JP2001100796A - Audio signal encoding device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress degradation of the tone quality by optimizing bit assignment in accordance with an aural sensitivity to reduce quantization distortion in an audio signal sub-band compression encoding device. SOLUTION: An audio signal inputted to a sub-band division device 101 has the frequency divided to obtain a sub-band signal. The signal level is obtained by a level calculation device 402. An assigned bit calculation device 104 corrects the signal level by the aural sensitivity described in an equisignal curve table 103 and determines the number of quantization bits to be assigned to each sub-band in accordance with the level after correction. A quantization device 105 normalizes each sub-band signal by the signal level to perform quantization. Level information on individual sub-bands and quantized sub-band signals are shaped and coupled by a bit stream generator 106 and are outputted as a bit stream.

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, and more particularly, to an encoding apparatus for dividing an audio signal into sub-bands, compressing the divided sub-bands, and transmitting the compressed data.

[0002]

2. Description of the Related Art As a conventional audio signal subband compression encoding apparatus, as shown in FIG.
There is an audio signal encoding device disclosed in Japanese Unexamined Patent Publication (Kokai) No. H11-15095.
This is a lossless encoding device that compresses and encodes an audio signal with high efficiency and completely decodes it. In FIG. 7, a sub-band dividing device 201 divides an input audio signal into predetermined sub-bands. Level calculation device 202
Is to detect the maximum value of the amplitude of the signal of each subband. The allocation bit calculation device 203 is a device that determines the number of bits for quantizing the signal of each subband based on the output of the level calculation device. The quantization device 204 is a device that quantizes the output signal of the subband division device with the number of bits determined by the allocation bit calculation device. The bit stream generation device 205 is a device that shapes the output of the level calculation device and the output of the quantization device to generate a bit stream. In this device, the frequency domain information generating section divides the digital audio signal into bands, and then, based on the energy values obtained for the individual bands, quantizes the sample values quantized by bit allocation according to the amount of information in the frequency domain. Encode as information. In the time domain information generation unit, the quantized sample value supplied from the frequency domain information generation unit is inversely quantized. In order to re-convert this into a time domain signal, the residual signal between the signal after band synthesis and the original signal is encoded as time domain information. By multiplexing the two in a multiplexing unit, lossless compression encoding is performed.

[0003]

However, in the above-mentioned conventional audio signal encoding apparatus, only the amplitude of each sub-band is used to determine the number of quantization bits, so that the auditory sensitivity is high. When the amplitude of the band is small, there is a problem that the bit allocation is insufficient and quantization distortion is perceived. This becomes remarkable when the bit rate of the transmission system is low, and there is a problem that sound quality is deteriorated.

An object of the present invention is to solve the above-mentioned conventional problems and to provide an audio signal encoding apparatus with less deterioration of sound quality.

[0005]

According to the present invention, there is provided an audio signal encoding apparatus which divides an audio signal into sub-bands and compresses the sub-band. Isosensitivity curve table corresponding to the level, level calculating means for calculating the level of the subband signal, comparing means for comparing the signal level calculated by the level calculating means with the value of the isosensitivity curve table, and a comparison result by the comparing means Means for obtaining an optimum bit allocation amount according to the above. With this configuration, an appropriate number of quantization bits can be obtained even when the amplitude in a band where the auditory sensitivity is high is small.

For high frequencies, a means is provided for setting the sensitivity coefficient described in the isosensitivity curve table to 0 when the input signal level is lower than a certain threshold value. With such a configuration, generally, the number of quantization bits in the high frequency range where the audio component is small is suppressed, and bits are assigned to the middle and low frequency range where the audio component is concentrated, so that when the bit rate of the transmission system is low. In addition, deterioration of sound quality can be suppressed.

For high frequencies, a means for making the sensitivity coefficient described in the isosensitivity curve table non-linear is provided.
With this configuration, it is possible to reduce a change in sound quality when the level of the high frequency rises and falls near the threshold.

Further, there are provided means for determining whether the input voice signal is a voiced sound or unvoiced sound, and means for reducing the high-frequency bit allocation when the input voice signal is a voiced sound. With this configuration, even when the input is limited to speech, higher-quality sound encoding can be performed.

Further, the audio signal decoding apparatus is provided with an isosensitivity curve table corresponding to an expected output level of the audio signal after decoding, a comparing means for comparing scale factor information with a value of the isosensitivity curve table, and a comparison result. Means for calculating the bit allocation amount of each subband according to With this configuration, a high-quality encoded bit stream can be decoded with high quality.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below in detail with reference to FIGS.

(First Embodiment) In a first embodiment of the present invention, a value of an isosensitivity curve table corresponding to an expected output level of a decoded audio signal, an output of a level calculating device, Is an audio signal encoding device that obtains an optimal bit allocation amount according to the result of comparison.

FIG. 1 is a functional block diagram of an audio signal encoding device according to a first embodiment of the present invention.
In FIG. 1, a sub-band dividing device 101 divides an input audio signal for each frequency band to generate a sub-band signal. The level calculator 102 calculates the level of each subband signal. The isosensitivity curve table 103 describes the auditory sensitivity at the expected output level of the decoded audio signal for each frequency band. Assigned bit calculation device
Numeral 104 is for correcting the signal level of each sub-band with auditory sensitivity and determining the number of quantization bits to be allocated to each sub-band according to the corrected level. Quantizer 105
Normalizes each subband signal at the signal level,
The quantization is performed with a predetermined quantization bit number. The bit stream generation device 106 shapes the level information of each sub-band and the quantized sub-band signal to form a bit stream.

The operation of the audio signal encoding apparatus according to the first embodiment of the present invention configured as described above will be described. First, the audio signal is divided into sub-band splitters
The signal is input to 101, is divided for each predetermined frequency band, and becomes a sub-band signal.
Entered in 05. The level calculation device 402 obtains a signal level from the maximum value of the amplitude of the input subband signal. In the allocated bit calculation device 104, the isosensitivity curve table 1
The signal level is corrected with the auditory sensitivity described in 03, and the number of quantization bits allocated to each subband is determined according to the corrected level. In addition, the quantization device 105 normalizes each subband signal according to the signal level, and performs quantization using the number of quantization bits determined by the allocation bit calculation device.
The level information of each sub-band and the quantized sub-band signal are shaped and combined in a bit stream generation device 106, and output as a bit stream.

FIG. 2 is a diagram showing the relationship between the isosensitivity curve table and the input signal. The horizontal axis in FIG. 2 indicates the frequency, and the vertical axis indicates the amplitude level. FIG. 2 plots an auditory sensitivity curve when the expected output level of the decoded audio signal is 70 dBSPL, and an example of the input signal. At this time, the difference between the auditory sensitivity curve and the input signal is the corrected signal level. As shown by the bold line in FIG. 2, even when the input signal levels are equal, the corrected signal levels are different if the auditory sensitivities are different.

As described above, according to the first embodiment of the present invention, the audio signal encoding apparatus is provided with an iso-sensitivity curve table corresponding to the expected output level of the audio signal after decoding and the level calculation. Since the optimum bit allocation amount is determined according to the result of comparison with the output of the apparatus, even if the amplitude in a band where the auditory sensitivity is high is small, an appropriate number of quantization bits is determined to obtain a code of high sound quality. Can be performed.

(Second Embodiment) In a second embodiment of the present invention, when the input signal level is lower than a certain threshold value in the high frequency range, the sensitivity described in the isosensitivity curve table is used. This is an audio signal encoding device that sets the coefficient to 0.

FIG. 3 is a functional block diagram of the audio signal encoding device according to the second embodiment of the present invention.
In FIG. 3, a sub-band dividing device 401 divides an input audio signal for each frequency band to generate a sub-band signal. The level calculation device 402 calculates the level of each subband signal. When the input signal level is smaller than a certain threshold value for the high frequency sub-band, the band limiting device 403 sets the level of the sub-band to 0. The isosensitivity curve table 404 describes the auditory sensitivity at the expected output level of the decoded audio signal for each frequency band. The allocation bit calculation device 405 corrects the signal level of each sub-band based on auditory sensitivity, and determines the number of quantization bits to be allocated to each sub-band according to the corrected level. The quantization device 406 normalizes each subband signal at a signal level, and then performs quantization with a predetermined number of quantization bits. The bit stream generation device 407 shapes the level information of each sub-band and the quantized sub-band signal to form a bit stream.

FIG. 4 is a diagram for explaining band limiting processing of the audio signal encoding device according to the second embodiment of the present invention. The horizontal axis in FIG. 4 is the input level, and the vertical axis is the output level. A broken line 501 is a line graph showing the band limiting processing characteristic according to the second embodiment.

The operation of the audio signal encoding apparatus according to the second embodiment of the present invention configured as described above will be described. First, the audio signal is divided into sub-band splitters
The signal is input to 401, is divided for each predetermined frequency band, and becomes a sub-band signal.
Entered in 06. The level calculation device 402 obtains a signal level from the maximum value of the amplitude of the input subband signal.

The signal level is input to the band limiting device 403. If the input signal level is smaller than a predetermined threshold value for a sub-band higher than a predetermined frequency, the level of the sub-band is set to 0. As shown by the broken line 501 in FIG.
If the input signal level is less than 20 dB, the signal level in that subband is forced to zero.

The allocated bit calculation device 405 corrects the signal level based on the auditory sensitivity described in the isosensitivity curve table 404, and determines the number of quantization bits allocated to each subband according to the corrected level. In the quantization device 406, each sub-band signal is normalized by a signal level,
The quantization is performed using the number of quantization bits determined by the allocation bit calculation device. The level information of each sub-band and the quantized sub-band signal are shaped and combined in a bit stream generation device 407, and output as a bit stream.

As described above, in the second embodiment of the present invention, the audio signal encoding device
When the input signal level is lower than a certain threshold value, the sensitivity coefficient described in the isosensitivity curve table is set to 0, so that in general, the number of high-frequency quantization bits with few audio components is suppressed. By allocating bits to the middle and low frequencies where audio components are concentrated, it is possible to suppress deterioration of sound quality even when the bit rate of the transmission system is low.

(Third Embodiment) A third embodiment of the present invention is an audio signal encoding apparatus for making the sensitivity coefficient described in the isosensitivity curve table non-linear for a high frequency band.

The configuration of the audio signal encoding device according to the third embodiment of the present invention is the same as that of the second embodiment. In FIG. 3, when the input signal level is lower than the first threshold value for the high-band sub-band, the band limiting device 403 sets the level of the sub-band to 0, and the input signal level becomes higher than the second threshold value. If it is larger, the level of the subband is equal to the input level, and the level coefficient changes non-linearly from the first threshold to the second threshold.

In FIG. 4, a broken line 502 is a non-linear graph showing the band limiting processing characteristic in the third embodiment.

The operation of the audio signal encoding apparatus according to the third embodiment of the present invention configured as described above will be described. First, the audio signal is divided into sub-band splitters
The signal is input to 401, is divided for each predetermined frequency band, and becomes a sub-band signal.
Entered in 06. The level calculation device 402 obtains a signal level from the maximum value of the amplitude of the input subband signal.

The signal level is input to the band limiting device 403. If the input signal level of the sub-band higher than the predetermined frequency is smaller than the first threshold, the level of the sub-band is set to 0 and the second sub-band is set to the second level. When the input signal level is higher than the threshold value, the level of the subband is equal to the input level, and the level coefficient changes nonlinearly between the first threshold value and the second threshold value. Curve 50 in FIG.
As shown by 2, when the input signal level is smaller than 20 dB, the signal level in the subband is forcibly set to zero. If the input signal level is greater than 40 dB, the signal level in that subband is equal to the input level. When the input signal level is between 20 dB and 40 dB, the level is continuously changed.

The allocated bit calculation device 405 corrects the signal level based on the auditory sensitivity described in the isosensitivity curve table 404, and determines the number of quantization bits allocated to each subband according to the corrected level. Further, the quantization device 406 normalizes each subband signal according to the signal level, and performs quantization using the number of quantization bits determined by the allocation bit calculation device. The level information of each sub-band and the quantized sub-band signal are shaped and combined in a bit stream generation device 407 and output as a bit stream.

As described above, in the third embodiment of the present invention, the audio signal
Since the sensitivity coefficient described in the isosensitivity curve table is configured to be non-linear, it is possible to reduce a change in sound quality when the level of the high frequency rises and falls near the threshold.

(Fourth Embodiment) In a fourth embodiment of the present invention, it is determined whether an input voice signal is a voiced sound or an unvoiced sound. It is an audio signal encoding device for reducing.

FIG. 5 is a functional block diagram of an audio signal encoding device according to the fourth embodiment of the present invention.
In FIG. 5, a sub-band dividing device 601 divides an input audio signal for each frequency band to generate a sub-band signal. The level calculation device 602 calculates the level of each subband signal. The voiced / unvoiced determination device 603 determines whether or not the input signal is a voiced sound. Bandwidth limiter 60
4 indicates that if the input signal level is lower than a certain threshold value for the high-frequency sub-band,
It is assumed that. The isosensitivity curve table 605 describes the auditory sensitivity at the expected output level of the decoded audio signal for each frequency band. The allocation bit calculating device 606 corrects the signal level of each subband with the auditory sensitivity, and determines the number of quantization bits to be allocated to each subband according to the corrected level.
The quantization device 607 normalizes each subband signal at a signal level, and then performs quantization with a predetermined number of quantization bits. The bit stream generation device 608 shapes the level information of each sub-band and the quantized sub-band signal to form a bit stream.

The operation of the audio signal coding apparatus according to the fourth embodiment of the present invention configured as described above will be described. First, the audio signal is divided into sub-band splitters
601 and is divided into predetermined frequency bands to become sub-band signals.
07 Entered. The level calculation device 602 obtains a signal level from the maximum value of the amplitude of the input subband signal.
The signal level is determined by the voiced / unvoiced determination device 603 and the band limiter 6
Entered in 04. The voiced / unvoiced determination device 603 extracts the feature amount of the input signal level and determines whether the input signal is a voiced sound, an unvoiced sound, or a non-voiced sound. If it is determined that the sound is voiced, the band limiting device 604
Then, components higher than the band corresponding to the voice are suppressed. The allocated bit calculation device 606 corrects the signal level based on the auditory sensitivity described in the isosensitivity curve table 605, and determines the number of quantization bits allocated to each subband according to the corrected level. In addition, the quantization device 607 normalizes each subband signal according to the signal level, and performs quantization using the number of quantization bits determined by the allocation bit calculation device. The level information of each sub-band and the quantized sub-band signal are shaped and combined in the bit stream generator 608,
Output as a bit stream.

As described above, in the fourth embodiment of the present invention, the audio signal encoding apparatus determines whether the input speech signal is a voiced sound or an unvoiced sound. , The bits are allocated to the middle and low frequencies where voice components of voiced sounds are concentrated, so that deterioration of sound quality can be suppressed even when the bit rate of the transmission system is low.

(Fifth Embodiment) In a fifth embodiment of the present invention, a value of an isosensitivity curve table corresponding to an expected output level of a decoded audio signal is compared with scale factor information. This is an audio signal decoding device that obtains a bit allocation amount for each subband according to a result of the above operation.

FIG. 6 is a functional block diagram of an audio signal decoding device according to the fifth embodiment of the present invention. In FIG. 6, a bit stream decomposing device 701 decomposes an input bit stream into level information and sample information. The isosensitivity curve table 702 describes auditory sensitivity at an expected output level of the decoded audio signal for each frequency band. The allocated bit calculating device 703 corrects the level information of each subband with auditory sensitivity, and estimates the number of quantization bits allocated to each subband according to the corrected level. The inverse quantization device 704 performs inverse quantization with a predetermined number of quantization bits, and then normalizes each subband signal at a signal level. The subband synthesizing device 706 synthesizes the original audio signal from the subband signal.

The operation of the audio signal decoding apparatus according to the fifth embodiment of the present invention configured as described above will be described. First, the bit stream is input to the bit stream decomposition device 701, decomposed into level information and sample information, the level information is input to the assigned bit calculation device 703 and the inverse quantization device 704, and the sample information is transmitted to the inverse quantization device 704. Is entered. The allocated bit calculation device 703 corrects the level information with the auditory sensitivity described in the isosensitivity curve table 702, and estimates the number of quantization bits allocated to each subband according to the corrected level. In addition, the inverse quantization device
In 704, after performing inverse quantization with the number of quantization bits estimated by the allocation bit calculation device, each subband signal is normalized at the signal level. The dequantized sub-band signal is input to the sub-band synthesizing device 706, and is synthesized with the original audio signal.

As described above, according to the fifth embodiment of the present invention, the audio signal decoding apparatus is provided with a function of the isosensitivity curve table corresponding to the expected output level of the decoded audio signal and the scale factor information. And the bit allocation amount of each sub-band is obtained according to the result of comparing the bit stream output encoded with high sound quality with
It can be decoded with high sound quality.

[0038]

As is apparent from the above description, according to the present invention, an audio signal encoding apparatus that divides an audio signal into sub-bands and compresses the divided sub-bands corresponds to an expected output level of the decoded audio signal. An isosensitivity curve table, a level calculating means for calculating the level of the sub-band signal, a comparing means for comparing the signal level calculated by the level calculating means with the value of the isosensitivity curve table, and an optimum according to the comparison result by the comparing means. Means for determining the optimal bit allocation amount, even when the amplitude in a band with high auditory sensitivity is small, the optimum number of quantization bits is determined according to the sound volume perceived by the listener, and high sound quality is obtained. Can be obtained.

For high frequencies, means for setting the sensitivity coefficient described in the isosensitivity curve table to 0 when the input signal level is lower than a certain threshold value is provided. Suppresses a small number of high-frequency quantization bits,
Allocate bits to the middle and low frequencies where audio components are concentrated,
Even when the bit rate of the transmission system is low, the effect that deterioration of sound quality can be suppressed can be obtained.

In the high frequency range, a means for making the sensitivity coefficient described in the isosensitivity curve table non-linear is provided, so that a change in sound quality when the level of the high frequency rises or falls near the threshold value is reduced. The effect that it can be obtained is obtained.

Further, since means for determining whether the input voice signal is a voiced sound or unvoiced sound and means for reducing the high-frequency bit allocation when the voice signal is voiced are provided, the input is limited to voice. In this case, it is possible to obtain an effect that encoding with higher sound quality can be performed.

Further, the audio signal decoding apparatus is provided with an isosensitivity curve table corresponding to an expected output level of the decoded audio signal, a comparing means for comparing the scale factor information with the value of the isosensitivity curve table, and a comparison result. And means for calculating the bit allocation amount of each subband according to the above. Therefore, it is possible to obtain an effect that a high-quality encoded bit stream can be decoded with high quality.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of an audio signal encoding device according to a first embodiment of the present invention;

FIG. 2 is a diagram showing a relationship between an isosensitivity curve and an input signal of the audio signal encoding device according to the first embodiment of the present invention;

FIG. 3 is a functional block diagram of an audio signal encoding device according to a second embodiment of the present invention;

FIG. 4 is a band limitation processing diagram of the audio signal encoding device according to the second and third embodiments of the present invention;

FIG. 5 is a functional block diagram of an audio signal encoding device according to a fourth embodiment of the present invention;

FIG. 6 is a functional block diagram of an audio signal decoding device according to a fifth embodiment of the present invention;

FIG. 7 is a functional block diagram of a conventional audio signal encoding device.

[Explanation of symbols]

 101 Subband splitter 102 Level calculator 103 Isosensitivity curve table 104 Assigned bit calculator 105 Quantizer 106 Bitstream generator 201 Subband splitter 202 Level calculator 203 Assigned bit calculator 204 Quantizer 205 Bitstream generation Apparatus 301 Isosensitivity curve 302 Level of input signal 303 Level after correction 401 Subband splitter 402 Level calculator 403 Band limiter 404 Isosensitivity curve table 405 Assigned bit calculator 406 Quantizer 407 Bitstream generator 501 Second Band limitation in the third embodiment 502 Band limitation in the third embodiment 601 Subband splitting device 602 Level calculating device 603 Voiced / unvoiced determining device 604 Band limiting device 605 Isosensitivity curve table 606 Assigned bit calculating device 607 Quantizer 608 Bitstream generator 701 Bitstream decomposer 703 Isosensitivity curve table 704 Assigned bit calculator 705 Dequantizer 706 Subband splitter

Claims (7)

[Claims] 1. An audio signal encoding apparatus which divides an audio signal into sub-bands and compresses the sub-bands, and calculates an iso-sensitivity curve table corresponding to an expected output level of a decoded audio signal and a sub-band signal level. Level calculating means, comparing means for comparing the signal level calculated by the level calculating means with the value of the isosensitivity curve table, and means for obtaining an optimum bit allocation amount according to the comparison result by the comparing means. An audio signal encoding device comprising: 2. A high frequency band, comprising: means for setting a sensitivity coefficient described in the isosensitivity curve table to 0 when an input signal level is lower than a predetermined threshold value. 2. The audio signal encoding device according to 1. 3. The audio signal encoding apparatus according to claim 2, further comprising means for making the sensitivity coefficient described in the isosensitivity curve table non-linear for a high frequency band. 4. The apparatus according to claim 1, further comprising means for determining whether the input voice signal is a voiced sound or unvoiced sound, and means for reducing high-frequency bit allocation when the input voice signal is a voiced sound. 3. The audio signal encoding device according to claim 1. 5. An isosensitivity curve table corresponding to an expected output level of an audio signal after decoding, comparison means for comparing scale factor information and values of the isosensitivity curve table, and each subband according to the comparison result. Means for calculating the bit allocation amount of the audio signal. 6. A digital wireless microphone system comprising: the audio signal encoding device according to claim 1; and the audio signal decoding device according to claim 5. 7. A speaker system comprising the digital wireless microphone according to claim 6. Description: