JP2006100882A - Voice packet reproducing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a voice packet reproducing system capable of reproducing voices while reducing a sense of incongruity being the influence provided to voices to be reproduced from a reduced voice packet. <P>SOLUTION: The division ratio of a voice reproduction timing signal is controlled in accordance with correction amount information corresponding to the difference in clock frequency between the voice reproduction timing signal to be a reference for reproducing a voice from the received voice packet and the reception timing of the received voice packet, and a reproduction request signal for reproducing the voice signal from the received voice packet is generated. The received voice packet is temporarily recorded in a voice buffer, the reading reproduction interval of the recorded voice packet is controlled by the reproduction request signal, and the reproduction timing of the voice packet to frame data is varied per reproduction clock unit, thereby reproducing without any change, discarding or inserting voice sample data per the reproduction clock unit to generate reproduced data. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a voice packet reproduction method in a system that performs voice communication via a packet on a network.

In this type of conventional voice packet reproduction method, voice packets received from the network are stored in a voice buffer for reproducing the voice at regular intervals based on the voice sample clock by absorbing variations in arrival time. In the audio buffer, audio reproduction is started when a certain number of packets are accumulated. When the voice packet reception interval becomes long or packet loss occurs and the packets in the buffer are exhausted, alternative packets are inserted. Further, when the voice packet reception interval becomes short and the buffer overflows, the packet is discarded (see Patent Document 1).
JP 2001-45065 A

  In such a conventional voice packet reproduction method, even in a good communication state in which there is no packet loss due to the clock frequency difference between the reception timing and the reproduction timing of the voice packet, the packet in the voice buffer is Accumulation or depletion has progressed, and voice buffer overflows or underflows occur periodically. Since voice data is discarded or inserted in units of voice packets each time, there is a problem that the sense of discomfort of the reproduced voice is large.

  An object of the present invention is to provide an audio packet reproduction method capable of performing audio reproduction while reducing a sense of incongruity that is an influence on audio reproduced from a received audio packet.

In order to solve this problem, an audio packet reproduction method according to the present invention includes a timing generation unit that outputs an audio reproduction timing signal serving as a reference for reproducing audio from a received audio packet;
A frequency difference detection unit that detects a clock frequency difference between the reception timing of the received audio packet and the audio reproduction timing signal and outputs correction amount information corresponding to the clock frequency difference;
A correction control unit that controls a frequency division ratio of the audio reproduction timing signal according to the correction amount information corresponding to the detected frequency difference and generates a reproduction request signal for reproducing an audio signal from the received audio packet;
The received voice packet is temporarily recorded, and the read / playback interval of the recorded voice packet is controlled by the playback request signal, so that the playback timing of the voice packet can be changed in units of playback clock. An audio buffer;
A voice decoding unit that converts the frame data of the voice packet into voice sample data under the control of the reproduction request signal;
An audio reproduction unit for generating reproduction data by reproducing, discarding, or inserting the audio sample data in units of the reproduction clock based on the reproduction request signal;
It has the composition provided with.

  According to the present invention, since the reproduction of audio data can be discarded and inserted in units of reproduction clocks, it is possible to expect an improvement in the quality of reproduced audio.

FIG. 1 is a block diagram showing Embodiment 1 of the present invention. FIG. 2 is a time chart for explaining the operation of the method of the present invention.
[LAN interface 1]
The LAN interface 1 transfers packet data received from the IP network to the packet reception control unit 2.
[Packet reception control unit 2]
As shown in FIG. 3, the packet reception control unit 2 includes a packet data reception unit 2-1, a reception buffer control unit 2-2, and a reception timing information generation unit 2-3.
The packet data receiving unit 2-1 extracts the reception timing from the voice packet received from the LAN interface 1, and outputs the sequence number, time stamp, and encoded voice frame data a.
The reception buffer control unit 2-2 transfers the encoded audio frame data a to the audio buffer 3.
The reception timing information generation unit 2-3 outputs a transfer completion timing signal indicating that the transfer to the audio buffer 3 has been completed to the frequency difference detection unit 5, and the sequence number recorded in the audio packet and Output timestamp.
[Audio buffer 3]
The audio buffer 3 stores the encoded audio frame data a transferred from the packet reception control unit 2.
The accumulated encoded audio frame data b is accumulated in a certain amount necessary for jitter absorption, and is then controlled by a reproduction request signal e from the correction control unit 6 and transferred to the audio decoding unit 7.

[Timing generator 4]
The timing generator 4 divides a reference clock from a sound reproduction side reference clock source (not shown) to generate a sound reproduction timing signal f.
[Frequency difference detector 5]
The frequency difference detector 5 detects the clock frequency difference between the transfer completion timing signal of the received voice packet and the reference reproduction timing signal f, and outputs correction amount information corresponding to this frequency difference. That is, the frequency difference detection unit 5 monitors the transition of the time given to the voice packet configured based on the voice coding method common between the transmission and reception points, and determines the received packet and the playback packet per certain time. By calculating the amount of change in time, the clock frequency difference between transmission and reception is calculated, and correction amount information corresponding to this is generated [correction control unit 6]
The correction control unit 6 divides the audio reproduction timing signal f from the timing generation unit 4 to generate a reproduction cycle as a reference for audio decoding and audio reproduction. Further, as shown in FIG. 5, the frequency division ratio control circuit 6-1 and the frequency division counter 6-2 are used to determine the reference reproduction cycle based on the correction amount information from the frequency difference detection unit 5. Thus, in order to perform timing correction control of data for one reproduction clock, a reproduction request signal e with a changed frequency division ratio is output. Further, the correction control unit 6 generates a reproduction request signal e for controlling the packet transfer interval from reception to reproduction, that is, the audio reproduction interval, based on the correction amount information corresponding to the clock frequency difference.

[Audio decoding unit 7]
The audio decoding unit 7 decodes the decoded audio data corresponding to the time series at the sampling rate actually used based on the encoded audio frame data b transferred from the audio buffer 3 in synchronization with the reproduction request signal e. Convert to c and output.
[Audio playback unit 8]
The audio reproduction unit 8 reproduces the decoded audio data c as it is corresponding to the change in the period of the reproduction request signal e, or discards or inserts the data for one reproduction clock and outputs the reproduction audio data d To do. That is, the audio reproduction unit 8 reproduces the audio sample data string as it is based on the correction amount information in conjunction with the cycle for extracting the packet from the audio buffer 3 or discards data for one reproduction clock. The inserted audio data d1, d2, and d3 are taken out.

An outline of the operation of the method of the present invention will be described.
In FIG. 1 showing an embodiment of the present invention, the frequency difference detection unit 5 monitors the transition of time given to a voice packet configured based on a voice coding method common between transmission and reception points, By calculating the amount of change in the time between the received packet and the reproduced packet, the clock frequency difference between transmission and reception is calculated, and correction amount information corresponding to this is created. Further, the correction control unit 6 generates a reproduction request signal e for controlling the packet transfer interval from reception to reproduction, that is, the audio reproduction interval, based on the correction amount information corresponding to the clock frequency difference.

  The audio buffer 3 temporarily changes the cycle of extracting packets from the audio buffer 3 based on the correction amount information in units of audio sample clocks.

  The voice decoding unit 7 converts the extracted packet into a voice sample data string based on the voice packet encoding method.

  The audio playback unit 8 discards or inserts one sample of data into the audio sample data string based on the correction amount information in conjunction with the cycle of extracting the packet from the audio buffer 3, and reproduces the audio data. Remove d.

FIG. 4 is a block diagram showing a configuration example of the frequency difference detection unit 5 used in the present invention.
The count counter circuit 501 performs a continuous counting operation in response to the audio reproduction timing signal f, and writes the count value 515 at the time when the transfer completion timing signal 211 is input to the memory 502.
The memory 502 sequentially stores the count value 515 together with the sequence number 212 of the received packet and the time stamp 213.
The difference circuit 503 obtains a difference from the count values 520 and 521 for the two packets read from the memory 502 and calculates a packet reception interval 522.
The memory 504 sequentially stores the reception intervals 522 obtained by the difference circuit 503.
The arithmetic circuit 505 reads the individual packet reception intervals 523 stored in the memory 504, performs an averaging process on a certain number, and calculates an average reception interval 524.
The reception interval register 509 temporarily stores an average reception interval 524 calculated from a fixed number of received packets, is sequentially updated as time elapses, and is output as a packet average reception interval 525.
The difference circuit 506 calculates a difference from the time stamps 516 and 517 for the two packets read from the memory 502 and calculates a packet transmission interval 508.
The transmission interval register 507 stores the packet transmission interval 518 calculated by the difference circuit 506.
The arithmetic circuit 508 calculates a frequency difference from the packet average reception interval 525 and the packet transmission interval 518 and outputs correction amount information 526.

The frequency difference detection unit 5 operates the count counter circuit 501 with the audio reproduction timing clock, and records the value of the count counter circuit 501 when the reception completion timing signal output from the packet reception control unit 2 is generated. Record the sequence number and time stamp of the corresponding packet.
Here, as shown in FIG. 6, two packet transmission intervals TI on the transmission side are determined by the difference value of the time stamps of two packets having two sequence numbers (k, k + n) indicating a constant difference n. Ask for.
On the other hand, the actually received time interval RIk is obtained from the difference value of the count counter values (Tk, Tk + n) corresponding to the two packets. Similarly, for 2n consecutive packets, n difference values of count counter values corresponding to two packets indicating sequence number difference n are obtained, and an average time interval RI is obtained.
The transition of the difference between TI and RI is sequentially monitored, a correction time interval at which one sample of audio data is to be discarded or inserted is calculated, and the correction control unit is notified as correction amount information.

  The present invention can be widely applied when reproducing voice packets in a system that performs voice communication via packets on a network.

It is a block diagram which shows one Example of this invention. It is a time chart for demonstrating operation | movement of the audio | voice reproduction part used for this invention. It is a block diagram which shows the structural example of the packet reception control part used for this invention. It is a block diagram which shows the structural example of the frequency difference detection part used for this invention. It is a block diagram which shows the structural example of the correction control part used for this invention. It is a figure for demonstrating operation | movement of the frequency difference detection part used for this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 LAN interface 2 Packet reception control part 2-1 Packet data reception part 2-2 Reception buffer control part 2-3 Reception timing information generation part 3 Voice buffer 4 Timing generation part 5 Frequency difference detection part 6 Correction control part 7 Voice decoding part 8 Audio playback unit 9 Buffer monitoring unit 211 Transfer completion timing signal 212 Sequence number 213 Time stamp 501 Count counter circuit 502 Memory 503 Difference circuit 504 Memory 505 Operation circuit 506 Difference circuit 507 Transmission interval register 508 Operation circuit 509 Reception interval register 515 Count value 516 Time stamp 517 Time stamp 518 Packet transmission interval 519 Packet transmission interval 520,521 Count value 522 Packet reception interval 523 Packet reception interval 524 Average reception interval 525 Average encoded reception interval a encoded audio frame data b encoded audio frame data c encoded audio data d reproduction audio data e reproduction request signal f audio reproduction timing signal

Claims (1)

A timing generation unit that outputs a sound reproduction timing signal serving as a reference for reproducing sound from the received sound packet;
A frequency difference detection unit that detects a clock frequency difference between the reception timing of the received audio packet and the audio reproduction timing signal and outputs correction amount information corresponding to the clock frequency difference;
A correction control unit that controls a frequency division ratio of the audio reproduction timing signal according to the correction amount information corresponding to the detected frequency difference and generates a reproduction request signal for reproducing an audio signal from the received audio packet;
The received voice packet is temporarily recorded, and the read / playback interval of the recorded voice packet is controlled by the playback request signal, so that the playback timing of the voice packet can be changed in units of playback clock. An audio buffer;
A voice decoding unit that converts the frame data of the voice packet into voice sample data under the control of the reproduction request signal;
An audio reproduction unit for generating reproduction data by reproducing, discarding, or inserting the audio sample data in units of the reproduction clock based on the reproduction request signal;
Voice packet playback system with

Images