JPH08335882A - Encoder and decoder - Google Patents

Abstract

PURPOSE: To give a wide-band and high dynamic range by providing an encoder with an encoding means which transmits an input acoustic signal after encoding it by a sampling frequency which is a specific number or more times as high as the audible band and providing a decoder with a decoding means which takes out and reproduces a signal of a prescribed band. CONSTITUTION: The input acoustic signal is encoded by >=4-fold sampling frequency of the audible band, and the dynamic range of the audible band is extended by ΔΣ modulation if necessary. That is, a ΔΣ modulator 2 is provided which is operated by >=4-fold sampling frequency of the audible band and has >=8-bit output and has the shaping characteristic varied by mode control and consists of an adder 4, a local quantizer 5, etc., and an inputted digital signal is encoded by the ΔΣ modulator 2 and is transmitted together with the mode control ID. At the time of decoding, the cut-off frequency of a low pass filter is controlled in accordance with the characteristic of the ΔΣ modulator 2 based on the mode control ID by a digital filter.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a coding device and a decoding device for transmitting digital signals with high quality.

[0002]

2. Description of the Related Art In recent years, music reproduction using digital signals such as compact discs has been widely performed. However, compact discs are determined to have a sampling frequency of 44.1 kHz and data of 16 bits, and 22.
In principle, it is impossible to reproduce sound exceeding 05 kHz or obtain a dynamic range exceeding 98 dB. However, an actual music signal contains components exceeding 22.05 kHz, and there are many dynamic signals exceeding 100 dB. In particular, it is considered that α-waves in brain waves are more likely to be emitted with components having a frequency exceeding 20 kHz, and it has been pointed out that the super high frequency components in the reproduced signal are important.

Therefore, as shown in the Radio and Experiment Magazine, February 1995 issue, pages 100 to 101, published by Seibundo Shinkosha, LSB of 16-bit data is used, and 2 bits are used for this bit.
A method of recording music signal information of 2.05 kHz or higher by using ADPCM (method 1), an AIA publisher, Radio Technical Journal, April 1991, pages 147 to 15
As shown on page 0, a method has been proposed in which the quantization noise is driven to 15 kHz to 22.05 kHz by using noise shaping to improve the audible dynamic range (system 2).

[0004]

However, in the above-mentioned configuration, the reproduction band is widened in the system 1, but the dynamic range is reduced by 6 dB, and in the system 2, 15 kHz to 20 kHz is in the audible band for humans. However, there is a problem that the dynamic range in the high frequency range is significantly reduced.

The present invention solves the above problems, and an object of the present invention is to provide a coding device and a decoding device having a wide band and a high dynamic range.

[0006]

In order to achieve this object, a coding apparatus according to the present invention comprises a coding means for coding an input acoustic signal at a sampling frequency of four times or more of an audible band and transmitting it. The decoding device is provided with a decoding unit that extracts and reproduces a signal in a predetermined band including the audible band from the encoded acoustic signal, and the coding unit is a ΔΣ modulation that replaces the quantization noise outside the audible band. , Performing predetermined ΔΣ modulation based on a band identification signal given in advance, transmitting the band identification signal, and controlling the frequency band to be reproduced by the decoding means based on the transmitted band identification signal; Is amplified in advance by the emphasis means and transmitted together with the emphasis identification signal indicating the emphasis state, and the frequency of the signal reproduced by the decoding means is reproduced based on the emphasis identification signal. Controlling the numerical characteristics, the encoding means encodes and transmits the acoustic signal at a predetermined sampling frequency based on a sampling frequency identification signal given in advance, and the decoding means performs decoding based on the sampling frequency identification signal. It is also possible to do.

[0007]

The acoustic signal input as described above is coded at a sampling frequency which is four times or more the audible band, and the dynamic range of the audible band is expanded by ΔΣ modulation as necessary. ~ 20k
It is possible to obtain an acoustic signal with a high dynamic range without degrading the dynamic range at Hz, and because the sampling frequency is 4 times the audible band or more, it is possible to obtain a super high frequency range outside the audible band that is meaningful for α waves. An acoustic signal can also be generated.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a coding device according to a first embodiment of the present invention. To explain this figure, the oversampling filter 1 has an acoustic input (here, the sampling frequency is 96 kHz, and the word length is 24
Bit) and 2)
Converts to a 4-bit, 192 kHz digital signal. ΔΣ
The modulator 2 bit-compresses the input 24-bit, 192 kHz signal into a 16-bit, 192 kHz digital signal based on the mode control signal, outputs the converted acoustic signal as an encoder output, and transmits it to a decoding device.

As the delta-sigma modulator 2, for example, one having a configuration as shown in the drawing is effective. That is, the local quantizer 5 rounds the input 24-bit data into 16 bits (requantizes) and outputs the rounded data. The subtracter 6 takes out the quantization error Vq generated by the local quantizer 5 and supplies it to the feedback circuits 7 and 8 having the transfer functions H1 (z) and H2 (z). The selector 9 selects and outputs one of the outputs of the feedback circuits 7 and 8 based on the mode control signal, and the adder 4 inputs the output from the oversampling filter 1 24.
The bit and the signal of 192 kHz are added and sent to the local quantizer 5. Here, the transfer function H1 of the feedback circuits 7 and 8
(z) and H2 (z), for example, H1 (z) = z ⁻¹ , H2
If something like (z) = z ⁻² −z ⁻¹ is used, signals having dynamic ranges of 116 dB and 128 dB in the audible band can be obtained as encoder outputs.

As described above, by changing the transfer function of the ΔΣ modulator 2 according to the word length of the sound input and the sampling frequency, the dynamic range in the audible band can be expanded without narrowing the reproduction band. in this case,
For example, the dynamic range of 70 kHz or more is reduced, but originally there is almost no large-amplitude signal in this band, and there is practically no problem. If a signal with a large amplitude is absolutely necessary in the ultra-high range, the transfer function of the feedback circuit 7 may be set to H1 (z) = 0. Even in this case, a dynamic range of 104 dB can be obtained in the audible band, and a sound quality superior to that of the current compact disc can be obtained.

FIG. 2 is a block diagram showing a decoding device of the coding device and the decoding device according to the first embodiment of the present invention.
In the decoding device 3, the converted acoustic signal transmitted from the encoder output is extracted, and 16-bit, 192 kHz
Is input to the oversampling filter 10 as a decoder input. In the oversampling filter 10, this signal is twice oversampled, converted into 384 kHz, 24 bits, and D /
It is input to the A converter 12. The D / A converter 12 converts this signal into an analog signal, and then the low pass filter 1
The aliasing distortion is removed by 3 and is output as an analog output.

Considering the analog output output from the decoding device 3, the decoder input has a constant word length and a constant frequency regardless of the word length and sampling frequency of the acoustic input input to the encoding device. As long as the oversampling filter 10 is capable of sufficiently dropping the component of 96 kHz or more, which is the Nyquist frequency, even if the characteristic is fixed, reproduction can be performed without substantially lowering the quality of the acoustic input given to the coding device. You can It is also possible to directly D / A convert the decoder input in the D / A converter 12 without using an oversampling filter. In this case, the low-pass filter 13 may be one that sufficiently sharply drops the components of the Nyquist frequency of 96 kHz or higher. Even in this case, the characteristics may be fixed regardless of the sampling frequency of the acoustic input and the word length.

As described above, the transmission between the coding device and the decoding device is performed by the digital signal exceeding 4 times the audible band, so that even if the characteristic of the transfer system in the decoding device is constant, it is input to the coding device. It is possible to exhibit the performance corresponding to the change of the sound input band and the dynamic range.
By the way, generally, an error occurs in a signal transmitted on a transmission line. Usually, for example, a compact disc or the like is provided with an error correction code to cope with this, so that even if some error occurs, it can be surely repaired. However, if the error occurrence exceeds a certain amount, it cannot be repaired. In this case, for example, the previous value hold or the linear interpolation processing by the previous and next values is performed. The generation of abnormal noise when the error cannot be corrected can be treated as something that cannot be helped to some extent, but it is important for the user that at least the hearing is not damaged and the device such as the speaker is not damaged. Taking these reasons into consideration, it is desirable that the level of abnormal noise is -40 dB or less of the sound signal reproduction level. Normally, the average recording level of music recorded on compact discs is -6 dB to -18 dB, so abnormal noise is -46 dB or less, that is, the number of transmission bits is 8.
It is desirable that the number is a bit or more. In the present invention, since the number of bits to be transmitted is 16 bits, which is a large number of bits, the occurrence of abnormal noise can be made extremely small even when a simple process such as a linear interpolation process is taken when error correction becomes impossible. There is an effect.

By the way, in the encoding device, the mode control signal may be output as a sub-code together with the converted acoustic signal. In this case, the decoding device is as a second embodiment of the present invention as shown in FIG. Configure. That is, in the decoding device, the converted acoustic signal transmitted from the encoder output and the mode control signal as a subcode are extracted, and the 16-bit, 192 kHz acoustic signal is input to the oversampling filter 10 as a decoder input. input. In the oversampling filter 10, this signal is oversampled by a factor of 384 kHz,
It becomes 24 bits and is given to the digital filter 11 in the subsequent stage. The digital filter 11 performs predetermined filtering based on the mode control signal previously extracted as a subcode. It is assumed that the relationship between the mode control signal and the band of the sound input input to the encoding device is predetermined.

Assuming that the sound input band is 0 kHz to 48 kHz, the 384 kHz, 24-bit signal obtained by the oversampling filter 10 is sharpened by the mode control signal to a signal in the band of 48 kHz or more by the digital filter 11. Perform the removal process. The output of the digital filter 11 is input to the D / A converter 12, converted into an analog signal, and then the aliasing distortion is removed by the low-pass filter 13 and output as an analog output. Even in this case, D /
The performance corresponding to the change in the band of the acoustic input given to the encoding device and the dynamic range can be exhibited without changing the circuit after the A converter 12.

A plurality of types of sampling frequencies, for example, 192 kHz and 96 kHz, are prepared for encoding in the encoding device, and the information about the sampling frequencies is converted into the sound with the sampling frequency identification signal as a subcode. It may be output together with the signal. In this case, in the decoding device 3, the converted acoustic signal transmitted from the encoder output and the sampling frequency identification signal as a subcode are taken out, and originally 192 kHz
The sampling frequency is changed to, for example, 96 kHz and the device is operated. With such a configuration, even if the transmission medium has the same capacity, it becomes possible for the decoding device to reproduce the sound input twice as long as the time. Of course, at this time, the reproducible band becomes 1/2, but when the width of the band is not important, it is effective, for example, for conversation transmission. In the decoding device 3, the clock frequency of the master clock supplied to the device may be halved, and the oversampling ratio of the oversampling filter 10 may be increased to 2 to 4 times.

The encoding device shown in FIG. 1 may be used for emphasis to emphasize high frequencies for encoding. In this case, for example, the acoustic input is input to the oversampling filter 1 via the emphasis circuit, or the output of the oversampling filter 1 is input to the ΔΣ modulator 2 via the emphasis circuit. Then, together with the converted acoustic signal, an emphasis identification signal indicating ON / OFF of emphasis is output as a subcode. In the decoding device, a de-emphasis circuit having a frequency characteristic opposite to that of the emphasis circuit is used as the oversampling filter 10,
It is inserted between either the digital filter 11 or the D / A converter 12, and is turned on / off based on the emphasis identification signal extracted as a subcode. This allows
For example, even if the quantization noise in the ultra-high frequency range is increased by the ΔΣ modulation, the noise can be reduced, and the signal level input to the speaker or the like during reproduction can be suppressed to a certain value or less. In particular, a speaker capable of reproducing a super high frequency generally has a low withstand voltage, and this method is effective.

The transfer function H1 of the feedback circuits 7 and 8
It is needless to say that (z) and H2 (z) are not limited to those shown in the embodiment, and higher ones or fractional polynomials may be used. The value of sampling frequency = 192 kHz is not limited to this, and the point is that the sampling frequency should be four times or more the audible band. Regarding de-emphasis, of course, this identification signal may be transmitted to switch on / off, but since the signal level in the band exceeding 20 kHz is not so large originally, only the on mode may be used.

[0019]

As described above, according to the present invention, the encoding means for encoding and transmitting the input acoustic signal at the sampling frequency of 4 times or more of the audible band, and the encoded acoustic signal in the audible band. And a decoding means for retrieving and reproducing a signal of a predetermined band including the encoding means, the quantization noise is replaced by ΔΣ modulation outside the audible band, and the band identification signal, the emphasis identification signal, and the sampling frequency identification signal are converted. Since the signal is transmitted and the decoding means performs decoding based on these signals, the audible band is 15 kHz to 20 kHz.
It is possible to obtain a high dynamic range acoustic signal without degrading the dynamic range in
Since the sampling frequency is 4 times the audible band or more, it is possible to generate an acoustic signal in the ultra-high range outside the audible band, which is meaningful for α-waves, and the circuit on the decoding device side is almost changed. The dynamic range and bandwidth can be expanded and contracted as needed.

[Brief description of drawings]

FIG. 1 is a block diagram showing an encoding device according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a decoding device in the embodiment.

FIG. 3 is a block diagram showing a decoding device according to a second embodiment of the present invention.

[Explanation of symbols]

 1, 10 Oversampling filter 2 ΔΣ modulator 3 Decoding device 4 Adder 5 Local quantizer 6 Subtractor 7, 8 Feedback circuit 9 Selector 11 Digital filter 12 D / A converter 13 Low-pass filter

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kei Kageyama 1006 Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Industrial Co., Ltd. (72) Inventor Katsuyoshi Fujii 1006 Kadoma, Kadoma City, Osaka Matsushita Electric Industrial Co., Ltd. 72) Inventor Yasunori Tani 1006 Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Industrial Co., Ltd.

Claims (7)

[Claims] 1. An encoding device comprising an encoding means for encoding and transmitting an input acoustic signal at a sampling frequency which is four times or more of an audible band, and the encoded acoustic signal includes an audible band. A decoding device provided with a decoding means for taking out and reproducing a signal in a predetermined band. 2. A coding apparatus according to claim 1, wherein the coding frequency used by the coding means for coding is 60 kHz or more. 3. The number of code bits output by the encoding means is 8.
The coding device according to claim 1, which is the above. 4. The encoding means performs ΔΣ modulation for substituting quantization noise outside the audible band.
The coding device described. 5. The encoding means performs predetermined ΔΣ modulation on the acoustic signal based on a band identification signal given in advance, transmits the band identification signal, and the decoding means reproduces based on the transmitted band identification signal. The encoding device and the decoding device according to claim 1, which control a frequency band. 6. The encoding means transmits a sound signal together with an emphasis identification signal indicating an emphasis state by enhancing the high frequency band by the emphasis means in advance, and the decoding means reproduces the frequency characteristic of the signal reproduced based on the emphasis identification signal. The encoding device and the decoding device according to claim 1, which are controlled. 7. The encoding means encodes and transmits the acoustic signal at a predetermined sampling frequency based on a sampling frequency identification signal given in advance, and the decoding means performs decoding based on the sampling frequency identification signal. The encoding device and the decoding device according to claim 1.