JP2005534984A - Voice communication unit and method for reducing errors in voice frames - Google Patents

Abstract

A voice communication unit (100) having a voice encoder (134) capable of representing an input voice signal, wherein the voice encoder (134) is a transmission path (281) for transmitting a number of voice frames to a voice decoder. The speech encoder (134) is characterized by a virtual transmission path (282) for transmitting one or more references to a number of speech frames transmitted over the transmission path (281), The plurality of references relate to alternative voice frames in a number of voice frames transmitted over the transmission path (281) that are used as replacement frames if a frame is received in error, the voice communication unit ( 100). The voice communication unit (100) provides at least one advantage that a more accurate replacement frame mechanism is provided, thus reducing the risk of hearing undesirable artifacts in the recovered voice frame.

Description

  The present invention relates to speech coding and methods for improving the performance of speech codecs in speech communication units. The present invention is applicable to error reduction in an audio codec, but is not limited to this.

  Many current voice communication systems, including Global Mobile Communication System (GSM), Cellular Telephone Standard System, and Terrestrial Radio (TETRA) System, for personal mobile radio users encode and encode voice patterns. An audio processing unit is used for decoding. In such a voice communication system, a voice encoder in the transmission unit converts an analog voice pattern into a digital format suitable for transmission. An audio decoder in the receiving unit converts the received digital audio signal into an audible analog audio pattern.

  Because the frequency spectrum for such wireless voice communication systems is a valuable resource, it is desirable to limit the channel bandwidth used by voice signals to maximize the number of users per frequency band. . Therefore, the main objective in using speech coding techniques is to reduce the capacity occupied by speech patterns as much as possible by using compression techniques without compromising performance.

  A further approach for voice and data communication systems is to provide substantially less protection for voice signals when compared to comparable data signals. This approach introduces significantly more errors in voice packets than data packets and leads to an increased risk of losing the entire voice packet.

In the audio decoder,
(I) There are too many bit errors in the received voice frame; or (ii) Internet Protocol (IP) based intra-network data packets (which may contain voice information) are lost ;
In such cases, error mitigation techniques are commonly used, for example, to improve the performance of voice communication units.

  "Bad frame" mitigation techniques are necessary to minimize the audible effects of erroneously received frames. “Received incorrectly” means herein received with an error or not received at all. Such a technique reproduces estimates of lost speech frames, and does not introduce either silence or noise into the decoded speech. Such techniques typically involve using the statistical stationary properties of speech. One frame in error is usually well estimated by replacing it with similar parameters including energy, pitch, spectrum and voicing from previous frames. However, the voice is not truly steady (eg, the voice comes out) and the plosive is a very short event. Thus, this simple “replacement” technique is often unnatural and therefore can lead to undesirable artifacts.

In an ideal world, it is preferable to interpolate data from both sides of the transmission interruption, ie take data before and after the bad frame sequence and interpolate between them. However, such an approach introduces undesirable delays and is unacceptable in a voice communication system.

  When several bad frames are received, the energy of the speech signal is often reduced to zero after a few frames. Since it is useful to change what is repeated based on whether the voice is voiced or not, a “voiced” parameter is often included. In principle, for voiced speech, it is desirable to simply repeat the periodic component. In contrast, for unvoiced speech, it is desirable not to be too periodic, but to produce similar energy with a similar audio spectrum.

  The inventors of the present invention have realized and correctly recognized the limitations of using such a simple “replacement” frame mechanism as a bad frame mitigation strategy. In particular, the inventors of the present invention have realized that replacement frames are really only suitable frames in rare cases. Moreover, if many frames are received in error, this can often occur in low quality wireless communication links, but such a replacement frame mechanism is even more unacceptable.

  Therefore, there is a need to provide an improved error mitigation technique that mitigates at least some of the aforementioned drawbacks when using such audio codecs.

In a first aspect of the present invention, a voice communication unit according to claim 1 is provided.
In a second aspect of the present invention, a voice communication unit according to claim 11 is provided.
According to a third aspect of the present invention, there is provided a method for performing bad frame error mitigation in a voice communication unit according to claim 13.

According to a fourth aspect of the present invention, there is provided a voice communication unit according to claim 14.
In a fifth aspect of the present invention, a wireless communication system according to claim 15 is provided.
Further aspects of the invention are defined in the dependent claims.

  In summary, it is an object of the present invention to provide a communication unit and method with a voice codec that alleviates at least some of the above-mentioned drawbacks associated with current bad frame error mitigation techniques. This assumes that a voice frame on the transmission path has been received in error, sends a voice frame on the transmission path, and is sent on a virtual transmission path to indicate an alternative replacement voice frame to be used by the voice decoder. This is mainly achieved by using a reference / pointer. By using an additional virtual transmission line that ideally has a different error statistic (eg, another FEC scheme), the reference / pointer will not receive the same error as the voice frame it references. In addition, buffering techniques are used at the encoder to select alternative voice frames from many previously transmitted voice frames. Many previously transmitted speech frames exhibit characteristics similar to the selected alternative speech frame referenced.

  Illustrative embodiments of the invention will now be described with reference to the drawings.

Referring now to FIG. 1, a block diagram of a wireless subscriber unit adapted to support the inventive concept of a preferred embodiment of the present invention, hereinafter referred to as mobile station (MS) 100, is shown. The MS 100 is preferably a duplex filter, antenna switch, or circulator 104 that provides isolation between the receiver and transmitter chain within the MS 100.
The antenna 102 is preferably coupled to the antenna 102.

  As is well known in the art, the receiver chain generally comprises a scanning receiver front-end circuit 106 (effectively providing reception, filtering, and intermediate or baseband frequency conversion). The scan front end circuit 106 is coupled in series with the signal processing function 108. The output from the signal processing function 108 is supplied to an appropriate output device 110 such as a speaker via the audio processing unit 130.

  The voice processing unit 130 has a voice encoding function 134 that encodes a user's voice into a format suitable for transmission on a transmission medium. The audio processing unit 130 also includes an audio decoding function 132 that decodes received audio into a format suitable for output via the output device (speaker) 110. The voice processing unit 130 operates in combination with the storage device 116 and the timer 118 via the controller 114. In particular, the operation of the audio processing unit 130 is adapted to support the inventive concept of the preferred embodiment of the present invention. In particular, the speech processing unit 130 is adapted to select a replacement speech frame from a number of previously transmitted speech frames. Next, the audio processing unit 130, ie, the signal processor 108, starts transmitting a reference / pointer signal (indicating the selected replacement audio frame) on the alternative virtual transmission line for the main transmission line. This adaptation of the audio processing unit 130 will be described in more detail with reference to FIG.

  For completeness, the receiver chain also includes a received signal strength indication (RSSI) circuit 112 (shown coupled to the scanning receiver front end 106), but the RSSI circuit 112 is included in the receiver chain. You may arrange | position in any other place. The RSSI circuit is coupled to the controller 114 to manage control of the entire subscriber unit. Controller 114 is also coupled to scanning receiver front-end circuit 106 and signal processing function 108 (generally implemented by a DSP). Accordingly, the controller 114 may receive bit error rate (BER) or frame error rate (FER) data from the recovered information. Controller 114 is coupled to memory device 116 to store modes of operation such as decoding / encoding. A timer 118 is typically coupled to the controller 114 to control the timing of operations within the MS 100 (sending or receiving time-dependent signals). In connection with the present invention, the timer 118 indicates the timing of the audio signal in the transmission (encoding) path and / or the reception (decoding) path.

  With respect to the transmit chain, the transmit chain substantially comprises an input device 120, including a microphone converter, coupled in series to a transmitter / modulation circuit 122 via a speech encoder 134. Thereafter, any transmitted signal is radiated from the antenna 102 through the power amplifier 124. Transmitter / modulation circuit 122 and power amplifier 124 operate in response to the controller, and the output from power amplifier 124 is coupled to duplex filter or circulator 104. The transmitter / modulation circuit 122 and the scanning receiver front-end circuit 106 have a frequency up-conversion function and a frequency down-conversion function (not shown).

  Of course, the various components within the MS 100 can be placed in any suitable functional topology that makes the inventive concept of the present invention available. Further, the various components within MS 100 may be implemented in discrete component formats or in integrated component formats, and the final structure is merely optional.

It is the present invention that preferred buffering or processing of audio signals can be performed in software, firmware, or hardware, preferably using a software processor (or digital signal processor (DSP)) that performs audio processing functions. Is within the expected range.

  With reference now to FIG. 2, a block diagram of a code-excited linear prediction (CELP) speech encoder 134 is shown in accordance with a preferred embodiment of the present invention. The acoustic input signal to be analyzed is directed to the voice coder 134 at the microphone 202. The input signal is then directed to the filter 204. Filter 204 will generally exhibit bandpass filter characteristics. However, if the audio bandwidth is already sufficient, the filter 204 may have a direct wire connection.

  The analog audio signal from filter 204 is then converted into a sequence of N pulse samples, and the amplitude of each pulse sample is then digital-analog (A / A) as is well known in the art. D) Represented by the digital code in the converter 208. The sampling rate is determined by the sample clock (SC). The sample clock (SC) is generated together with the frame clock (FC).

  Thereafter, the digital output of A / D 208, which can be represented as input speech vector s (n), is directed to coefficient analyzer 210. This input speech vector s (n) is obtained iteratively in individual frames, ie in blocks of time. The length of time is determined by the frame clock (FC), as is well known in the art.

For each block of speech, the coefficient analyzer 210 generates a set of linear predictive coding (LPC) parameters according to a preferred embodiment of the present invention. Generated speech coder parameters may include LPC parameters, long-term prediction (LTP) parameters, excitation gain factor (G ₂ ) (with best stochastic codebook excitation codeword I). Such speech coding parameters are directed to the multiplexer 250 and sent over the channel for use by the speech synthesizer at the decoder. Input speech vector s (n) is also directed to subtractor 230. The function of the subtracter 230 will be described below.

  In the conventional CELP encoder of FIG. 2, the codebook search controller 240 uses the adaptive codebook in block 216 to generate a minimum weighting error for the selected excitation vector sum used to represent the input speech samples. The best index and gain are selected from the probabilistic codebook in block 214. The outputs from probabilistic codebook 214 and adaptive codebook 216 are input to respective gain functions 222 and 218. Thereafter, the gain-adjusted outputs are added by the adder 220 and input to the LPC filter 224 as is well known in the art.

Initially, the adaptive codebook or long-term predictor component is calculated as l (n). This is characterized by a delay and a gain factor “G ₁ ”.
For each individual stochastic codebook excitation vector u _i (n), a reconstructed speech vector s ′ _i (n) is generated for comparison with the input speech vector s (n). Gain block 22 calculates the excitation gain factor “G ₂ ” and summing block 220 adds an adaptive codebook component. Such a gain may be pre-calculated by the coefficient analyzer 210 and used to analyze all excitation vectors, or along with a search for the best excitation code name I generated by the codebook search controller 240. It may be optimized.

The calculated excitation G ₁ l (n) + G ₂ u _i (n) is then filtered by the linear predictive coding filter 224 that constitutes a short-term prediction (STP) filter, and the reconstructed speech vector s ′ _i (n) Is generated. The reconstructed speech signal for the i th excitation code vector is compared with the same block of the input speech vector s (n) by subtracting these two signals in the subtractor 230.

The difference vector e _i (n) represents the difference between the original block of speech and the reconstructed block. The difference vector is perceptually weighted by a weighting filter 232 that utilizes weighting and filter parameters (WTP) generated by the coefficient analyzer 210. Perceptual weighting emphasizes that frequency when the error is more important to the human ear and reduces other frequencies.

The energy calculator function in the codebook search controller 240 calculates the energy of the weighted difference vector e ′ _i (n). The codebook search controller compares the i th error signal of the current excitation vector u _i (n) with the previous error signal to determine the excitation vector that produces the smallest error. Thereafter, the code of the i th excitation vector with the smallest error is output to the channel as the best excitation code I.

The calculated excitation G ₁ l (n) + G ₂ u _I (n) is stored in 216 long-term predictor memory for future use.
In the alternative, the codebook search controller 240 may determine a specific code name that provides an error signal having a number of predefined criteria that meet a predetermined error threshold.

  A more detailed description of the functionality of a typical speech coding unit can be found in AM Kondoz's “Digital speech coding for low-bit communication systems” published by John Wiley in 1994. bit rate communications systems)) ”.

  In the preferred embodiment of the present invention, error mitigation techniques are applied to speech frames after multiplexer 250. The present invention utilizes an alternative virtual transmission line 282, preferably parallel, that is used to send pointers to previously encoded speech frames sent from the encoder over the main transmission line 281.

  In the context of the present invention, the expression “virtual” is defined as the transmission path supplied from the encoder to the decoder in addition to the main transmission path supporting voice communication. The “virtual” transmission paths may be in the same bit stream, or in the same time frame or multiframe in a time division multiplex scheme, and may be on different communication paths such as VoIP systems, for example. It may be through. Ideally, by utilizing an additional virtual transmission line with different error statistics (eg, another FEC scheme), the reference / pointer will not receive the same error as the voice frame it references.

  One notable difference to the known encoder is that there is a second minimized section after the multiplexing operation. Such a circuit evaluates the speech parameter data held in the buffer and selects the one closest to the current speech frame.

  In an enhanced embodiment, parallel virtual transmission lines use forward error correction (FEC) protection that is different from that used by the voice coder on the main transmission line. Thus, by using independent FEC paths, voice data packets are subject to different error statistics. This difference between the main transmission path and the parallel virtual transmission path helps to improve robustness against errors.

Multiplexer 250 outputs the data packet / frame to buffer 260, which holds the previous multiplexed frame. Demultiplexer 270 accesses the buffer frame of the multiplexed signal held in buffer 260. In this regard, demultiplexer 270 separates excitation parameter 274 from LPC parameter 272. Note that the long-term predictor memory used to generate the excitation parameters must be the same as the first long-term predictor 216 of the frame.

Thus, for each block of multiplexed speech, a set of linear predictive coding (LPC) parameters for the current and previous frames is generated. In the preferred embodiment of the present invention, each set of quantized LPC and excitation parameters forms a reconstructed speech vector s ′ _j (n) for the j th previous frame of buffer data. The reconstructed speech vector s ′ _j (n) is compared with the previously buffered speech vector s (n) by subtracting these two signals at the subtractor 262.

The difference vector e _j (n) represents the difference between the original block of speech and the previously buffered block. The difference vector is perceptually weighted by the LPC weighting filter 264. As already indicated, perceptual weighting emphasizes that frequency when the error is more important to the human ear and reduces other frequencies.

The energy calculator function in the codebook search controller 266 calculates the energy of the weighted difference vector e ′ _j (n). The codebook search controller 266 compares the jth error signal of the current excitation vector u _j (n) with the previous error signal to determine the excitation vector that produces the smallest error. The codebook search controller 266 then selects the “best index for frame data” to provide the least weighted error. The encoder then transmits to the decoder a pointer to the previous frame determined to provide the least weighting error between itself and each speech frame in the main transmission path.

  In effect, the reference speech frame (ideally with a different time and frame number from the current transmission frame) is most similar to the frame encoded by the encoder (in the sense of perceptually weighted errors). Compose frames within a particular moving window of audio. Thus, if a speech frame is received in error, the reference speech frame represents the best match (pointer) to the current frame used for error mitigation procedures. This representation or pointer will be described in more detail in FIG.

  Referring now to FIG. 3, a buffer timing diagram 300 illustrating the preferred process of the present invention is shown. The timing diagram shows frame 0 310 as received at the audio decoder and determined to be incorrect. The decoder then accesses an alternative virtual transmission path to determine the most appropriate frame to replace frame 0 310. As shown in FIG. 3, an alternative virtual transmission path includes a pointer to frame-4 320 as the preferred replacement for frame 0 310. Replacing frame 0 310 with frame-4 320 has minimal impact on speech quality in the speech decoding process.

  The inventors of the present invention recognize and utilize the fact that the immediately preceding preceding frame is all (generally) spoken by the same speaker, i.e., the speech frame exhibits a similar pitch and formant position. . Therefore, it is highly possible that a previous speech frame similar to the current speech frame can be found.

In accordance with a preferred embodiment of the present invention, given a set of parameters for each frame in memory, the minimum weight is determined by evaluating the segmental signal to noise (SEGSNR) or average weighted SNR for each of the buffer frames. Perceptual errors are found. Preferably, the segment is defined at the voice codec subframe level.

  This determination is made at the encoder. It is envisioned that if there is a small pitch error, significantly different SEGSNR values can occur. This is because the source audio and buffer signal can move quickly and out of phase. Thus, in an enhanced embodiment of the invention, a resolution smaller than the sample (usually 1/3 or 1/4 sample) is used to find the pitch period of the buffer frame and its surroundings (eg +/- 5%). However, it has been proposed to employ the highest SEGSNR value.

  In a further enhancement of the present invention, the frame used to mitigate bad reception of the frame itself is now erroneously received, as shown in FIG. 4, even if the frame itself is received incorrectly. Would be the best source of audio information for frames. Accordingly, FIG. 4 shows a timing diagram showing how multiple errors are handled. The data from frame 0 410 is known to be incorrect. The proposed error mitigation process uses an alternative virtual transmission line that shows data frame-4 420 as a suitable replacement. However, data frame-4 420 is determined to be incorrect. In that case, the pointer indicates data from frame-6 430 as the frame that is most similar to corrupted frame-4 420. Thus, frame-6 450 is used to replace frame-4 420 and is suitable to replace frame-1 410. In this way, multiple frame errors can be handled to overcome the problem of referencing out of memory.

  As a result, the reference (pointer) can effectively fall out of the storage window in the end. However, if the wrong value in the window is updated by eliminating the need for multiple references, this need not be a problem.

  Instead, when a replacement frame is stored in the buffer, if frame-4 420 is the current frame, it would have been replaced with frame-6430 (and then frame-2) in the buffer. As a result, the buffer always contains only usable data.

  In summary, the reference or pointer is sent to the decoder in an alternative bit stream for the main bit stream. The reference or pointer indicates the previously transmitted frame that best matches the currently transmitted frame. References or pointers are preferably transmitted in parallel bit streams. This reference or pointer is used in the frame replacement error mitigation process if the frame is received in error by the audio decoder. Thus, frame mitigation is enhanced by extending the known immediately or immediately following frame replacement mechanism from many frames to any frame. In this regard, the number of frames used in the process is limited only by the buffering / storage mechanism and / or processing power required to determine the minimum weighted error frame.

As shown, the speech coder speech parameter buffering / storing process is performed for many frames. For example, for a GSM Enhanced Full Rate (EFR) codec that is less than 12 kb / s, the amount of storage for 3 seconds of speech is only 5 kilobytes. Thus, the most difficult task is to identify the closest frame match from the 150 possible frames. Thus, in one embodiment of the present invention, the minimum weight error selection technique described above may be applied to parameters derived from a subset of parameters or synthesized speech, rather than all parameters of a speech coder frame. In other words, for storage in memory and for comparison processing, not the exact coder parameters, but the LPC filter parameters (LSF) and the energy of the synthesized speech frame (speech derived from synthesized speech calculated by both the encoder and decoder) Parameter) will be referenced (or pointed).

In this regard, speech frames contain many parameters, so the proposed technique can be applied in principle to any number of them. Examples of such parameters include the following for CELP coders:
(I) a line spectrum pair (LSP) representing LPC parameters;
(Ii) Long-term predictor (LTP) delay for subframe-1;
(Iii) LTP gain for subframe-1;
(Iv) Codebook index for subframe-1;
(V) codebook gain for subframe-1;
(Vi) long-term predictor delay for subframe-2;
(Vii) LTP gain for subframe-2;
(Viii) codebook index for subframe-2;
(Ix) codebook gain for subframe-2;
(X) long-term predictor delay for subframe-3;
(Xi) LTP gain for subframe-3;
(Xii) codebook index for subframe-3;
(Xiii) codebook gain for subframe-3;
(Xiv) long-term predictor delay for subframe-4;
(Xv) LTP gain for subframe-4;
(Xvi) codebook index for subframe-4;
Or (xvii) codebook gain for subframe-4.

  It is within the assumption of the invention that a pointer can be sent that refers to the set of LSPs of the previous frame that matches the LSP of the current frame, rather than the entire set of parameters. Alternatively, it would be possible to have a pointer to each of the many parameters described above.

  In a wireless communication system, parallel virtual transmission lines are preferably block-coded reference words (7 bits equal to about 2.5 seconds to support 128 frame buffers) within the unprotected bits of the data payload. Will be sufficient). The reference word may be encoded with a 15-bit BCH block code (at an equivalent rate of 75 bits / second), providing up to 2 bits of error correction.

  Instead, it is contemplated that an alternative virtual transmission line may provide a combination of error correction and error detection functions. Error detection may be useful because lack of receipt of references can lead to bad mitigation. If the reference word is received badly, the scheme can default to the previous frame repetition. The channel speed of 75 bits / second only slightly reduces the net bit rate of the GSM full speed channel from 22.8 kbps to 22.725 kbps. This is a negligible loss of sensitivity.

In another embodiment, such as voice over over an Internet Protocol (VoIP) communication link, an alternative virtual transmission path may be achieved by sending multiple packet streams. In this regard, it is desirable that the overall traffic does not increase substantially because the probability of dropping packets increases.

  The preferred mechanism would be to send a reference to the previous frame as described above only when a transition occurs and the speech is non-stationary. If the speech is stationary and if the prior art is relatively successful, no reference is sent. In this way, the packet network is not overloaded and most performance gains are achieved. The degree to which the audio signal is steady can be generated as a variable, which can be adjusted to improve the playback quality in the case of lost packets.

Since the decoder function is substantially the inverse of the encoder (without additional circuitry following the multiplexer), it will not be described in detail here. A description of the functionality of a normal speech decoding unit is also provided by AM Kondoz's “Digital speech coding for low-bit rate communications systems” published by Jo Win Wiley in 1994. ) ”. At the decoder, the decoder follows the standard decoding process until a bad frame is determined. When a bad frame is detected, the decoder evaluates the alternative virtual transmission path to determine the alternative frame indicated by each of the references / pointers. The decoder then searches for “similar” frames, as indicated by the reference / pointer transmission. The previously indicated frame is then used to replace the received frame to synthesize speech.

Advantageously, the inventive concepts described herein can be adapted to existing codecs by stealing bits from an already configured FEC scheme.
It is within the contemplation of the present invention that any audio processing circuit will benefit from the inventive concepts described herein.

It will be appreciated that the bad frame error mitigation mechanism has at least the following advantages, as described above.
(I) A more accurate replacement frame mechanism is provided, thus reducing the risk of undesired artifacts being heard in recovered speech frames.
(Ii) The alternative virtual transmission path can be adapted to an existing codec, for example by stealing bits from an already configured FEC scheme.
(Iii) The existing bad frame error mitigation techniques are used if a reference to the previous frame is sent only if the transition occurs and the speech is non-stationary, so the additional data required for the present invention Can be minimized.
(Iv) Misreceived parameters can be detected by cross-referencing data received in a given frame with frames referenced in this scheme.

  Although the preferred embodiment discusses applying the present invention to a CELP coder, the inventors have other benefits that can benefit from the inventive concepts contained herein where transmission errors can occur. Any audio processing unit is also envisaged. The inventive concepts described herein include, among other things, Universal Mobile Telecommunications System (UMTS) units, Global Mobile Telecommunications System (GSM), Terrestrial Based Radio (TETRA) communication units, Digital Interchange for Information and Signaling (Digital Interchange Applications for speech processing units for wireless communication units such as the Information of Signaling Standard (DIIS) and Voice over Internet Protocol (VoIP) units are particularly conceivable.

(Device of the present invention :)
The voice communication unit includes a voice encoder capable of representing an input voice signal. The speech encoder has a transmission path for transmitting many speech frames to the speech decoder. The speech encoder further comprises a virtual transmission path for transmitting one or more references for many speech frames transmitted over the transmission path. One or more references relate to alternative voice frames within the many voice frames transmitted over the transmission path that are used as replacement frames if the frame is received in error.
A voice communication unit, for example a voice communication unit as described above with a voice encoder, comprises a voice decoder adapted to receive a number of voice frames on the transmission line and one or more replacement voice frame references on the virtual transmission line. ing. The one or more references relate to alternative voice frames within the many voice frames received on the transmission path that are used as replacement frames if the frame is received in error.

(Method of the present invention :)
A method for performing bad frame error mitigation in a voice communication unit consists of sending a number of voice frames over a transmission path to a voice decoder by a voice encoder in the voice communication unit. A speech encoder transmits one or more references for many speech frames transmitted over a transmission path over a virtual transmission path. One or more references relate to alternative voice frames within the many voice frames transmitted over the transmission path that are used as replacement frames if the frame is received in error.

Thus, if a speech frame is received in error, an improved replacement frame can be selected from many speech frames.
Thus, a bad frame error mitigation technique and associated voice communication units and circuits have been described that substantially alleviate at least some of the above-mentioned drawbacks associated with at least known error mitigation techniques.

1 is a block diagram of a wireless communication unit with a voice coder adapted to support various inventive concepts of a preferred embodiment of the present invention. 1 is a block diagram of a code-excited linear prediction speech coder adapted to support various inventive concepts of a preferred embodiment of the present invention. Use of a reference mechanism indicated by an alternative virtual transmission line in which a replacement frame is selected from many other frames according to a preferred embodiment of the present invention. Enhanced use of alternative virtual transmission lines to address multiple errors occurring in the main transmission line, according to a preferred embodiment of the present invention.

Claims (15)

  A voice communication unit (100) having a voice encoder (134) capable of representing an input voice signal, wherein the voice encoder (134) is a transmission path (281) for transmitting a number of voice frames to a voice decoder. The speech encoder (134) has a virtual transmission path (282) for transmitting one or more references to a number of speech frames transmitted on the transmission path (281), The one or more references relate to alternative audio frames within the many audio frames transmitted over the transmission path (281) that are used as replacement frames if the frame is received in error. Communication unit (100). The speech encoder (134)
A multiplexer (250) for multiplexing the number of speech frames;
A buffer (260) that operates in combination with the multiplexer (250) and stores multiplexed audio data; and a current audio in the buffer (260) that operates in combination with the buffer (260). A processor (130, 270) for characterizing a frame and selecting an alternative voice frame that exhibits characteristics similar to the current voice frame, the reference to the alternative voice frame being a virtual transmission path (282) Sent to the decoder at the processor (130, 270);
The voice communication unit (100) of claim 1, further characterized by:   The processor has a demultiplexer function (270) for accessing one or more audio frames in the buffer (260), and the LPC parameters of the audio frames buffered to select audio frames exhibiting similar characteristics The voice communication unit (100) of claim 2, wherein the excitation parameter (274) is separated from the (272).   The voice communication unit (100) according to any of claims 1 to 3, wherein the virtual transmission line (282) is included in the same bit stream of the transmission line (281).   The transmission line (281) uses a first forward error correction protection scheme, and the virtual transmission line (282) uses a second forward error correction protection different from that used in the transmission line (281). The voice communication unit (100) according to any one of claims 1 to 4.   The voice communication unit (100) according to any of claims 2 to 5, wherein the processor (130, 266, 270) selects an alternative replacement frame to provide a minimum weighting error.   The processor (130, 266, 270) determines a minimum weighting error by evaluating a weighted segmental signal to noise (SEGSNR) or average weighted SNR for each of a plurality of buffer frames. Item 7. The voice communication unit (100) according to item 6.   The speech communication unit (100) according to claim 6 or 7, wherein the processor (130, 266, 270) determines a minimum weighting error of a subset of speech coding parameters.   The processor (130, 266) substantially searches the pitch period of the buffered speech frame and its surroundings and selects the frame showing the highest SEGSNR value. The voice communication unit (100) according to claim 1. d   The voice communication unit (100) according to any of claims 1 to 9, wherein the alternative voice frame (320) is referenced to the current voice frame only when a transition occurs and the voice is non-stationary.   Features an audio decoder (132) adapted to receive a number of audio frames on the transmission line (281) and one or more replacement audio frames (320) references on the virtual transmission line (282). The one or more references relate to alternative voice frames (320) within the many voice frames received on the transmission path (281) that are used as replacement frames if the frames are received in error. The voice communication unit (100) according to any of claims 1 to 10, wherein:   If the substitute audio frame (420) is received in error, a frame (430) is selected as the substitute frame for the incorrectly received substitute frame (420) and the current audio frame (410) received in error and The voice communication unit (100) of claim 11, wherein the voice communication unit (100) is used in place of a falsely received alternative voice frame (420). A method for performing bad frame error mitigation in a voice communication unit (100) comprising:
Sending a number of audio frames over the transmission path (281) to the audio decoder by the audio encoder (134) in the audio communication unit (100);
The method comprising:
Transmitting one or more references for a number of voice frames transmitted over a transmission path (281) on a virtual transmission path (282), wherein the one or more references are received when the frame is erroneously received. A step of relating to an alternative voice frame within the number of voice frames transmitted over the transmission path (281), which is used as a replacement frame in the case of
A method characterized by comprising.   Voice communication unit (100) adapted to carry out the two steps of the method according to claim 13.   A wireless communication system adapted to support the use of a transmission line (281) and a virtual transmission line (282) according to claims 1-14.

Images