JPH07182781A - Digital sound recorder - Google Patents

Abstract

PURPOSE:To prevent fluctuation of high frequency sound occurring at the time of band compression encoding without setting a sampling frequency to a lower value than the standard value. CONSTITUTION:A low pass filter 1 restricts the maximum frequency of an input voice signal S1 to fc. A sampling clock generator 2 generates a sampling clock signal Ss having the sampling frequency fs. An A/D converter 3 digitizes a voice signal S2 passing through the low pass filter 1 by the sampling clock signal Ss. A band compressor 4 divides a digitized voice signal S3 into plural sub-bands, and encodes them allocating preferentially bits for sub-bands of low and intermediate frequencies side. The maximum frequency fc which can be passed through the low pass filter 1 is set to a lower value than a half of the sampling frequency fs.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital recorder, and more particularly to a digital recorder for dividing a voice signal into sub-bands, band-compressing and encoding them for efficient recording.

[0002]

2. Description of the Related Art Generally, in a digital recorder, in order to efficiently record a voice signal on a magnetic medium, a magneto-optical medium, or the like,
After the audio signal is digitized, it is divided into a plurality of subbands and encoded by ADPCM as shown in FIG. In this case, since human hearing has a hearing effect of being lowered in the low range and the high range, when dividing and encoding into sub-bands, the bit allocation for the sub-bands in the low and middle range is given priority. I try to keep the rate constant. For this reason, the number of bits assigned to the subband on the treble side increases or decreases depending on the amount of information in the low-mid range, and fluctuations occur in the high-range sound when the recorded sound is reproduced, resulting in unnatural sound. hear.

In order to reduce such a phenomenon, in a conventional digital recorder, the high frequency range of a voice signal is suppressed to a necessary minimum by a low pass filter, the bit rate is kept constant and the sampling frequency is lowered. There is. For example, the maximum frequency of the voice signal is 10 kHz, and the sampling frequency is 24 kHz, which is about twice the maximum frequency of the voice signal.

[0004]

As described above, in the related art, when dividing a band into sub-bands and performing band compression coding, high-range sounds generated by giving priority to bit allocation to the sub-bands in the low and middle range are set. To prevent fluctuations,
The maximum frequency and sampling frequency of the audio signal are reduced to the necessary minimum. For example, the maximum frequency of audio signals is 10
The sampling frequency is 24 kHz.

However, the sampling frequency of the audio signal is specified for each device. For example, in the device of AES / EBU standard, the sampling frequencies are 32 kHz, 44.1 kHz, 4 kHz.
It is regulated to 8 kHz. For this reason, the output of a digital recorder whose sampling frequency is set to 24 kHz is converted to the AES whose sampling frequency is set to 48 kHz.
It has a problem that it cannot be directly connected to a digital device of the / EBU standard.

It is an object of the present invention to provide a digital recorder capable of preventing fluctuations of high frequency sounds that occur during band compression coding without changing the sampling frequency from a specified value.

[0007]

SUMMARY OF THE INVENTION A digital recorder according to the present invention comprises a low pass filter for limiting the maximum frequency of an input audio signal to fc and a sampling clock generator for outputting a sampling clock signal having a sampling frequency fs. An A-D converter for digitizing the audio signal that has passed through the low-pass filter by the sampling clock signal, and the audio signal digitized by the A-D converter is divided into a plurality of subbands. A band compressor for band-compressing and coding by preferentially allocating bits to the sub-bands on the low-midrange side, and the highest frequency fc of the low-pass filter is set lower than 1/2 of the sampling frequency fs. There is.

[0008]

The present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the present invention, and FIG. 2 is a low pass filter 1 and a sampling clock signal S.
It is a figure which shows the frequency relationship with s.

The low-pass filter 1 is a filter for suppressing the frequency component of the input audio signal S1 to a necessary minimum,
Here, the maximum frequency fc = 10 kHz or less is passed. The sampling clock generator 2 generates a sampling clock signal Ss having a sampling frequency fs, and here, AES / E.
The fs of the BU standard is 48 kHz. A-D converter 3
Converts the audio signal S2 output by the low-pass filter 1 into digital signals by the sampling clock signal Ss. Band compressor 4
Divides the digital audio signal S3 into a plurality of sub-bands, prioritizes bit allocation to the sub-bands in the low and middle tone range, and encodes by ADPCM so that the bit rate becomes constant. The recording device 5 can record the signal S4 output from the band compressor 4 on a magnetic disk, a magneto-optical disk, a magnetic tape, a semiconductor memory, or the like, and read the recorded signal.

The band expander 6 expands the signal S5 read by the recording device 5 based on the sampling clock signal Ss and outputs it as a reproduced digital signal S6. The DA converter 7 converts the reproduced digital signal S6 into an analog signal by the sampling clock signal Ss. The low-pass filter 8 has the same characteristics as the low-pass filter 1, removes unnecessary frequency components included in the output signal S7 of the DA converter 7, and outputs the output audio signal S8.

By the way, since the signal S2 input to the AD converter 3 does not contain a frequency component higher than 10 kHz, the band compressor 4 divides the digital audio signal S3 into a plurality of sub-bands and encodes them. At this time, bit allocation to subbands higher than 10 kHz is not performed. Therefore, it is possible to suppress the fluctuation of the allocated bits of the high-frequency side subband due to the information amount of the low-mid frequency range, and thus it is possible to reduce the fluctuation of the high frequency sound without lowering the sampling frequency.

In this embodiment, the maximum frequency fc passed by the low-pass filter 1 is set to 10 kHz, and the sampling frequency fs of the sampling clock signal Ss is set to 48k of the AES / EBU standard, but fs> 2fc is satisfied. It is optional.

[0014]

As described above, according to the present invention, after passing through the low-pass filter that limits the maximum frequency fc of the audio signal to less than 1/2 of the sampling frequency fs, it is divided into a plurality of subbands. By dividing and band-compressing by giving priority to the bit allocation for the low-midrange subband,
It is possible to prevent fluctuations of high-frequency sounds that occur during band compression encoding without setting the sampling frequency lower than a specified value as in the conventional case. Therefore, the reproduced voice output of the digital recorder can be directly connected to another digital device having the same sampling frequency.

[Brief description of drawings]

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

FIG. 2 is a diagram showing a frequency relationship between a low pass filter 1 shown in FIG. 1 and a sampling clock signal Ss.

FIG. 3 is a diagram showing an example of dividing bits into a plurality of subbands and allocating bits.

[Explanation of symbols]

 1 Low-pass filter 2 Sampling clock generator 3 A-D converter 4 Band compressor S1 Input voice signal Ss Sampling clock signal

Claims (1)

[Claims] 1. A low-pass filter for limiting the maximum frequency of an input audio signal to fc, a sampling clock generator for outputting a sampling clock signal of a sampling frequency fs, and an audio signal passed through the low-pass filter. An A / D converter for digitizing with the sampling clock signal and a voice signal digitized by the A / D converter are divided into a plurality of sub-bands, and bit allocation to the sub-bands in the low and middle frequency range is prioritized. And a band compressor for encoding, wherein the highest frequency fc is set to a value lower than 1/2 of the sampling frequency fs.