JP2638522B2 - Audio coding device - 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 coding apparatus, and more particularly to a speech coding apparatus used in a digital mobile communication type mobile telephone system.

[0002]

2. Description of the Related Art As shown in FIG. 2, in a conventional voice coding apparatus used in a digital mobile communication type mobile telephone system, a data receiving section (2) inputs coded data from an input terminal (1). And the power index (2a), the adaptive code index (2b), and the noise code index (2
c), gain code index (2d) and LSP (Line S
(pectrum Pair) index (2e) and output. Further, the data receiving section (2) outputs the input encoded data to a VOX (Voice Operated Transmitter) unique word detector (3) (2f).

[0003] In order to reduce power consumption, a mobile station may transmit a unique word of a VOX and stop transmitting an encoding index when an encoder determines that an uplink speech is a silent section.

[0004] The VOX unique word includes a postamble for notifying the stop of the transmission of the coding index and a preamble for notifying the start of the transmission of the coding index.

[0005] When a silent section continues, encoded data for periodically updating the postamble and the background noise is transmitted.

A VOX unique word detector (3) detects a postamble and a preamble output from a mobile station, and VOX MOD indicating a state of a sound section or a silent section.
E information (referred to as "VOX mode information") is output.

The VOX mode information includes, for example, a state "-1" when a preamble is detected, a state "1" when a postamble is detected, and a state "2" when a coded index following the postamble is received. The state until reception is set to state “3”, and the state “0” is output when the encoded index is received after the detection of the preamble.

[0008] The power code decoder (4) derives a quantized power value (4) from the power codebook based on the power index (2a).
Output a).

[0009] The LSP code decoder (5) converts the LSP codebook to the LPC (5) based on the LSP index (2e).
a) and the prediction residual gain (5b) are output.

The gain code decoder (6) receives the quantized power value (4a), the prediction residual gain (5b) and the gain code index (2d), and adapts the gain code book based on the gain code index. A code gain coefficient (6a) and a noise code gain coefficient (6b) are output.

[0011] The noise code signal decoder (7) generates a noise code vector signal (7) based on the noise code index (2c).
Output a).

The adaptive code signal decoder (8) generates an adaptive code vector signal (8) based on the adaptive code index (2b).
Output a).

The adaptive codebook used in the adaptive code signal decoder (8) is updated by the excitation signal (11a) to generate a new adaptive codebook.

A multiplier (9) multiplies the adaptive code vector signal (8a) by the adaptive code gain coefficient (6a) and a noise code gain coefficient (6b) by the noise code vector signal (7a). The output signal multiplied by (10) is added by an adder (11) to generate an excitation signal (11a).

The synthesis filter (12) is an LPC (5a)
And an excitation signal (11a), and outputs a synthesized voice signal (12a).

Further, the perceptual quality of the synthesized voice signal (12a) is enhanced by the adaptive spectrum postfilter (13), and the voice signal is output from the output terminal (16).

When the VOX mode information is a silent section, background noise is generated with each index in the encoded data following the postamble as an initial value.

[0018]

In this conventional speech coding apparatus, a unique word is transmitted when uplink speech from a mobile station is silent, and transmission of a coding index to a base station is stopped. However, if the silent state continues, the coding index is periodically transmitted to the base station as information on ambient noise.

The speech coding apparatus of the base station generates background noise based on the information on the ambient noise.

Accordingly, there is a problem that the background noise that changes periodically becomes a sound that is very uncomfortable, unlike a sound obtained by encoding and decoding ordinary noise.

SUMMARY OF THE INVENTION The present invention has been made in view of such problems, and an object of the present invention is to provide a voice coder used in a digital mobile communication type automobile telephone system, and a mobile station VOX (or VOX). VAD: Voice Acti
It is an object of the present invention to provide a speech coding apparatus that reduces a sense of incongruity with respect to a speech output of a decoder caused by a periodic temperature control change of background noise transmitted upon detection of silence in vity detection processing.

[0022]

In order to achieve the above object, a speech coding apparatus according to the present invention receives coded data, and receives a power index, an adaptive code index, a noise code index, a gain code index, LSP (Li
ne Spectrum Pair) A data receiving unit that separates and outputs an index, and outputs the input encoded data, and receives encoded data output from the data receiving unit,
A VOX unique word detector that outputs OX mode information, a power code decoder that receives the power index and the VOX mode information and outputs quantized power, a noise code index and the VOX mode information , A random code decoder that outputs a random code vector signal, and the LSP index and the VOX mode information, and an LPC (Linear Predictive Coefficien).
t) and an LSP code decoder that outputs a prediction residual gain, and inputs the gain code index, the VOX mode information, the quantized power, and the prediction residual gain, and calculates an adaptive code gain coefficient and a noise code gain coefficient. An output gain code decoder;
An adaptive code signal decoder that receives the adaptive code index, the VOX mode information, and an excitation signal and outputs an adaptive code vector signal; a signal obtained by multiplying the adaptive code vector signal by the adaptive code gain coefficient; An adder that adds a signal obtained by multiplying the vector signal by the noise code gain and outputs the excitation signal;
A speech filter that receives a PC and outputs a synthesized speech signal; and a post filter that inputs the synthesized speech signal and the LPC and outputs a speech signal. A power controller for inputting VOX mode information and the audio signal, and performing power control of an audio signal output from the post filter when the VOX mode information generates background noise in a silent section. The present invention provides an audio encoding device that performs

In the speech coding apparatus of the present invention, preferably, the quantized power, the VOX mode information and the speech signal are inputted, and the VOX mode information is used when the background power of the silent section is generated. It is characterized by including a power controller that averages a long period of time and controls the power of the audio signal output from the post-filter so that the averaged power is obtained over a long period of time.

In the speech encoding apparatus of the present invention, preferably, the synthesized speech signal, the LPC and the VO
A post filter for inputting X-mode information and controlling so that filtering is not performed when the VOX mode information generates background noise in a silent section is provided.

In the speech coding apparatus of the present invention, it is preferable that the adaptive code index, the VOX mode information, and the excitation signal are input, and the adaptive code input when the VOX mode generates background noise in a silent section. An adaptive code signal decoder that switches an index to a fixed index and outputs an adaptive code vector signal is provided.

[0026]

According to the present invention, state control is performed in a sound section and a silent section based on VOX mode (VOX MODE) information output from a VOX unique word detector, thereby causing a periodic tone change of background noise. Discomfort can be reduced.

According to a preferred aspect of the present invention, when the VOX mode information indicates a silent state, the frequency enhancement of background noise is performed in an adaptive spectral postfilter used to enhance the auditory quality of a synthesized speech signal. Curb,
By controlling the power value in the silent section by averaging for a long time, it is possible to suppress the sudden level fluctuation of the background noise from occurring, and to generate an adaptive code vector signal with a fixed index in the adaptive code decoder. In addition, the periodic change of the background noise can be suppressed.

[0028]

Next, the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the speech encoding apparatus according to the present invention.

In FIG. 1, 1 is an input terminal, 2 is a data receiving unit, 3 is a VOX unique word detector, 4 is a power code decoder, 5 is an LSP code decoder, 6 is a gain code decoder, 7 is noise. 8 is an adaptive code signal decoder, 9 and 10 are multipliers, 11 is an adder, 12 is a synthesis filter, 13 is a post filter (adaptive spectrum post filter), 14 is a power controller, and 15 is a power controller. A multiplier 16 is an output terminal.

The operation will be described with reference to FIG. The data receiving unit (2) receives encoded data from the input terminal (1), and outputs a power index (2a), an adaptive code index (2b), a noise code index (2c), a gain code index (2d), and an LSP ( (Line Spectrum Pair) index (2e) and output separately.

Further, the data receiving section (2) outputs the inputted encoded data to the VOX unique word detector (3) (2f).

The VOX unique word detector (3) detects the postamble and the preamble output from the mobile station, and VOX MOD indicating the state of a sound section or a silent section.
E information (referred to as "VOX mode information") is output.

As VOX mode information, for example, the state at the time of detecting a preamble is set to state “−1”, the state at the time of detecting a postamble is set to state “1”, the state at the time of receiving an encoded index following the postamble is set to state “2”, and thereafter the preamble is set. The state until reception is set to state “3”, and the state “0” is output when the encoded index is received after the detection of the preamble. Among them, the state “2” and the state “3” correspond to generation of background noise in a silent section.

The power code decoder (4) outputs a quantized power value (4a) from a power codebook (not shown) based on the power index (2a).

The LSP code decoder (5) converts an LSP codebook (not shown) from the LSP codebook (not shown) based on the LSP index (2e).
The PC (5a) and the prediction residual gain (5b) are output.

The gain code decoder (6) receives the quantized power value (4a), the prediction residual gain (5b) and the gain code index (2d), and obtains a gain code based on the gain code index (2d). From the book (not shown), the adaptive code gain coefficient (6
a) and the noise code gain coefficient (6b).

The noise code signal decoder (7) generates a noise code vector signal (7) based on the noise code index (2c).
Output a).

The adaptive code signal decoder (8) uses the adaptive code vector signal (8) based on the adaptive code index (2b).
Output a). The adaptive codebook used in the adaptive code signal decoder (8) is updated by the excitation signal (11a),
A new adaptive codebook is generated.

The multiplier (9) generates a signal obtained by multiplying the adaptive code vector signal (8a) by the adaptive code gain coefficient (6a), and the multiplier (10) converts the noise code vector signal (7a) into a noise code signal (7a). A signal multiplied by the gain coefficient (6b) is generated, an output signal of the multiplier (9, 10) is input to the adder 11, and an excitation signal (11a) is output.

The synthesis filter 12 receives the LPC (5a) and the excitation signal (11a) and outputs a synthesized voice signal (12a). Furthermore, an adaptive spectrum post filter (13)
Increases the auditory quality of the synthesized speech signal.

Adaptive spectrum post filter (13)
Receives the synthesized speech signal (12a), LPC (5a), and VOX mode information, does not perform filtering when the VOX mode information is in the background noise in a silent section, and performs filtering when the VOX mode information is in a sound section. Control.
With this control, the post filter (13) suppresses frequency emphasis when generating background noise.

When the VOX mode information is a silent section, background noise is generated with each index in the encoded data following the postamble as an initial value.

The adaptive code signal decoder (8) receives the adaptive code index (2b) and VOX mode information,
When the OX mode generates background noise in a silent section, the input adaptive code index (2b) is switched to a fixed index (for example, an index not using an adaptive codebook), and an adaptive code vector signal (8a) is output.

The power controller (14) has a quantization power value (4
a), the VOX mode information, and the audio signal (13a) output from the post-filter 13, and calculates the power value of the audio signal. If the VOX mode information is in a silent section, the quantization power becomes longer. The average is taken, and the level of the audio signal is adjusted via the multiplier (15) so that the power value of the audio signal becomes the averaged power for a long time, thereby suppressing a sudden level fluctuation of the background noise. .

As described above, according to the present embodiment, by performing state control in a sound section and a silent section based on the VOX mode information output from the VOX unique word detector (3), the periodicity of the background noise is reduced. Discomfort caused by a change in tone can be reduced.

Although the present invention has been described with reference to the above embodiments, the present invention is not limited only to the above embodiments.
Various modes according to the principle of the present invention are included.

[0048]

As described above, according to the present invention, V
By performing state control in a sound section and a silent section based on the VOX mode information output from the OX unique word detector, an unnatural feeling caused by a periodic tone change of background noise can be reduced.

More specifically, according to a preferred embodiment of the present invention (claims 2 to 4), when the VOX mode information indicates a silence state, the adaptation used to enhance the auditory quality of the synthesized speech signal. In the spectral postfilter,
By suppressing the frequency emphasis of the background noise and controlling the power value in a silent section by averaging for a long time, it is possible to suppress the sudden level fluctuation of the background noise from occurring.In the adaptive code decoder, a fixed index is used. Generating an adaptive code vector signal has the effect of suppressing periodic changes in background noise.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an embodiment of a speech encoding device according to the present invention.

FIG. 2 is a block diagram illustrating a configuration of a conventional speech encoding device.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Input terminal 2 Data reception part 3 VOX unique word detector 4 Power code decoder 5 LSP code decoder 6 Gain code decoder 7 Noise code signal decoder 8 Adaptive code signal decoder 9, 10 Multiplier 11 Adder 12 Synthesis Filter 13 Post filter (adaptive spectrum post filter) 14 Power controller 15 Multiplier 16 Output terminal

Claims (4)

(57) [Claims] A coded data is inputted, and a power index, an adaptive code index, a noise code index, a gain code index, an LSP (Line Spectrum) are input.
A data receiving unit that separates and outputs a Pair index and outputs the input encoded data; a VOX unique word that receives the encoded data output from the data receiving unit and outputs VOX mode (VOXMODE) information A detector, a power code decoder that inputs the power index and the VOX mode information and outputs quantized power, and a noise code that inputs the noise code index and the VOX mode information and outputs a noise code vector signal A decoder, an LSP code decoder that inputs the LSP index and the VOX mode information, and outputs an LPC (Linear Predictive Coefficient) and a prediction residual gain; and the gain code index, the VOX mode information, and the quantization power. And the prediction residual gain, and a gain code for outputting an adaptive code gain coefficient and a noise code gain coefficient. An adaptive code signal decoder that inputs the adaptive code index, the VOX mode information, and an excitation signal and outputs an adaptive code vector signal; and a signal obtained by multiplying the adaptive code vector signal by the adaptive code gain coefficient. An adder that adds a signal obtained by multiplying the noise code vector signal to the noise code gain and outputs the excitation signal; a synthesis filter that receives the excitation signal and the LPC and outputs a synthesized speech signal; And a post-filter that receives the synthesized speech signal and the LPC and outputs a speech signal. A speech encoding apparatus comprising: a receiving unit that receives the quantized power, the VOX mode information, and the speech signal; A power controller for controlling the power of the audio signal output from the post filter when generating background noise in a silent section in which mode information is provided. Speech coding apparatus according to claim. 2. The apparatus according to claim 1, wherein said quantized power, said VOX mode information, and said audio signal are inputted, and said VOX mode information is averaged for a long time during generation of background noise in a silent section. 2. The speech coding apparatus according to claim 1, further comprising: a power controller that controls power of a speech signal output from the post-filter so as to obtain power averaged for a long time. 3. A post-filter for inputting the synthesized speech signal, the LPC, and the VOX mode information, and controlling the VOX mode information so as not to perform filtering when background noise in a silent section is generated. 3. The speech encoding device according to claim 2, comprising: 4. The adaptive code index and the VOX
Inputting mode information and the excitation signal;
4. The speech code according to claim 3, further comprising an adaptive code signal decoder that switches an input adaptive code index to a fixed index and outputs an adaptive code vector signal when a background noise in a silent section is generated. Device.