JP4510742B2 - Voice packet receiving and reproducing method and apparatus, and program recording medium therefor - Google Patents

Description

In the present invention, when a digital audio signal, an audio signal such as a music signal (in this application document, the audio signal is generically) is transmitted through a packet communication network such as the Internet, the receiving side The present invention relates to an audio packet reception and reproduction method and apparatus used for reproducing an audio signal with stable quality, and a program recording medium thereof.

In recent years, there has been a demand for a technology for transmitting and receiving a higher quality audio signal by using a Voice over IP (VoIP) technology.
FIG. 8 shows a configuration example in which voice signals are converted into voice packets and communicated in real time through a packet communication network such as an IP communication network. The input audio signal input to the audio signal transmitting unit 70 is divided by a frame dividing unit 71 at regular intervals called frames, and the audio encoding unit 72 encodes the audio signal. Here, the frame length is generally 10 to 20 milliseconds in many cases. The encoded audio signal is converted into an audio packet by the audio packet converter 73. In this way, the input voice is converted into voice packets in the voice signal transmission unit 70 and transmitted to the IP communication network 74.

The voice packet includes a voice code obtained by converting a voice signal divided into frames by a voice coding method, and a time stamp or a frame number indicating the time order of the packet. As a typical speech coding method, G.I. which is a standard of ITU-T (International Telecommunication Union). There are 711 methods, but any other method may be used. Since the time stamp and the frame number can be converted to each other, the time stamp or the frame number is hereinafter simply referred to as a frame number.
Voice packets from the IP communication network 74 are received by the voice signal receiver 75, and the voice packets received by the voice signal receiver 75 are converted into voice signals and output.

FIG. 9 shows a specific configuration example of the audio signal receiving unit 75. Voice packets from the IP communication network are received by the packet receiver 80 and further stored in the reception buffer 81. The voice packet decoding unit 82 takes out the voice packet from the reception buffer 81, decodes it into a voice signal, and outputs the voice signal to the output terminal 88 through the changeover switch 89.
Here, the reception buffer 81 is also called a fluctuation absorbing buffer, and absorbs fluctuation when the arrival time of the packet is fluctuated depending on the state of the IP communication network, and prevents the reproduced audio signal from being interrupted. Also, when packet loss (packet loss) occurs in the communication network or when fluctuation exceeding the fluctuation absorption amount of the reception buffer 81 occurs and the reception buffer 81 is temporarily emptied, the voice packet decoding is performed. The unit 82 cannot extract the voice packet to be decoded from the reception buffer 81. In these cases, the control unit 84 operates the loss compensation processing unit (packet loss concealment unit) 83, and the loss compensation processing unit 83 generates a compensation audio signal and outputs it to the output terminal 88 through the changeover switch 89. Note that the loss compensation processing unit 83 is a method of generating an audio signal from which the deterioration of the reproduced audio is not noticeable from the already received audio signal when packet loss occurs. G. has been established as a standard of T (International Telecommunication Union). 711 Appendix I (Gee Dot 711 Appendix One) or the method shown in Non-Patent Document 1 is known.

The changeover switch 89 is switched by the control unit 84 depending on whether the audio signal is output from the audio packet decoding unit 82 or the loss compensation processing unit 83.
FIG. 10 shows an audio signal receiving unit 91, an audio signal transmitting unit 92, a speaker 93 for reproducing an audio signal from the audio signal receiving unit 91, a microphone 94 for receiving a transmission signal, and a microphone 94. A configuration example of an apparatus including an echo canceller 95 that suppresses an echo signal from a sound signal that has been sounded and transmits the signal to the sound signal transmission unit 92 will be described. The speaker 93 and the microphone 94 are a general acoustic speaker and a general microphone, respectively, which are equipped with both the audio signal receiving unit 91 and the audio signal transmitting unit 92, and are used for two-way communication like telephone communication. It is a structural example. In addition, the received voice signal (received signal) x (k) from the voice signal receiving unit 91 is reproduced and emitted from the speaker 93, and the reproduced voice is also collected by the microphone 94 through the echo path 96 and transmitted as an acoustic echo. Is done. Therefore, the audio signal x (k) supplied to the speaker 93 is also input to the echo canceller 95 and is passed through the pseudo echo path in the echo canceller 95 to generate a pseudo acoustic echo, which is collected by the microphone 94. A speech signal which is subtracted from the signal y (k) in the echo canceller 95 and the echo signal is suppressed is input to the audio signal transmission unit 92.

The pseudo echo path in the echo canceller 95 is adaptively controlled so as to approach the characteristics of the true echo path. In this way, the echo canceller 95 is operating normally, with the received signal x (k) and the collected sound signal y (k) being input so as to be correctly synchronized. If a state in which the reception signal x (k) is not input occurs or the synchronization is lost, it does not operate normally. That is, the echo is not correctly suppressed.
The voice signal transmission unit 92 includes a transmission method in which a voice packet is not transmitted at all when an input voice signal is silent, that is, a call is connected but a transmitter speaker is not speaking. This method is called silence compression or Discontinuous Transmission (hereinafter referred to as DTX), and is used as a method of efficiently using the bandwidth of the packet communication network.
Omuro Naka, et al. “Improvement of burst packet loss tolerance by parallel transmission of voice features”, IEICE Technical Report (Japan Society for Electronics and Information Engineers) Communication, SP2004-77, pp. 35-40, 2004

There are two problems to be solved by the present invention.
First, when the fluctuation of the packet becomes larger than the fluctuation absorption amount of the reception buffer as described above, the reception buffer is temporarily emptied. In order to prevent this phenomenon, it is possible to increase the amount of fluctuation absorbed in the reception buffer, that is, to increase the amount of packets stored in the reception buffer, but if the amount of packets stored in the reception buffer is increased, packets are received. After that, the delay until the audio signal is reproduced increases, and it becomes difficult to speak in the case of a two-way call. For this reason, it is common to tolerate a state in which the reception buffer is temporarily emptied when fluctuations fluctuate.

  When the reception buffer becomes empty, the loss compensation processing unit is operated as described above to perform packet loss compensation processing. However, if this process is applied continuously for many seconds, there is a problem that the state of the reception buffer becomes unstable when resuming packet arrival. In other words, in general, program control may be performed so that the amount stored in the reception buffer becomes a predetermined appropriate amount. In this case, as described above, packet loss compensation processing is continuously performed for many seconds. When it is in a state to perform, the operation for setting the accumulation amount to an appropriate amount does not work well, the delay becomes inappropriate, or the reproduced sound deteriorates due to the process of recovering it. In addition, for a long time, the sound reproduced by the packet loss compensation process has a poor sound quality.

  Second, in order to prevent the problem that the reception buffer becomes unstable, loss compensation processing is not performed, and the process of decoding voice packets is stopped until packets accumulate in the reception buffer. It is a problem that malfunction occurs in the operation. That is, in FIG. 10, when the reception signal x (k) input to the echo canceller 95 is stopped, only the collected sound signal y (k) is input to the echo canceller 95, and the echo canceller 95 operates correctly. However, in this case, a phenomenon occurs in which the transmission sound, that is, the reproduction sound of the other party is deteriorated.

  Whether or not there is a voice packet to be extracted from the reception buffer for each frame is determined in the first determination process, and if it is determined that there is no voice packet, whether to perform packet loss compensation processing or silence generation processing When it is determined in the second determination process and it is determined that there is no voice packet, a signal is generated by performing either the packet loss compensation process or the silence generation process based on the determination in the second determination process. That is, until N frames (N is an integer equal to or greater than 0) continue to be determined that there is no voice packet, it is determined that the packet loss compensation process is performed in the second determination process, and it is determined that there is no voice packet. When the state continues beyond N frames (N is an integer of 0 or more), it is determined that the silence generation process is performed in the second determination process.

If necessary, observe a series of received voice packets, estimate whether the received voice packets use silence compression in the DTX detection process, and if not use silence compression, If it is estimated, N ≧ 1, and if it is estimated that silence compression is used, N = 0.
If necessary, every time an audio packet is extracted from the reception buffer and decoded, the corresponding frame number is stored in the frame number counter, and each time the packet loss compensation process is performed for one frame, the frame number counter Every time the count value is incremented by 1 and the silence generation process is performed, the count value of the frame number counter is maintained as it is without being increased, and it is determined that there is an audio packet to be extracted in the first determination process. In the frame number determination process, it is determined whether or not the frame number of the packet is equal to or less than the frame number indicated by the frame number counter. If the determination result is the following, the received voice packet is discarded, and the determination result is If not, the extracted voice packet is decoded and a voice signal is output.

[Example 1]
FIG. 1 shows a functional configuration example of a voice packet reception / playback apparatus according to the best mode for carrying out the present invention. In this figure, parts corresponding to those in FIG. The packet receiving unit 80 receives a voice packet from the IP communication network and sends it to the reception buffer 81.
When the voice packet is stored in the reception buffer 81, the voice packet decoding unit 82 extracts one (one frame) having the smallest frame number from the voice packets in the reception buffer 81 and converts it into the voice packet. At the same time as decoding the included speech code, the frame number of the extracted packet is set in the frame number counter 85 as a count value. The decoded audio signal is sent to the output terminal 88 and output.

  When the reception buffer 81 is empty and the voice packet decoding unit 82 tries to extract the voice packet from the reception buffer 81 and the extracted voice packet is not accumulated, the control unit 84 causes the loss compensation processing unit 83 to operate. The compensation audio signal is sent to the output terminal 88 and output. Every time packet loss compensation processing is performed for one frame, the count value of the frame number counter 85 is incremented by “1” (incremented by “1”). When the packet loss compensation processing is completed, it is checked again whether or not the voice packet can be taken out from the reception buffer 81. When there is no voice packet to be extracted and the loss compensation processing unit 83 is operated N times continuously, and thereafter there is no voice packet that can be extracted from the reception buffer 81, the silence generating unit 87 is operated to generate a silence signal. Then, it is sent to the output terminal 88 and outputted. N is an integer of 0 or more. Thereby, the first problem can be solved. The signal supplied to the output terminal 88 is also supplied to the echo canceller. Therefore, since the signal is always input to the echo canceller also from the reception signal side, the echo canceller can be operated correctly, thereby solving the second problem.

When the frame length is 20 milliseconds, the value of N is preferably 2 to 5 frames, and when the frame length is 10 milliseconds, the value of N is preferably 4 to 10 frames. It has been experimentally confirmed that the number of frames corresponds to 40 to 100 milliseconds, and preferably the number of frames corresponding to 40 to 60 milliseconds. Note that if N is too small, it becomes a silent signal immediately, which is not preferable as voice, and conversely, if N is too large, voice quality deteriorates.
[Example 2]
An embodiment of the inventive device for the case of using DTX on the transmission side will be described.
In this case, a DTX detector 86 indicated by a one-dot chain line in FIG. 1 is provided. The DTX detection unit 86 monitors the packet reception state in the packet reception unit 80, and continuously estimates whether or not the transmission side transmits a voice packet using DTX. Whether or not DTX is used can be estimated, for example, by referring to the header information of the packet.

When the transmission side is transmitting voice packets using DTX, the reception buffer 81 on the reception side is emptied, and when the voice packet decoding unit 82 cannot extract the voice packets, the fluctuation is larger than the expected fluctuation. It is not possible to determine at this time whether it is a case of using DTX or a case of using DTX. If the phenomenon that the reception buffer 81 is emptied due to the use of DTX is regarded as packet loss (incorrect determination) and loss compensation processing is performed, reproduced sound deteriorates.
Therefore, when the reception buffer 81 becomes empty, if the transmission side does not use DTX, as described above, the packet loss compensation processing and, if necessary, silence signal generation are performed, and the reception packet uses DTX. If so, a silence signal is immediately generated.
Specific Example of Control Unit Next, a configuration example of specific functions of the control unit 84 in FIG. 1 is shown in FIG. A signal indicating whether or not there is a voice packet that the voice packet decoding unit 82 can extract from the reception buffer 81 is generated by the packet extraction enable / disable signal generation unit 20 from the voice packet accumulation state in the reception buffer 81. The packet extraction enable / disable signal generation unit 20 outputs “1” in FIG. 2 if there is an audio packet that can be extracted, and outputs “0” if there is no audio packet that can be extracted. The packet extraction enable / disable signal is sent to the voice packet decoding instruction unit 22, the loss compensation instruction unit 24, and the silence generation instruction unit 26, respectively. If the packet extraction enable / disable signal is “1”, the voice packet decoding instruction unit 22 instructs the voice packet decoding unit 82 to perform decoding processing.

A detection signal from the DTX detection unit 86 is sent to the DTX flag register 28. In this example, the DTX flag in the DTX flag register 28 is set to “1” when DTX is used, and the DTX flag is set to “0” when DTX is not used. The DTX flag signal is sent to the loss compensation instruction unit 24 and the silence generation instruction unit 26.
Since the packet loss compensation processing can be performed continuously up to N frames (N is an integer of 0 or more), a compensation continuation enable / disable signal generation unit 34 is provided. The further compensation continuation enable / disable signal is sent to the loss compensation instruction unit 24 and the silence generation instruction unit 26.

For example, as shown in FIG. 3, the compensation continuation enable / disable signal generator 34 is incremented by 1 every time the n counter 34e reset to zero instructs to perform the packet loss compensation process for one frame. The count value n of 34e is compared with the value N of the register 34c by the comparison unit 34d. If the count value n is less than N according to the comparison result, the compensation continuation enable / disable signal is set to “1”, and it is greater than or equal to N. In this case, the compensation continuation enable / disable signal is output as “0”. The n counter is reset to 0 by a decoding instruction signal from the voice packet decoding instruction unit 22.
Alternatively, the loss compensation processing unit 83 stores the count value F ₁ of the frame number counter 85 at the start of the packet loss compensation process, and every time the packet loss compensation process is performed for one frame, the count value F _c of the frame number counter 85 F ₁ may be subtracted from and a determination may be made as to whether or not the subtraction result is less than N.

These configurations are also referred to as loss compensation processing counters because they count the number of frames of packet loss compensation processing that are continuously performed.
If the voice packet decoding unit 82 does not extract a voice packet from the reception buffer 81, the value of the DTX flag is “0”, and the number of consecutive loss compensation processes is less than N, that is, the compensation continuation permission signal is “ 1 ”, the loss compensation instruction unit 24 issues a loss compensation instruction and operates the loss compensation processing unit 83. That is, the loss compensation instruction unit 24 is configured by an AND circuit 38, for example, as shown in FIG. When the output of the AND circuit 38 becomes “1”, a loss compensation instruction is generated.

  Further, if the voice packet decoding unit 82 does not extract a voice packet from the reception buffer 81, the value of the DTX flag is 0, and the number of consecutive loss compensation processes is N or more, that is, the compensation continuation enable / disable signal is “ If it is “0” (condition 1), the silence generation instruction unit 26 issues a silence generation instruction to the silence generation unit 87 and activates the silence generation unit 87. Further, if the voice packet decoding unit 82 cannot extract the voice packet from the reception buffer 81 and the value of the DTX flag is 1, that is, if the transmission side does not transmit the voice packet in the silent period (condition 2), there is no sound. The generation instruction unit 26 issues a silence generation instruction to the silence generation unit 87 and activates the silence generation unit 87. That is, for example, as shown in the drawing, the silence generation instruction unit 26 detects the condition 1 with the AND circuit 40, the condition 2 with the AND circuit 42, and the output of either of the AND circuits 40 and 42 is “ If “1”, the silence generation instruction unit 26 outputs a silence generation instruction through the OR circuit 46.

When the decoding instruction is output from the voice packet decoding instruction unit 22, the control unit 84 switches the switch 89 in FIG. 1 to connect the output side of the voice packet decoding unit 82 to the output terminal 88, and the loss compensation instruction unit 24, when the loss compensation instruction is output, the switch 89 is switched, the output side of the loss compensation processing unit 83 is connected to the output terminal 88, and when the silence generation instruction unit 46 outputs the silence generation instruction, the switch 89 And the output side of the silence generator 87 is connected to the output terminal 88.
As described above, the frame number of the voice packet taken out by the voice packet decoding unit 82 from the reception buffer 81 is set as the count value in the frame number counter 85, and every time the packet loss compensation process is performed for one frame, The count value of the number counter 85 is increased by “1”, but when a silence signal is generated, the count value of the frame number counter 85 is not increased and the count value is left as it is.

  When receiving a voice packet, the packet receiver 80 temporarily stores it in the received packet register 80a as shown in FIG. 4, for example, and within the packet receiver 80, the frame number of the received voice packet and the frame number counter 85 The match determination unit 80b compares the count value, and if the frame number of the received voice packet is equal to or less than the count value of the frame number counter, the match determination unit 80b does not store the packet in the reception buffer 52 and the packet determination unit 80b 80c is activated and the received voice packet is discarded. If the frame number of the received voice packet is larger than the count value of the frame number counter, the matching determination unit 80b opens the gate 80d and stores the received voice packet in the reception buffer 81 through the gate 80d.

  Such a process of consistency between the received packet and the already received packet and the frame number may be performed not on the received voice packet but on the voice packet extracted from the reception buffer 81. In this case, as shown in FIG. 4, a functional configuration for this purpose is provided in the reception buffer 81 as shown in parentheses. The voice packet taken out from the reception buffer 81 is stored in the taken-out packet register 81a, and is input to the voice packet matching determination unit 81b, where consistency with the count value of the frame number counter 85 is determined. In this case, either the packet discarding unit 81c or the sent voice packet is sent to the voice packet decoding unit 82 through the gate 81d to be decoded.

Since this configuration example is provided with the DTX detection unit 86, the DTX flag register 28 and the AND circuit 42 and the OR circuit 46 in the silence generation instruction unit 26 are omitted when the DTX detection unit 86 is not provided. .
[Example 3]
With reference to FIG. 5 and FIG. 6, an embodiment of the method of the present invention in the case where the DTX detector 86 is provided will be described. Accumulation of voice packets in the reception buffer 81 and extraction from the reception buffer 81 operate asynchronously. FIG. 5 shows a flow of accumulation operation in the reception buffer. First, the DTX flag register 28 is reset to 0 (S201), and when the packet receiver 80 receives a voice packet (S203), the DTX detector 86 observes the packet header (S205). As a result, it is determined whether or not the received voice packet uses DTX (S207). If DTX is used, the DTX flag register 28 is set to 1 (S209), and if DTX is not used. The DTX flag register 28 is left as it is. The matching determination unit 80b refers to the count value of the frame number counter 85 (S211) and compares it with the frame number of the received voice packet (S213). If the frame is not already reproduced, that is, the frame of the received voice packet. If the number is larger, the packet is stored in the reception buffer 81 (S215). If the frame that has already been reproduced, that is, the frame number of the received voice packet is less than or equal to the count value of the frame number counter 85, the packet discarding unit 80c. Accordingly, the received voice packet is discarded (S217). Thereafter, it is checked whether or not the call is finished. If not finished, the process returns to the step of receiving the packet in step S203. If the call is finished, the storing process in the reception buffer is finished (S219).

  Next, the extraction of voice packets from the reception buffer 81 and the subsequent processing flow will be described with reference to FIG. The loss compensation processing counter in the compensation continuation permission / prohibition signal generation unit 34 is reset to 0 (S301), and the voice packet decoding unit 82 performs an operation of taking out a voice packet from the reception buffer 81 (S303). It is determined whether there is a voice packet that can be taken out in the reception buffer 81 (S305). If there is a voice packet that can be taken out, the voice packet with the smallest frame number is taken out (S307). The extracted voice packet is decoded by the voice packet decoding unit 82 (S309), and the frame number of the decoded voice packet is set as a count value in the frame number counter 85 (S311). When the decoded audio signal is output to the output terminal 88 and the audio signal transmission unit is also provided, the decoded audio signal is also sent to the echo canceller (S313). Next, it is determined whether or not the call is ended (S315). If the call is not ended, the process returns to the step of taking out the voice packet from the reception buffer 81 in step S303.

On the other hand, if there is no voice packet to be extracted from the reception buffer 81 in step S305, the value of the DTX flag is determined (S317). If the value of the DTX flag is “0”, that is, if DTX is not used. Then, it is determined whether the number of consecutive packet loss compensation processes is less than N (S319). If the number of consecutive packet loss compensation processes is less than N, the loss compensation processing unit 83 is activated to generate a loss compensation signal for one frame (S321), and the count value of the frame number counter 85 is set to “ 1 "is increased (S323), and the generated compensation signal is output to the output terminal 88 (S313). In step S317, if the value of the DTX flag is 1, that is, if DTX is used, the silence generator 87 is activated (S325), and a silence signal is output to the output terminal 88 (S313). If it is determined in step S319 that the number of consecutive packet loss compensation processes is N or more, the silence generator 87 is activated (S325), and a silence signal is output to the output terminal 88 (S313). If it is determined in step S315 that the call has ended, the voice packet described above is extracted, and the process associated therewith is terminated.
[Example 4]
Next, with reference to FIGS. 5 and 7, an embodiment will be described which is performed when the consistency of the frame number of the received voice packet is taken out from the reception buffer. In this case, the process of accumulating voice packets in the reception buffer 81 omits steps S211, S213, and S217 as shown by the broken line in FIG. 5 and immediately moves from step S209 to step S215.

  On the other hand, in the process of extracting the voice packet from the reception buffer 81, the step corresponding to FIG. 6 in FIG. 7 indicates the same number, and in a different part, the voice packet decoding unit 82 extracts the voice packet having the smallest frame number. After that, the count value of the frame number counter 85 is referred to (S601), the count value is compared with the frame number of the extracted voice packet, and if the latter is larger, it is determined that it has not been reproduced (S603). In step S309, the extracted voice packet is decoded. If the frame number of the voice packet extracted in step S603 is less than or equal to the count value of the frame number counter, it is determined that the voice packet has been reproduced, and the extracted voice packet is discarded (S605), and the process proceeds to step S315.

In the above description, the silence signal is not a complete silence signal, and may be a small amplitude signal that does not affect the receiver.
The apparatus shown in FIGS. 1 and 2 may be operated by a computer. In this case, a program for causing the computer to function as the apparatus shown in FIGS. 1 and 2 is mounted on a recording medium such as a CD-ROM, a magnetic disk, or the computer via a communication line. It can also be downloaded and realized.
With the configuration of the present invention, when a voice packet does not arrive from the network and the reception buffer is empty, loss compensation processing is performed to prevent deterioration of reproduced sound. However, if the loss compensation process is performed for many seconds, it will be easy to cause deterioration of the reproduced sound. Therefore, if the loss compensation process is performed for N frames set in advance, a silence signal is generated and output to the echo canceller. Since the silent signal is also input, the echo canceller can be operated correctly. Furthermore, by providing a DTX detector, detecting whether the phenomenon that the reception buffer is emptied is due to the use of the DTX function or due to fluctuations more than expected, and performing processing suitable for it, An effect that does not deteriorate the sound quality can be obtained.

The block diagram which shows the function structural example of Example 1 of this invention. The figure which shows the specific functional structural example of the control part 84 in FIG. The block diagram showing each characteristic of the specific structural example of the compensation continuation permission signal generation part 34 in FIG. The block diagram which shows the function structure which determines the consistency of the frame number of a received packet or an extraction packet, and the frame number counter 85. FIG. The flowchart figure which shows the example of the accumulation | storage process to the reception buffer 81 of Example 3 and Example 4 of this invention. The flowchart figure which shows the example of the extraction process from the reception buffer 81 of Example 3 of this invention. The flowchart figure which shows the example of the extraction process from the reception buffer 81 of Example 4 of this invention. 1 is a block diagram showing an example of a system configuration for converting a voice signal of a prior art into a voice packet and communicating via a packet communication network. The block diagram which shows the structural example of the audio | voice signal receiving part of a prior art. The block diagram which shows the structural example of the transmitter / receiver provided with the echo canceller of FIG. 8 of a prior art.

Claims (6)

In the audio packet receiving and reproducing method for accumulating received audio packets in a reception buffer, sequentially extracting audio packets from the reception buffer in order of frame number, decoding the extracted audio packets, and outputting an audio signal.
A first determination step for determining whether there is an audio packet to be extracted from the reception buffer for each frame;
A second determination process for determining whether to perform a packet loss compensation process or a silence generation process when it is determined in the first determination process that there is no voice packet;
When it is determined in the first determination process that there is no voice packet, a signal is generated by performing either the packet loss compensation process or the silence generation process based on the determination in the second determination process, and is output. Process ,
Each time a voice packet is extracted from the reception buffer and decoded, a corresponding frame number is stored in a frame number counter;
Each time the packet loss compensation process is performed for one frame, the count value of the frame number counter is incremented by 1,
Each time the silence generation process is performed, the process of maintaining the count value of the frame number counter without increasing it,
It is determined that there is a voice packet to be extracted in the first determination process, and a frame number determination process for determining whether or not the frame number of the extracted voice packet is equal to or less than the frame number indicated by the frame number counter;
If the determination result of the frame number determination process is the following, the received voice packet is discarded, and if the determination result is not the following, the extracted voice packet is decoded and an audio signal is output. Have
The second determination process, in the upper Symbol first determination process until a state where it is determined that there is no voice packet (the N 0 or an integer) N frames consecutive, it is determined to perform packet loss compensation processing , when the condition of judging not voice packets are continuously beyond the N frames, the voice packet receiving and reproducing method characterized in that a determination to perform a no-sound generation processing. The method of claim 1, wherein
Having a DTX detection process of observing a series of received voice packets and estimating whether the received voice packets utilize silence compression;
In the above DTX detection process, if it is estimated that silence compression is not used, N ≧ 1,
In the DTX detection process, if it is estimated that silence compression is used, N = 0 is set. In claim 2 Symbol placement methods,
In the DTX detection process, when it is estimated that silence compression is not used, the N is set to the number of frames corresponding to 40 to 100 milliseconds. A computer-readable recording medium storing a program for causing a computer to execute each step of the voice packet receiving and reproducing method according to any one of claims 1 to 3 . Accumulating received voice packets to the receive buffer, from the reception buffer, the frame number order, sequentially extracts a voice packet, the extracted voice packets decrypt, and outputs the audio signal to the output terminal, the frame number Is stored in the frame number counter,
A packet extraction enable / disable signal generator for generating a packet extraction enable / disable signal indicating whether or not there is a voice packet to be extracted from the reception buffer for each frame;
The packet extraction permission signal generation unit by lipase packet extraction availability signal is input, the packet extraction permission signal to exhibit not taken out, N frames (N is an integer of 0 or more) compensated continuation permission indicating whether continuous A compensation continuation enable / disable signal generator for generating a signal;
For each frame, the packet extraction enable / disable signal and the compensation continuation enable / disable signal are input. If the packet extraction enable / disable signal cannot be extracted, and the compensation continuation enable / disable signal can continue to be compensated, a packet loss compensation process is performed. A loss compensator that executes the frame, generates an audio signal, outputs the audio signal to the output terminal, and increases the count value of the frame number counter by 1;
For each frame, the packet extraction enable / disable signal and the compensation continuation enable / disable signal are input. If the packet extraction enable / disable signal cannot be extracted and the compensation continuation enable / disable signal cannot be compensated, a silence signal is generated. A silence generator that outputs to the output terminal but does not increase the count value of the frame number counter ;
A matching determination unit that compares the frame number of the received voice packet with the frame number indicated by the frame number counter and determines whether the frame number of the voice packet is equal to or less than the frame number indicated by the frame number counter;
If the determination result of the matching determination unit is as follows, a packet discard unit that discards the received voice packet;
If the determination result is not below, a gate for storing the received voice packet in the reception buffer;
When a packet takeout enable / disable signal is input from the packet takeout enable / disable signal generator, and the packet takeout enable / disable signal indicates that it can be taken out, the voice packet stored in the reception buffer is taken out and decoded, and the voice signal is sent to the output terminal. An output voice packet decoding unit;
An audio packet receiving / reproducing apparatus comprising: Apparatus according to claim 5 Symbol mounting,
A DTX detector that receives the received voice packet and detects whether the transmission side uses silence compression;
If the detection output of the DTX detection unit is input and the detection output indicates that silence compression is not used, the N is set to 1 or more, and if the detection output indicates that the NTX is used, the N is set to 0. voice packet receiving and reproducing apparatus, characterized in that it comprises a means for setting the.

Images