JP2582947Y2 - Audio signal 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 an audio signal encoding apparatus used for digital recording such as an answering machine and an IC recorder.

[0002]

2. Description of the Related Art When digitally recording and recording audio for each block, a method of recording only the number of successive blocks without silence in general is often used in order to save storage capacity without recording silent portions as they are. Used. However, in the conventional method, silence is considered to be a signal near zero, and speech level or power (root mean square) is used to distinguish between speech and silence. If it is included, there is a risk that it is erroneously determined to be sound. In terms of this DC component, it is a frequent phenomenon that a DC component is added to an audio signal in an analog circuit such as an AD converter, and as a result of being regarded as sound by the DC component as a whole, However, there is a drawback in that the number of parts is reduced and the compression ratio of audio data is not improved.

[0003]

As described above, the conventional speech signal encoding apparatus has a problem that the DC component is regarded as a sound even if there is no sound. Was a factor.

[0004] In view of the above problems of the prior art, the present invention is intended to accurately determine the presence or absence of sound and to improve the data compression rate in speech signal encoding.

[0005]

According to the present invention, there is provided an encoding section for encoding an audio signal divided into a plurality of blocks for each block, a processing section for calculating a root mean square of each block, A processing unit for calculating the square of the average value of the block; a determination unit for determining whether the block is sound or silence based on a difference between the square mean and the square of the average value; and a code for each block in the case of sound. The memory comprises a memory for storing the number of silent blocks in the case of silence on the other hand while storing the converted audio signal.

[0006]

According to the above configuration, it is possible to execute audio signal encoding with an improved compression ratio for an audio signal containing a DC component.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS First, the principle of the present invention will be described. Assuming that the audio signal input to the audio encoding device is Xi, the difference between the root mean square and the root mean square, that is, the variance, is
It is represented as

[0008]

(Equation 1)

(Xi−ＸXi)
/ N) can be regarded as an amount obtained by removing the DC component ΣXi / N from the input signal Xi, that is, the variance can be considered as an amount of change obtained by removing the DC component from the input signal.

In the present invention, a variance is used instead of a root mean square for the determination of sound / no sound, so that a signal having only a DC component is determined to be silent, and only when the signal is changing is determined to be sound. Therefore, silence suppression works effectively even for a signal in which a DC component is biased.

FIG. 1 shows a block circuit of an embodiment of an encoding / decoding device employing the speech signal encoding device of the present invention. Referring to FIG. 1, reference numeral 1 denotes an A / D converter for converting an input analog audio signal into digital data; 2, a buffer for storing one block of the digital signal converted by the A / D converter 1; The mean square calculating unit 4 calculates the mean of the square of the data of one block stored in the buffer 2, and the mean is the mean of calculating the square of the mean of the data of one block stored in the buffer 2. The value square calculator 5 has an adder 51, calculates the difference between the outputs of the mean square calculator 3 and the mean square calculator 4, and determines whether the block is sound or no sound from the result. The voiced / silent determining unit 6 encodes one block at a time when it is determined that there is voice based on the determination result of the voiced / silent determining unit 5 and outputs the result. The encoded data Encoding unit for force 7 is a memory which stores a sound and silence of encoded data output from the coded unit 6, constituting up to this point as a coding device.

On the other hand, the decoding device takes in the coded data in the memory 7 and judges whether the data is sound or no sound. If there is sound, the decoding device decodes the coded data into a digital audio signal and outputs it. In this case, a decoding unit 8 for outputting a silence period indicated by encoded data indicating silence, a silence part insertion unit 9 for inserting the silence period into the decoded digital audio signal, The average value calculating unit 10 outputs the average value to the silent unit insertion unit 9 when a silent block is output by calculating the average value of the digital audio signal of the block, and the digital audio signal that has passed through the silent unit insertion unit 9 is output. It comprises a D / A converter 11 for converting to analog audio.

In the configuration of this embodiment, the sound / non-speech judging unit 5 uses the amount of change excluding the DC component of the audio signal as the basis for judgment as described in the above equation (1). If not, it is determined that there is no sound, and the result is sent to the encoding unit 6.

The decoding side receives voice code data and silence code data from the memory 7 and outputs a decoded voice signal in the case of voice, but outputs zero in the case of silence. Since the boundary between the sound block and the silence block is not well connected, the average value of the sound signal of the sound block up to the previous time stored in the average value calculation unit 10 is output, and the sound signal is properly processed at the boundary. The device is connected. For this reason, the average value calculation unit 10 calculates the average value of the audio signal in the case of a sound, and stores the average value until the next sound block enters the average value calculation unit 10.

That is, while the sound block is continuously output from the decoding unit 8, nothing is performed in the silent unit insertion unit 9, and is output to the D / A converter 11 as it is. On the other hand, when data indicating a silent period is output from the decoding unit 8, the average value calculating unit 10 does not update the average value, and outputs the previous average value to the silent unit inserting unit 9. The silent section insertion section 9 calculates the average value from the average value calculating section for a period corresponding to the silent period stored in the memory by D /
Output to the A converter 11.

[0016]

[Effects of the Invention] As described above, according to the present invention, silence cancellation works effectively by using the variance of the audio signal to determine the presence or absence of voice even for voices containing DC components. By using not the zero output but the average value of the previous sound block, the effect of reproducing the sound with a reasonable value close to the original sound is produced.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating the concept of a speech signal encoding apparatus according to the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 A / D converter 2 Buffer 3 Mean-square calculating part 4 Mean-square calculating part 5 Speech / non-speech judging part 6 Encoding part 7 Memory

Claims (1)

(57) [Scope of request for utility model registration] An encoding unit for encoding an audio signal divided into a plurality of blocks for each block, a processing unit for calculating a mean square of each block, and obtaining a square of an average value of each block A processing unit, a determination unit that determines whether the block is voiced or silent based on a difference between the root mean square and the square of the average value, and stores a voice signal coded for each block if the block is voiced. On the other hand, in the case of silence, a speech signal encoding device comprising a memory for storing the number of silence blocks.