JPH01241240A - Voice packet processor - Google Patents

Abstract

PURPOSE:To update background noise by sending a silence packet as background noise information. CONSTITUTION:Packet information 1-4 compressed by a sound presence/silence discrimination circuit 14 is combined as a packet and fed to a sound presence/ silence deciding circuit 18. In case of the silence packet, the result is outputted from a buffer 21 as packet information 109 corrected based on a 3rd packet information and a 7th packet information. Thus, it is not required to recognize the silence decision value at the transmission side and the processing in the deciding circuit 18 is simplified.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an audio packet processing device capable of reproducing high quality audio faithful to the original audio.

[Conventional technology]

In voice packet communication, a method of transmitting only voice packets (silence compression) is common in order to increase the efficiency of communication channel usage. However, when playing back compressed silence, if background noise is not taken into consideration, the decoded audio will sound as if the line has been disconnected at the silence.

It's extremely unnatural.

Conventionally, as a method to solve this unnaturalness, JP-A-57-
190999, a method has been proposed in which information predetermined on the receiving side is reproduced at a double tone level based on temporal information of a silent section in a header portion of a voiced packet.

[Problem to be solved by the invention]

In the above-mentioned conventional technology, no consideration was given to the noise level during silent sections during conversation. In other words, the background noise level generated on the receiving side is different from the actual background noise, and the audio level of background noise in normal conversation generally changes from moment to moment. Noise reproduction can be a factor in deteriorating sound quality.

An object of the present invention is to treat the background noise that changes from moment to moment as a piece of sound information, thereby making it possible to transmit the actual background noise level and update the background noise.

The objective is to improve the quality of reproduced audio.

[Means to solve the problem]

The configuration of the present invention for achieving the above object has the following features.

1. When input audio signals are packetized, each time it is determined that there is no sound, one or more packets are sent out as valid packets, and the contents are stored in the receiving side memory.

2. By repeatedly reproducing the information stored in this memory until the next sound packet arrives, a silent section of arbitrary length is reproduced according to the actual noise level.

[Effect]

A silent packet containing background noise information is sent every time it is determined that there is no sound, so even if a silent section occurs again, new noise information is sent as a packet and the information in the memory is replaced to eliminate the background noise. can be updated.

〔Example〕

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a conceptual block diagram showing an embodiment for implementing the present invention. In FIG. 1, the transmitting side is connected to an audio input terminal 11 and a PCM that encodes the input analog signal.
An encoder 11, a packet assembling unit 13 that divides the PCM signal into predetermined packet lengths, a voice/silence determination circuit that determines whether a voice packet is a voice packet or a silent packet; It consists of a packet transmitting section 15 that sends packets to. On the other hand, the receiving side includes a packet receiving unit 16 that receives packets, a voice/silence determination circuit 18 that determines voice/silence of the received packet, and controls the submemory 19 and selector 20 as described below. A sub-memory 19 for storing silent part information, which will be described later, under the control of
a selector 20 that selects the output information 108 and outputs it to a buffer 21; a buffer 21 that absorbs delay variations within the network of voice packets sent from the selector 20; a packet decomposer 22 that converts the packets into PCM signals; and an OCM. It consists of a PCM decoder 23 that converts the signal into an output analog signal and an audio output terminal 24.

Analog audio signal 101 input to audio input terminal 11
Lt, is converted into a PCM signal 102 by a PCM encoder 12. The PCM signal 102 is assembled into audio packet information 103 of a predetermined length by a packet assembly section 13. Further, this packet information 103 is judged as sound/silence by comparing the packet information value with a preset silence judgment value in the sound/silence judgment circuit 14. The packet determined to be valid packet information 104 is sent to the packet transmitter 15. On the other hand, the packet information determined to be silent following the packet information determined to be sound by the determination circuit 14 is sent to the packet transmitter 15 for background noise reproduction. The second and subsequent consecutive packets that are determined to be silent by the determination circuit 14 are discarded. As described above, the packet transmitter 15
All the packets sent to are valid packets, and the packet transmitter 15 sends these packets to the transmission path 16.
Sequentially.

The voice packet arriving at the packet receiving section 17 is sent to the voice/silence determining circuit 18. The received audio packet 105 is sent to the comparison circuit 18-(b) in the judgment circuit 18, and at the same time it is sent to the selector 20 from the judgment circuit 18.
Output to 106. It is sent as output 107 to submemory 19. In the determination circuit 18, the silence determination value storage unit 1
The output 113 of 8-(a) is compared with the size of the audio packet 105 to determine whether or not there is a sound. The audio packet 105
When there is a sound, the determination circuit 18 selects the sub-memory control line 11.
5 prohibits writing of the output 107 to the sub-memory 19, and instructs the selector 20 to select the output 106 using the selector control line 1]4. On the other hand, when the audio packet 105 is silent, the sub memory 19 of the output 107 is controlled by the sub memory control line 115.
the selector control line 11.
4 instructs to select the output 108 from the sub-memory 19. The control by the selector control line 114 is continued until a new bucket l- arrives, and for example, when silent part compression is performed on the transmitting side as described above, the first sound packet arrives at the determination circuit 18. The packet group 110 whose delays have been aligned in the 109° buffer 21 is converted into a PCM signal 111 by the packet decomposition unit 22, and converted into an analog signal by the PCM decoder 23. After being decoded, it is output from the audio terminal 24.

FIG. 2 is a time chart schematically showing behavior 9 of voice packets on the network. In FIG. 2, in each packet group, packets with sloped lines indicate voice packets, and white packets indicate silent packets. Further, the numbers attached to the packets indicate the order in which each packet was created.

In FIG. 1, the packet information output 103 generated by the packet assembling section 13 is sent to the speech/silence determination circuit 14. The determination circuit 14 determines whether the packet information output 103 is voiced or not, and discards the second and subsequent silent packet information 209, 210° 211 of the silent section. That is, the packet information 104 is assembled into a packet and sent to the transmission path 16.

The packet information 105 that has arrived at the receiving side is sent to the sound/silence determination circuit 18 . First packet information 201 which is sound information. The second packet information 202 is sent to the buffer 21 after receiving a voice determination in the determination circuit 18 . The third packet 203 (a), which is a silent packet, is determined to be silent by the determining circuit 18, is written into the sub-memory 19, and is sent to the buffer 21. Said sub memory 19
The information 203 (b) of the third packet stored in the fourth packet 204 . is stored in the fourth packet 204 . It is read out as the information of the fifth packet 205. The sixth packet information 206, which is new voice packet information, is sent to the determination circuit 1.
If it is determined that there is a sound at step 8, the packet 206 is sent to the buffer 21. The seventh silent packet is a new silent packet.
When the packet information 207 (a) is determined to be silent by the determination circuit 18, it is newly written to the sub-memory 19 and then output to the buffer 21. Said sub memory 1
The seventh packet information 207 (b) stored in the packet 9 is read out as the information of the eighth packet 208 according to the instruction from the determination circuit 18 .

The packet information 104 compressed by the sound/silence determination circuit 17 as described above is divided into the third packet information 203(a), the seventh packet information 203(a),
The buffer 21 outputs the packet information 109 corrected based on the packet information 207 (b).

Table 1 shows an example of the length of silence during a standard conversation.
This figure shows the influence of this embodiment (one packet containing background noise information is sent in a silent section) on silence compression compared to the conventional method that did not send silence information.

Table 1 Here, μ represents the average value and σ represents the S standard deviation.

Assuming that the encoder uses 32 Kbps ADPCM and 31 Bytes are used for one packet of audio data, one packet of audio corresponds to 7675 μsec.

Average silent time μ is 1.7 seconds (time equivalent to 219 packets)
This is extremely long, and the loss in this embodiment is only 0.46%. Also, even if we take into account the dispersion of the data,
The loss of silence compression is 0.70% or less. If the loss of silent section compression increases significantly, it will have a large effect on the network utilization efficiency, but with a loss of the level of this embodiment, this effect is extremely small.

Further, as an example of application of the present embodiment, in FIG.
One possible method is to attach the determination result to the header of No. 4 (for example, 1 for sound, 0 for no sound). In this case, a 1-bit judgment is sufficient for the speech/non-speech judgment of the received packet 105 in the receiving side speech/silence judgment circuit 18, and furthermore, the above-mentioned judgment does not require prior knowledge of the silence judgment value on the transmitting side. The processing in the circuit 18 can also be simplified.

In this embodiment, only the first packet of the silent portion is sent, but the number of packets to be sent is a matter of silence compression efficiency and is not essential to the present invention.

Furthermore, although this embodiment shows the reproduction of a silent part following a sound part, for example, even if a silent part occurs at the same time as a call is set up, the silent part can be detected and the background noise information can be sent to reproduce the silent part. Reproduction is possible.

〔Effect of the invention〕

According to the present invention, since silent packets are sent as background noise information, it is possible to approximate the actual noise and update the noise, thereby improving the quality of reproduced audio.

[Brief explanation of the drawing]

FIG. 1 is a conceptual block diagram showing the configuration of an apparatus according to an embodiment of the present invention, and FIG. 2 is a time chart schematically showing the behavior of voice packets on the network. 11... Audio input terminal, 12... PCM encoder, 1
3... Packet assembling section, 14... Sending side sound/silence determination circuit, 15... Packet transmitting section, 16... Transmission path, 17... Packet receiving section, 18-... Receiving side Sound/silence determination circuit, 1B (a)... Silence determination value storage section, 18(b)... Comparison circuit, 19... Sub memory,
2o...Selector, 21 Buffer, 22...Packet disassembly unit, 23...PCM decoder, 24-mount audio output terminal, 114(・-1,6(\ secret)

Claims (1)

[Claims] 1. An audio packet that digitizes an analog audio signal, divides it into blocks of a certain length, determines whether each block is sound or no sound, and creates a packet based on the result and sends it to a transmission path. An audio packet processing device, characterized in that the processing device includes means for transmitting one or more blocks determined to be silent as packets. 2. means for determining whether a received packet is a voice packet or a silent packet; means for storing the received packet based on the result of the determination means; 2. The audio packet processing device according to claim 1, further comprising means for selecting output from the storage means. 3. The audio packet processing device according to claim 1, further comprising means for adding information indicating that the transmitted packet is a packet generated from a block determined to be silent to the transmitted packet. .