed p, characterized in that the means for creating an unauthorized set of transitions from one speed to another includes means for creating an unauthorized switch from one speed to another frame with quarter speed, followed by a frame with full speed type 1.

33. System for error detection speed p, characterized in that the means for creating an unauthorized set of transitions from one speed to another includes means for creating an unauthorized switch from one speed to another frame with quarter speed, followed by a frame with half speed type 1.

34. System for error detection speed p, characterized in that the means for creating an unauthorized set of transitions from one speed to another includes means for creating an unauthorized switch from one speed to another frame with the speed of 1/8, followed by a frame with full speed type 1.

35. Systemidle error detection speed p, characterized in that the means for creating an unauthorized set of transitions from one speed to another includes means for creating an unauthorized switch from one speed to another frame with the speed of 1/8, followed by a frame with half speed type 1.

36. System for error detection speed p, characterized in that the means for creating an unauthorized set of transitions from one speed to another includes means for creating an unauthorized switch from one speed to another frame with the speed of 1/8, followed by a frame with quarter speed, followed by the frame with the speed of 1/8.

37. System for error detection speed p, characterized in that the means for creating an unauthorized set of transitions from one speed to another includes means for creating an unauthorized switch from one speed to another frame with the speed of 1/8, followed by a frame with half speed, followed by the frame with the speed of 1/8.

38. System for error detection speed p, characterized in that the means for creating an unauthorized set of transitions from one speed to another includes means for creating an unauthorized switch from one speed to another for the frame rate is 1/8, followed by a frame with full speed, followed by the frame with the speed of 1/8.

39. System for error detection speed A.25, characterized in that the means for detecting errors in the provided rate includes means for encoding at least one spare bit speech frame by assigning a preset value and a means to verify compliance with the specified value of the received backup bits encoded backup bat.

40. System for error detection speed A.25, characterized in that the means for detecting errors in the provided rate includes means for encoding unused ID filter type for the frame of speech with assignment of the setpoint and the means to verify compliance with the specified value unused type identifier encoded filter ID filter type.

41. System for error detection speed by clause 37, wherein the means for detecting errors in the provided rate includes means for analysis of the relationship between the transfer coefficients of the fixed code tables and the transfer coefficients of the linear prediction to generate the threshold curve and the means to confirm the accuracy of a received speech on the threshold curve.

2. System for error detection speed A.25, characterized in that it further comprises means for perceptually smoothing effects of the detected error rate by performing processing for the restoration of the original state of the memory.

43. System for error detection speed A.25, characterized in that the means for processing, it restores the original state of the memory includes means for overwriting the memory values to zero.

44. System for error detection speed A.25, characterized in that the means for processing, it restores the original state of the memory includes means for overwriting memory values on initialization values.

45. System for error detection speed A.25, characterized in that the means for processing, it restores the original state of the memory includes means for performing processing for the restoration of the original state memory, in memory quantitate vector moving average for gear ratio : fixed code tables.

46. System for error detection speed A.25, characterized in that the means for processing, it restores the original state of the memory includes redtwo to perform processing, it restores the original state of the memory, in the memory of the excitement.

47. System for error detection speed A.25, characterized in that the means for processing, it restores the original state of the memory includes means for performing processing for the restoration of the original state memory, in memory synthesis for encoding the linear prediction method (CLP).

48. System for error detection speed A.25, characterized in that the means for processing, it restores the original state of the memory includes means for performing processing for the restoration of the original state memory, in memory synthesis post-filter.

49. System for error detection speed A.25, characterized in that the means for processing, it restores the original state of the memory includes means for overwriting memory values on initialization values.

50. The receiver system to detect errors when determining the coding rate of the data, and the encoding rate of the data is a variable containing the receiver of the encoded speech signal, the controller can determine the speed of the encoding frame using the algorithm for determining the speed and detector error rate for testing the government is a major referred to determine the speed of the coding frame by using the information about the coding rate, related to the previous frame.

51. System for error detection rate for item 50, wherein the detector error rate includes generator unauthorized transitions from one speed to another to create a set of unauthorized transitions from one speed to another for successive frames of speech on the basis of information classification of speech and phonetic features of speech and determine the occurrence of unauthorized transitions from one speed to another.

52. System for error detection rate for item 50, wherein the detector error rate includes an encoder for encoding at least one spare bit frame of speech with assignment of the setpoint and the error detector speed to confirm compliance with the specified value of the received backup bits encoded backup bat.

53. System for error detection rate for item 50, wherein the detector for detecting the error rate includes unused ID filter type for the frame of speech with assignment of the setpoint and the error detector speed to confirm compliance with the specified value unused type identifier encoded filter ID filter type.

54. System for error detection speed is according to item 50, characterized in that the detector error rate includes the detector error rate for detecting errors in the provided rate by analyzing the relationship between the transfer coefficients of the fixed code tables and the transfer coefficients of the linear prediction to generate the threshold curve and validate the received speech on the threshold curve.

55. System for error detection rate for item 50, wherein the detector error rate includes the detector error rate for perceptual aliasing effects detected error rate by performing processing for the restoration of the original state of the memory.