JPH024064A - System for reproducing silent section for voice packet communication - Google Patents

Abstract

PURPOSE:To reduce a difference between the level of a background noise contained in a sound section and the level of a signal in a silent section by repeating a sample stored in a silence storing part in the silent section, sending it to a digital/analog converting part and reproducing the silent section. CONSTITUTION:At a transmitting side, one part of the signal in the silent section is transmitted, the signal is received at a receiving side, and thereafter, it is stored in a silence storing part (memory) 9. In the silent section, the sample stored in the silence storing part 9 is repeated until a block in a next sound section is received, it is sent to a digital/analog converting part 12, it is converted to an analog signal, and the silent section is reproduced.

Description

[Detailed Description of the Invention] [Industrial Application Field] This invention digitizes an analog audio signal on the transmitting side, packetizes only the sound section of the digitized sample sequence, and sends it to the transmission path. The present invention relates to a silent section reproduction method in a voice communication system in which communication is performed by extracting a sample string of a sound section from a received packet and converting it back into an analog audio signal.

[Conventional technology]

The transmitting side digitizes the analog audio signal, packetizes only the active section of the digitized sample string, and sends it to the transmission path.The receiving side extracts the sample string of the active section from the received packet and converts the analog audio signal into a packet. In voice communication systems that communicate by converting signals back to signals, if there is line noise in the analog signal on the transmitting side or when the caller is talking in a noisy environment, if the silent section is played back to zero on the receiving side. , line noise and background noise are completely eliminated during the silent section, making the call unnatural. In order to suppress this unnaturalness, a fixed level of noise is generally inserted into the silent section on the receiving side.

[Problem to be solved by the invention]

In the method of inserting a fixed level of noise, the noise may become unnatural if the background noise of the active section (line noise or background noise when the caller is talking in background noise) and the inserted noise are different. There is a problem with that.

An object of the present invention is to digitize an analog audio signal on the transmitting side, packetize the sound section of the digitized sample string, and send it to a transmission path, and on the receiving side, digitize the sample string of the sound section from the received packet. In a voice communication system that performs communication by extracting an analog voice signal and converting it back to an analog voice signal, this method solves the problem that the level of background noise included in a voice section and the level of noise inserted into a silent section on the receiving side are thick (different). An object of the present invention is to provide a silent section reproduction method for voice packet communication.

[Means to solve the problem]

In this invention, the transmitting side transmits a part of the signal during the silent section, and the receiving side stores this signal in the silent storage section (memory) after receiving it. The samples are repeatedly sent to a DAC (digital-to-analog converter) section until the next block of a sound section is received, and the samples are converted into an analog signal and the silent section is reproduced.

[Effect]

In this invention, the transmitting side transmits a part of the signal during the silent section, and the receiving side stores this signal in the silent storage section (memory) after receiving it. The samples are repeatedly sent to the DAC (digital-to-analog converter) section until a sound section block is received and converted into an analog signal to reproduce the silent section, so that the level of background noise included in the sound section can be adjusted. The level of the signal inserted into the silent section on the receiving side does not vary greatly.

[Embodiment] FIG. 1 is a block diagram showing the transmitting side of an embodiment of the present invention, and FIG. 2 is a block diagram showing the receiving side.

In these figures, 1 is an analog signal input terminal, 2 is a low-pass filter section, 3 is an ADC (analog-digital conversion) section, 4 is an audio detection section, 5 is a packet assembly section, and 6
is a transmission line for transmission, 7 is a transmission line for reception, 8 is a packet decomposition unit, 9 is a silent storage unit (memory), 10 is a delay fluctuation compensation unit, 11 is a switching unit, and 12 is a DAC (digital-to-analog conversion) unit. , 13 is a low-pass filter section, and 14 is an analog signal output terminal.

Next, an example of operation will be explained.

First, the operation on the transmitting side will be described with reference to FIG.

The analog signal input from the analog signal input terminal 1 has its high-frequency components cut by the low-pass filter section 2, and then
It is converted into a digital signal by the ADC section 3, and then sent to the audio detection section 4.
and is sent to the packet assembly section 5.

The voice detection section 4 determines whether there is a sound section or a silent section based on the digitized sample sequence, and sends the determination result to the packet assembly section 5. Methods for determining whether there is a voiced section or a silent section include a method that determines a voiced section when the short-term average power of the voice exceeds a threshold value, and a method that combines this with the number of zero crossings. be.

If it is a sound section, the packet assembling unit 5 divides the sample sequence into audio blocks corresponding to the packet length, adds information necessary for block transfer control to each block as a header, and transmits it as a packet for transmission. Sending out from road 6,
If it is a silent section, the silent section of an appropriate length following the sound section is also treated as a sound section, and after dividing it into blocks corresponding to the packet length, each block is divided into blocks with information necessary for block transfer control as a header. , and sends it out from the transmission line 6 as a packet. This is equivalent to increasing the hangover time of the voice detection section 4.

Next, the operation on the receiving side will be explained with reference to FIG. The packet decomposition unit 8 sends the block of the sound section received from the reception transmission line 7 to the silence storage unit 9 and the delay fluctuation compensator 10. In addition, as a method for identifying the final block of a sound interval, the packet disassembly unit 8 divides the block into
There is a method in which a packet is transferred with an identifier indicating that it is the last block of a sound section, and the receiving side determines the last block of the sound section based on this identifier.

The silence storage section 9 is composed of memory for a plurality of samples,
The most recent samples received are stored in this memory while a block is being received. The delay fluctuation compensator 10 adds a delay to each block to make the different transfer delay time constant for each block, and then transfers the sample sequence in the block through the switching unit 11 to the DACAlB 12.
Then, the DAC AlB 12 converts the received sample string into an analog audio signal. This analog audio signal is passed through the low-pass filter section 13 to the analog signal output terminal 14.
is output from.

When the packet decomposition unit 8 finishes sending the last block of the sound section to the silence storage section 8 and the delay fluctuation compensator 10, it sends a signal indicating that this block is the last block of the sound section (hereinafter referred to as the final block signal). ) is sent to the delay fluctuation compensation unit 10 and the silence storage unit 9. At this time, the silence storage unit 9 stores a plurality of samples after the end of the sound section, that is, a plurality of samples after the silence section of an appropriate length following the true sound section. When the silent storage unit 9 receives the final block signal, it stops updating the memory until it receives a voiced section start signal, which will be described later.

The delay fluctuation compensator 10 passes all the samples of the received block along with the final block signal through the switching unit 11, and when the DAC AlB 12 finishes, it sends a signal indicating the end of the sound section (hereinafter referred to as the sound section end signal). , switching section 11
send to The switching unit 11 normally includes the delay fluctuation compensating unit 10
The DAC AlB 12 is connected to send the samples sent from the DAC AlB 12, but when it receives the sound section end signal, the connection is switched so that the samples from the silent storage section 9 are sent to the DAC section. The silence storage section 9 repeatedly sends the accumulated samples to the DAC section.

When the packet decomposition unit 8 receives a block of a sound interval after a silent interval, it sends a sound interval start signal to the silence storage unit 9 and also sends a sound interval start signal to the switching unit 11 via the delay fluctuation compensator 10.
send to When the silent section start signal is received, the silent storage section 9 starts updating the memory so that the latest plural samples received are stored in the memory. In addition, the switching unit 11
When receiving the sound interval start signal, the DAC AlB 12 switches the connection so that the sample sent from the delay fluctuation compensator 10 is sent to the DAC AlB 12.

FIG. 3 is an explanatory diagram showing an example of the configuration of a packet sent from the transmitting side to the transmission path.

In FIG. 3(a), packet 1, which is the first packet to be sent, is a packet that includes a sound section and a silent section (silent part) of an appropriate length, and the continuation (packet 2) is a packet with a similar structure. On the receiving side, the silent part (multiple samples) of each packet is stored in the silent storage section 9 each time, and is repeatedly converted into an analog signal by the DAC AlB 12 until the next block (packet 2) of the sound interval is received. This is a method used for reproducing a silent section, and is the method adopted in the above-mentioned embodiment.

In the method shown in Figure 3(b), the packet structure including a sound section and a silent section of an appropriate length is implemented only for the sound section immediately after the start of communication, and the receiving side The silent part (samples) of the packet is stored in the silent storage section 9, and is repeated in all subsequent silent intervals until the next block (packet) of the active interval is received. This is a method of converting it into a signal and using it to reproduce the silent section, and such a method can also be adopted.

In the method shown in FIG. 3(c), a silent period of an appropriate length immediately after the start of communication is also packetized and transmitted as packet 0, and on the receiving side, samples of the silent period are also stored in the silent storage section 9. This is repeated until the block (packet 1) of the first sound section is received.
AlB12 is a method for converting into an analog signal and using it for reproducing silent sections, and such a method may be adopted.

In the method shown in FIG. 3(d), a silent section of an appropriate length immediately after the start of communication is packetized and transmitted as packet 0, and thereafter only the sound section is packetized and transmitted. Samples of silent sections received immediately after the start of communication are stored in the silent storage section 9, and the samples are repeatedly converted into analog signals by the DAC AlB 12 until the next block (packet) of the active section is received in all silent sections. This method is used to reproduce silent sections, and such a method may be adopted.

In addition, in the above-mentioned embodiment of the present invention, a case was described in which a sample in which voice detection was digitized was used.
Even when audio detection is based on an analog input signal,
A similar implementation can be achieved by sending the judgment result of the voice detector to the packet assembling section 5.

In addition, in the embodiment of the present invention, it is assumed that the hangover time of the voice detection unit 4 is short, and a silent section of an appropriate length following a sound section is also packetized and transferred. However, if the hangover time is long, A voice detection unit is used to packetize and transmit only the sections determined to be audible, and the receiving side accumulates multiple samples at the end of the audible sections.
The present invention can also be implemented by repeatedly sending this signal to the DAC unit during silent periods.

〔Effect of the invention〕

According to this invention, the transmitting side transmits a part of the signal in the silent section, and the receiving side stores this signal as a sample in the silent storage section, and in the silent section, the sample stored in the silent section is stored as a sample. Repeat DAC until tone interval block is received
Since the silent section is played back by converting it into an analog signal sent to the receiver, the level of the background noise included in the active section and the level of the signal inserted into the silent section on the receiving side do not differ greatly, and the signal can be reproduced naturally. It has the advantage of being able to realize flexible voice communication.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the transmitting side of an embodiment of the present invention, FIG. 2 is a block diagram showing the receiving side, and FIG. 3 is an explanation showing an example of the structure of a packet sent from the transmitting side to the transmission path. Figure. Explanation of symbols 1: Analog input terminal, 2: Low-pass filter section, 3: ADC (analog-digital conversion) section,
4...Audio detection unit, 5...Packet assembly unit, 6...
- Transmission transmission line, 7... Reception transmission line, 8... Packet disassembly unit, 9... Silence storage unit (memory), 10...
- Delay fluctuation compensation section, 11... switching section, 12... D
AC (digital/analog conversion) section, 13...Low pass filter section, 14...Analog signal output terminal No.
3 Illustration agent Patent attorney Akio Namiki

Claims (1)

[Claims] 1) The transmitting side digitizes the analog audio signal, packetizes the digitized sample string with a sound section, and sends the packet to the transmission path, and the receiving side converts the sample string of the received sound section packet into a packet. In a voice communication system that performs communication by converting the audio signal back into an analog audio signal, the transmitting side also packetizes a part of the silent section that precedes or follows the sound section and sends it to the transmission path, and the receiving side , after receiving this, the silent section reproduction method for voice packet communication is characterized in that the packets are decomposed and stored in a memory, and the stored data is used to reproduce the required silent section.