JPH0530137A - Sound packet transmission device - Google Patents

Abstract

PURPOSE:To constitute a sound packet transmission device in which the circuit scale of a transmission-side transmission device is small and which has superior characteristic. CONSTITUTION:A background noise detection part 108 which detects the level of background noise by extracting background noise included in the sound signal of a ringing section by using that the normality of background noise is high in a reception-side transmission device, and the generation level of background noise is decided by the output of the background noise extraction part 108. Thus, a noise generation part 111 in a reception-side can generate the background noise of a silent section without transmitting a noise packet showing the level of background noise from the transmission-side transmission device.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice packet transmission device for digitally encoding voice from a voice terminal such as a telephone and packetizing it for transmission.

[0002]

2. Description of the Related Art Recently, a method of packetizing a digitized voice signal and transmitting it has been used. Since voice has a voiced section in which voice information is actually present and a silent section called gap, pause, or silence, it is possible to effectively use the transmission path by packetizing only the voiced section and transmitting it. .. In such a case, if the voice output in the silent section is completely silenced on the receiving side, the background noise included in the voice is also cut and it becomes unnatural in the sense of hearing. Play the noise.

A conventional voice packet transmission device will be described below. FIG. 2 shows the configuration of a conventional voice packet transmission device. On the transmission side of FIG. 2, 201 is an input terminal, 202 is an encoder for encoding voice, and 203 is a packet assembling unit for assembling voices digitally encoded by the encoder 202 into packets. Also, 20
Reference numeral 4 denotes a noise level detecting unit that detects the level of background noise from the silent section of the input voice, and 205 denotes a noise level detecting unit 204.
A noise packet assembler for assembling a noise packet for transmitting the noise level detected by the receiver to the receiving side, 20
Reference numeral 6 denotes a voice / sound determination unit that determines a voiced section and a silence section of the voice from the input voice, and 207 denotes a voiced / soundless determination section 206.
It is a packet sending unit that selects one of the outputs of the packet assembling unit 203 and the noise packet assembling unit 205 according to the determination result of (1) and sends it to the receiving side. Also, on the receiving side, 20
8 is a packet decomposing unit for decomposing the packet transmitted from the transmitting side, 209 is a decoder for decoding the voice code transmitted by the packet from the transmitting side, and 210 is noise transmitted by the noise packet from the transmitting side. A noise generation unit that reproduces background noise in a silent section according to the level, and a voice signal by switching a voice signal of the decoder 209 and a noise signal from the noise signal generation unit 210 according to the voice / silence instruction signal 211 from the packet decomposition unit 208. The switching unit outputs from the output terminal 213.

FIG. 3 shows a packet format used by the voice packet transmission device of FIG. In FIG. 3, reference numeral 301 is a voice packet that transmits a voice code in a voiced section, and 305 is a noise packet that transmits a background noise level in a silent section.

The operation of the voice packet transmission apparatus configured as shown in FIG. 2 will be described below. First,
In the voiced section of the voice, the voice code from the encoder 202 is assembled into a voice packet indicated by 301 in FIG.
Sent from. On the receiving side, a voice code is extracted from this voice packet by the packet decomposing unit 208, and the voice decoded by the decoder 209 is output from the voice output terminal 213 via the switching unit 212.

Further, in the silent section of the voice, the noise level detecting section 204 detects the background noise level. In order to transmit this noise level to the receiving side, the noise packet assembling unit 205 assembles the noise packet shown at 305 in FIG. At this time, the sound / silence determination unit 20
6 sends a silence detection signal to the packet sending unit 207,
The packet sending unit 207 replaces the voice packet 301 from the packet assembling unit 203 with the noise packet assembling unit 205.
Noise packet 305 from On the other hand, on the receiving side, the noise generation unit 210 generates background noise according to the background noise level extracted from the packet by the packet decomposition unit 208. The switching unit 212 is the packet decomposition unit 2
In accordance with the voice / silent instruction signal 211 from 08, the voice output from the voice signal from the decoder 209 is generated by the noise generating unit 21.
The background noise from 0 is switched to output to the audio output terminal 213.

As described above, even in the above-mentioned conventional voice packet transmission device, it is possible to reproduce the background noise in the silent section during the voice packet transmission.

[0008]

However, in the above-mentioned conventional voice packet transmission device, since the noise packet for notifying the receiving side of the background noise level must be transmitted even in the silent section, the utilization efficiency of the transmission line can be improved. It had a problem of making it worse.

The present invention solves the above-mentioned problems of the prior art. By transmitting packets only in the voiced section, the transmission path can be effectively used, and the function of performing noise packets on the transmission side is unnecessary. Therefore, it is an object of the present invention to provide an excellent voice packet transmission device capable of significantly reducing the circuit scale on the transmission side.

[0010]

In order to achieve the above object, the present invention detects the level of background noise which is included in a voice signal in a voiced section and has a characteristic of high stationarity on the receiving side, and thereby the silence is generated. The reproduction level of the background noise in the section is determined.

[0011]

With this structure, even if the noise packet indicating the background noise level is not transmitted from the transmitting side, the receiving side can generate the background noise in the silent section at an appropriate level.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

In the transmitting side of FIG. 1, 101 is an input terminal, 102 is an encoder for encoding voice, and 103 is a packet assembling unit for assembling voices digitally encoded by the encoder 102 into packets. Further, reference numeral 104 denotes a voiced / non-voiced determination unit that determines a voiced section and a voiceless section of voice from voice input, and 105 indicates a packet assembly section 1 only in the voiced section according to the determination result of the voice / non-voice determination section 104.
It is a packet sending unit that sends out the packet assembled by the packet No. 03. Also, on the receiving side, 106 is a packet decomposing unit for decomposing packets transmitted from the transmitting side, 1
Reference numeral 07 is a decoder for decoding the voice code transmitted from the transmission side in packets, 108 is a background noise extraction unit for detecting the level of background noise from the voice in the voiced section transmitted from the transmission side, and 111 is A noise generation unit for generating noise according to the background noise level determined by the background noise extraction unit 108, and 112, a voice signal from the decoder 107 and noise generation by the voice / silence instruction signal 113 from the cell decomposition unit 106. It is a switching unit that switches the noise signal from the unit 111 and outputs it from the audio output terminal 114.

The operation of the voice packet transmission apparatus configured as described above will be described with reference to FIG. First, the packetization of voice on the transmitting side and the reproduction from the voice packet on the receiving side in the voiced section are the same as the operation in the conventional voice packet transmission apparatus described with reference to FIG. Further, in this voiced section, the background noise extraction unit 108 on the receiving side determines the noise level used in the next silent section. The embodiment shown in FIG. 1 is characterized in that the audio signal has a non-stationary characteristic and thus has a low temporal correlation, whereas the background noise is relatively stationary and thus has a high correlation. By utilizing this, the background noise level is determined from the voiced section. That is, first, the autocorrelation calculation unit 109 calculates the autocorrelation coefficient in the voiced section. Next, the noise level determination unit 110 extracts only the autocorrelation coefficient for a time longer than a predetermined threshold value from the autocorrelation coefficients calculated by the autocorrelation calculation unit 109. Since the background noise has a high stationarity as described above, it can be said that the autocorrelation coefficient for a time longer than a certain level is characteristic of the background noise. Therefore, the noise level of the background noise can be determined from the magnitude of the coefficient extracted here. In accordance with this level, the noise generator 111 generates background noise to be reproduced in the next silent section.

Next, the operation of the voiceless section will be described. In the silent section, the transmitting side has a sound / silence determination unit 10
4 determines that there is no sound, the packet sending unit 105 stops sending packets in response to an instruction from the unit 104. On the receiving side, packets do not arrive from the transmitting side, so
The packet disassembling unit 106 determines that there is a silent section, and the switching unit 11
2 is notified by the voiced / non-voiced determination signal 113 that there is no voiced section. In the switching unit 112, the noise generating unit 111
Noise generated according to the background noise level determined by the background noise extraction unit 108 is output from the audio output terminal 114.

[0016]

As described above, the present invention determines the background noise level generated in the silent section by extracting the characteristics of the background noise included in the voice signal in the voiced section, and determines the background noise level according to the level. By providing a noise generation unit that generates noise, it is not necessary to send out noise packets in a silent section, effective use of the transmission path can be achieved, and an excellent voice packet transmission device with a small circuit scale on the transmission side can be provided. It can be realized.

[Brief description of drawings]

FIG. 1 is a block diagram of a voice packet transmission device according to an embodiment of the present invention.

FIG. 2 is a block diagram of a conventional voice packet transmission device.

FIG. 3 is a diagram showing a packet format used in a conventional voice packet transmission device.

[Explanation of symbols]

 102, 103 encoder 103, 203 packet assembling unit 204 noise level detecting unit 205 noise packet assembling unit 104, 206 voice / silence determining unit 105, 207 packet transmitting unit 106, 208 packet disassembling unit 107, 209 decoder 111, 210 noise generator 114,212 switching unit 301 voice packet 305 noise packet

Claims (1)

Claim: What is claimed is: 1. A packet decomposing unit that receives a voice packet sent from a transmitting side and decomposes the content of the packet.
A decoder for decoding the voice code extracted from the packet by the packet decomposing unit, a background noise extracting unit for determining the level of background noise included in the voice signal from the decoder and having a characteristic of high stationarity, A noise generator that generates background noise according to the output of the background noise extractor, and a voice / noise switching unit that switches and outputs the decoded voice by the decoder and the background noise generated by the noise generator. A voice packet transmission device comprising: