JP3121094B2 - Speech coder - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech encoder mainly applied to an ATM communication network.

[0002]

2. Description of the Related Art Conventionally, when a speech coder is applied to a multiplex transmission apparatus, it is coded with a constant information amount irrespective of a change in an inputted speech signal, so that the redundancy is large. Therefore, in order to reduce the redundancy as much as possible and to enable efficient voice transmission, there is known a voice coder equipped with functions such as compression of a silent section and varying the amount of information by an input signal. . The speech encoder is considered to be applied to an ATM communication network because information generation is asynchronous.

FIG. 6 is a block diagram showing an example of the configuration of a conventional ATM encoder, and FIG. 7 is an explanatory diagram showing a signal compression state. An audio signal from the telephone is sent to a silence detector 101.
, The silent portion is detected, and the voiced portion is encoded (coded A or B) by the variable speech encoder 102 into a signal of rate A or rate B.

[0004] The encoded signal is transmitted to the cell assembler 1.
03 and is sent out to the ATM line via the transmission buffer 104. The transmission of cells from the transmission buffer 104 and the matching with the ATM line are performed by an up / down counter 105. Therefore, as shown in FIG. 7, the audio signal is encoded at the rate A or B in the sound interval, and is not transmitted (no transmission) in the silent interval, so that it is compressed and transmitted.

[0005]

As described above, when the speech encoder is applied to an ATM communication network, compression of a silent section and control of a transmission rate are performed in order to reduce the redundancy of the speech signal. However, on the other hand, the naturalness of the voice is lost, which also causes quality deterioration. In particular, there is a problem that the quality is significantly deteriorated due to the start and end of the talk in the compression of the silent section.

The present invention has been made in view of such circumstances, and when an audio encoder is applied to an ATM communication network, an application range of silence compression is limited and redundancy of an audio signal is suppressed. It is an object of the present invention to provide an audio encoder that maintains the naturalness of an audio signal.

[0007]

FIG. 1 is a block diagram showing the principle of the present invention. As shown in FIG. 1, the present invention relates to a signal compressor 1 for compressing and encoding a signal from a telephone. A cell assembler 2 for assembling cells in response to an output from a signal compressor 1, a transmission buffer 3 for temporarily storing a plurality of assembled cells, and sequentially transmitting the cells to a line, and detecting a cell storage amount of the transmission buffer 3. And a comparator 5 that compares the detected cell storage amount with a predetermined value and controls the signal compressor 1 to increase the degree of signal compression when the cell storage amount is larger than the predetermined value. The present invention provides a speech coder.

It is preferable that the comparator 5 has a timekeeping function, and controls the signal compressor 1 so as to increase the degree of signal compression when the cell storage amount is longer than a predetermined value for a predetermined time.

[0009]

The signal compressor 1 compresses and encodes the signal from the telephone, and the cell assembler 2 receives the output from the signal compressor 1 and assembles the cell. Therefore, the transmission buffer 3 temporarily stores the assembled cells and sequentially sends them out to the line.

Then, the detector 4 detects the cell storage amount of the transmission buffer 3, and the comparator 5 compares the detected cell storage amount with a predetermined value. When the cell storage amount is larger than the predetermined value, the signal compression is performed. The signal compressor 1 is controlled so as to increase the degree. Therefore, when the cell storage amount of the transmission buffer 3 is smaller than the predetermined value, the degree of signal compression is suppressed, and the quality of the transmission signal is improved.

[0011]

FIG. 2 is a block diagram showing an embodiment of the present invention. Reference numeral 11 denotes a silence detector for receiving a voice signal from a telephone, and a voice larger than a set threshold for detecting sound / silence. Is detected as a voiced section, and a low voice section is detected and output as a silent section.

Reference numeral 12 denotes a variable-type audio encoder, which encodes an audio signal at a set signal transmission rate A or B and outputs it. Reference numeral 13 denotes a cell assembler for assembling a signal encoded by the variable speech encoder 12 into cells, and 14 a transmission buffer for temporarily storing a plurality of cells output from the cell assembler 13 and sequentially transmitting the cells to an ATM line , 15 are up-down counters for matching the transmission operation of the cells from the transmission buffer 14 to the situation of the ATM line, and count up each time a cell is stored in the transmission buffer 14, and every time a cell is transmitted to the ATM line. Count down to

Reference numeral 16 denotes a state monitor, which monitors the count state of the up / down counter 15. Reference numeral 17 denotes a variable rate control unit which controls the variable type speech encoder 12 based on the presence / absence of sound / non-speech information from the silence detector 11 and the information from the state monitoring unit 16 to set the signal transmission rate to A or B. Switch.

Reference numeral 18 denotes a sound / non-speech determining unit, which receives the information from the state monitoring unit 16 and makes a sound / no-sound judgment. 1
Reference numeral 9 denotes a variable part of a threshold for detecting sound / non-speech, which changes the threshold for sound / non-speech detection of the non-speech detector 11 in response to information from the state monitor 16.

In such a configuration, the voice signal from the telephone is coded in one of the following four modes as shown in FIG. 15 to the ATM line.

[First Mode] All voice signals are coded (coded A) at a fixed rate A without performing voice compression with voice / no voice. [Second mode] Speech signals are encoded at a rate A or B (A <B) without performing speech compression based on speech / non-speech.

[Third Mode] Voice / silence is detected by a small voice / silence detection threshold value, and transmission is not performed in a voiceless section, but only a voice signal in a voiced section is encoded at a rate A. . [Fourth mode] Speech / silence is detected by a large speech / silence detection threshold value, and transmission is not performed in a silence section, and only a speech signal in a speech section is encoded at a rate A.

Next, the operation of selecting these modes will be described with reference to the flowchart shown in FIG. First, when off-hook information is input from the telephone set to the silence detector 11 (step 201), a voice signal from the telephone set is encoded in the first mode, assembled into a cell, and transmitted via the transmission buffer 14 to the ATM. It is sent to the line (step 202).

When congestion occurs on the ATM line, the number of cells (buffer amount x) held in the transmission buffer 14 increases, but the buffer amount x is compared with threshold values a, b, and c by the state monitor 16. You.

The threshold values a, b, and c are set according to the size of the transmission buffer 14, as shown in FIG.
<B <c. Therefore, step 203
If x ≦ a, the first mode is continued (step 204). When a <x <b is satisfied (step 205), the variable rate control unit 17 operates to execute the second mode (step 206).

When b ≦ x <c (steps 205 and 2)
07), the sound / silence determining unit 18 operates the silence detector 11, and the third mode is executed (step 208).
Next, when x ≧ c, the threshold value for sound / non-sound detection is set to be larger by the operation of the variable unit 19, and the fourth mode is executed (step 209). When the on-hook information is input, all the operations are stopped.

As described above, an appropriate mode is selected according to the congestion state of the ATM line, and voice encoding is performed while the amount of information compression is suppressed to a minimum.

FIG. 8 is an explanatory diagram showing an example of a temporal change in a section in which sound information ch1 to ch8 is generated.
When the start time and the end time of the audio information ch1 to ch8 are instantaneously overlapped as shown in the period T, the compression ratio is frequently changed by detecting this and performing audio compression for a short time. However, the transmission efficiency is not improved, but only the voice quality is deteriorated. Therefore, in the above embodiment, in order to protect it, the state monitoring unit 16 is provided with a clock function, and the count value of the up / down counter 15 is set to a predetermined time (for example, the time required for generating three cells,
The state monitoring unit 16 changes the audio compression ratio only when the threshold value a, b, or c exceeds or falls below the threshold value a, b, or c.
7. It is desirable to output to the sound / non-sound determining unit 18 and the variable unit 19.

[0024]

According to the present invention, the speech quality degradation that has always occurred at the time of compressing the information amount of the speech encoder can be reduced by A
This can be suppressed only when congestion occurs on the TM line, and can contribute to maintaining the voice quality when the voice encoder is applied to the ATM communication network.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the principle of the present invention.

FIG. 2 is a block diagram showing one embodiment of the present invention.

FIG. 3 is an explanatory diagram illustrating a relationship between a transmission buffer size and a threshold according to the embodiment;

FIG. 4 is an explanatory diagram showing each mode of the embodiment.

FIG. 5 is a flowchart showing the operation of the embodiment.

FIG. 6 is a block diagram showing a conventional example.

FIG. 7 is an explanatory diagram showing a signal compression state in a conventional example.

FIG. 8 is an explanatory diagram showing an instantaneous change in the amount of information.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 Silence detector 12 Variable speech coder 13 Cell assembler 14 Transmission buffer 15 Up / down counter 16 State monitoring unit 17 Variable rate control unit 18 Voice / silence determination unit 19 Variable unit

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int. Cl. ⁷ , DB name) H04B 14/00-14/06 H04L 12/28 H04L 12/56

Claims (2)

(57) [Claims] 1. A signal compressor (1) for compressing and encoding a signal from a telephone, and a cell assembler (2) for assembling a cell by receiving an output from the signal compressor (1). A transmission buffer (3) for temporarily storing a plurality of cells and sequentially transmitting the cells to a line, a detector (4) for detecting a cell storage amount of the transmission buffer (3), and a detected cell storage amount and a predetermined value. An audio encoder including a comparator (5) for controlling the signal compressor (1) to increase the degree of signal compression when the cell storage amount is larger than a predetermined value. 2. The comparator (5) has a timing function, and controls the signal compressor (1) so as to increase the degree of signal compression when the cell storage amount is continuously larger than a predetermined value for a predetermined time. The speech encoder according to claim 1.