JPH02124622A - Quantizing error reducing device - Google Patents

Abstract

PURPOSE:To reduce the waveform distortion attended with quantization by passing only a component of a frequency band including a signal before quantization among frequency components of a digital signal after the quantization. CONSTITUTION:An m-bit digital signal 71 being an output of a decoder 7 is inputted to a quantization error reducing device 11 and its l-bit arithmetic result 81 is given to a D/A converter(DAC) 13 and a low pass filter 9 to obtain an analog signal output 10. On the other hand, A fourier transformation device 13 as the frequency detection means applies Fourier transformation to N-sample of the m-bit signal 71 to output an amplitude term output 14. In this case, a band limit digital filter 15 has a cut-off characteristic to pass through only a spectrum component f0 whose amplitude is larger than the level of average quantizing noise level. The bit number l of the output signal 81 of the filter 15 in this case is increased more than the bit number (m) of the signal 71 to reproduce a smooth waveform without any distorted component attended with quantization.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application This invention relates to digital signal processing in digital audio equipment.

Conventional technology In digital audio equipment, the recording system converts the original signal to digital and records it on a recording medium, and the playback system reads the digital signal from the recording medium, converts it to analog, and plays it back. This is an important deciding factor.

FIG. 6 shows the basic configuration of a recording/playback system of a conventional digital audio device. An input analog signal 1 is sampled by a sample and hold circuit 2 at a constant repetition period of 1/F s. The output signal of the sample and hold circuit 2 is converted into an m-bit digital signal 31 by a quantizer 3. The sample-and-hold circuit 2 and the quantizer 3 constitute an analog-to-digital converter 4. The m-bit digital signal 31 that is the output of the quantizer 3 is input to the encoder 5, where a code such as an error correction code is added. 6 is a recording medium for recording the signal encoded by the encoder 5. In this way, analog
The digital converter 4 and encoder 5 constitute a recording system. The decoder 7 performs processing such as error correction on the signal read from the recording medium 6 and decodes it into an m-bit digital signal 71. The m-bit digital signal 71 output from the decoder 7 is converted to a digital-to-analog converter [hereinafter referred to as
It is converted into an analog signal by a DAC]8. DAC
The output signal of 8 is passed through a low-pass filter 9 to F s /
After removing two or more unnecessary aliasing components, it is output as an analog signal 10. Figure 7 shows the signal waveforms of various parts when a low level signal is input as input analog signal 1, (a) is input analog signal 1, (b)
is the output waveform of the sample and hold circuit 2, (c) is the output waveform of the m-bit digital signal 31.71 and the DAC 8, and (d) is the output waveform of the analog signal 10.

Problems to be Solved by the Invention As described above, the recording and reproducing system of conventional digital audio equipment quantizes an analog signal into a finite digital signal of m bits, so especially when the signal level is low, the analog signal output 10 is The waveform will be far different from the input analog signal 1, and the influence of errors during quantization will be large, making it difficult to faithfully reproduce the analog signal 1 before quantization.

SUMMARY OF THE INVENTION An object of the present invention is to provide a quantization error reduction device that can reduce waveform distortion caused by quantization when the signal level of an input analog signal is small.

Means for Solving the Problems The quantization error reduction device of the present invention includes a filter that converts an input analog signal into a digital signal, further converts it into an analog signal, and then inserts the signal into a signal processing path for output. The present invention is characterized in that a frequency detection means for detecting the distribution of frequency components of a signal is provided, and the cutoff characteristic of the filter is varied by the output of the frequency detection means.

Effect: According to this configuration, among the frequency components of the quantized digital signal, only the components in the frequency band that includes the pre-quantized signal can be passed, and the other components can be blocked using a filter. .

Embodiments Hereinafter, embodiments of the present invention will be explained based on FIGS. 1 to 5.

FIG. 1 shows a recording and reproducing system of a digital audio device using a quantization error reduction device according to a first embodiment of the present invention. In addition, input analog signal 1, j - analog-digital converter 4, encoder 5, recording medium 6
, decoder 7, and low-pass filter 9 are the same as those shown in FIG. 6, which shows a conventional example. In FIG. 1, a quantization error reduction device 11 of the present invention is provided after the decoder 7. Here, since the bit loss of the output 81 is different from the bit number m of the input cuff 1, the number of input bits of the DAC 12 is also different from the DAC 8 of FIG.

The operation of the recording and reproducing system of the digital audio device according to the present invention configured as described above will be described below. Input analog signal 1 to analog-digital converter 4
The process is the same as the conventional example shown in FIG. 6 until the digital signal is quantized by the encoder 5, encoded by the encoder 5, recorded on the recording medium 6, and read out again from the recording medium 6 and decoded by the decoder 7. And the output of the decoder 7 is m
Quantization error reduction device 11 for bit digital signal 71
The operation result 81 of the rising bit is input to the DAC 12,
By passing it through a low-pass filter 9, an analog signal output 10 is obtained.

The quantization error reduction device 11 is constructed as shown in FIG. A Fourier transformer 13 serves as a frequency detection means, which performs Fourier transform on the m-bit input digital signal 71 and outputs an amplitude product output 14. Reference numeral 15 denotes a band-limiting digital filter for comparing the amplitude product output 14 of the Fourier transformer 13 with a certain constant value and passing only the frequency band containing the signal before quantization. Figure 3 shows the waveforms of each signal in Figure 2, (a)
is the waveform of the m-bit input digital signal 71, (b) is the amplitude product output 14 of the Fourier transformer 13, (c) is the cutoff characteristic of the band-limiting digital filter 15, and (d) is the waveform of the band-limiting digital filter 15. View bit output signal 81
This is the waveform of First, in FIG. 2, the Fourier transformer 13
performs Fourier transform on N samples of the m-bit input digital signal 71 to obtain the amplitude product output 14. As a result, the input digital signal 71 shown in FIG.
The spectrum of is shown in FIG. 3(b). and compares the amplitude product output 14 with the average quantization noise level,
As a result, fO, which is a spectral component with an amplitude larger than the average quantization noise level, is considered to be a frequency component originally included in the analog signal before quantization, rather than a waveform distortion component due to quantization. It will be done. In such a case, the band-limiting digital filter 15 is provided with a cutoff characteristic that allows only fO to pass. the result,
The band-limiting digital filter 1 is a band-pass filter that passes only the frequency fO, as shown in FIG. 3(c). Then, the bit loss of the output signal 81 of the band-limiting digital filter 15 at this time is expressed as the input signal 71.
By making the number of bits larger than m, it is possible to reproduce a smooth waveform free of distortion components caused by quantization as shown in FIG. 3(d).

FIG. 4 shows a second embodiment of the quantization error reduction device 11. 16 is a frequency detection digital filter for detecting whether or not the m-bit input digital signal 71 contains a signal component of a specific frequency band, and has a frequency detection output 17. Reference numeral 18 denotes a band-limiting digital filter that changes its cutoff characteristic in accordance with the frequency configuration 7 of the frequency detection digital filter. Fifth
The figure shows the waveform of each signal in Figure 4,
(,) is the waveform of the m-bit input digital signal 71, (
b) is the spectrum of the input digital signal 71, (c
) is the cutoff characteristic of the frequency detection digital filter 16,
(d) shows the spectrum of the frequency detection output 17, and (e) shows the cutoff characteristic of the band-limiting digital filter 18 when band-limiting is performed. (f) shows the waveform of the output signal 81 when the input digital signal 71 is band-limited by the band-limiting digital filter 18, and (g) shows the waveform (
f) spectrum.

First, the input digital signal 71 having the spectrum shown in FIG. 5(b) is filtered by the frequency detection digital filter 16 having the cutoff characteristic shown in FIG. get.

At this time, the spectrum of the frequency detection output 17 is shown in FIG.
)become that way. Figure 5(b).

As is clearer from (d), since the input digital signal 71 at this time does not include a signal component having a frequency equal to or higher than fc, the frequency detection output 17 contains only the quantization error component. Therefore, the band-limiting digital filter 18 detects that the frequency detection output 17 is only a quantization error component, and sets its cutoff frequency to fc. Therefore, as shown in FIG. 5(g), the input digital signal 71 has quantization error components greater than fc removed by the band-limiting digital filter 18. And as a result, the fifth
As shown in Figure (f), an output signal 81 with less waveform distortion due to quantization is obtained. When the input digital signal 71 contains a component equal to or higher than fc, the band-limiting digital filter 18 outputs the input digital signal 71 as is.

In the second embodiment, the case where there is only one frequency detection digital filter 16 has been described, but the cutoff characteristics of the band limiting digital filter 18 can be changed in three or more ways by using two or more frequency detection digital filters. By doing so, it is possible to perform even more optimal band limiting.

In the first and second embodiments described above, the filter for band limiting was explained as a digital filter 15.18, but the input digital signal is D/A converted to an analog signal, and then the band is limited using an analog filter. It is also possible.

Effects of the Invention As described above, according to the present invention, since the present invention includes a frequency detection means for detecting the distribution of frequency components of an input signal and a filter whose cutoff characteristic is changed according to the output of the detection means, the signal after quantization is By extracting and reproducing only specific frequency components included in the signal, it is possible to significantly reduce waveform distortion caused by quantization, especially when the signal level is low.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a recording/playback system of a digital audio device using the quantization error reduction device of the present invention, FIG. 2 is a block diagram of a first embodiment of the quantization error reduction device of the present invention, and FIG. The figure is a signal waveform diagram in the first embodiment, and the fourth
The figure is a block diagram of a second embodiment of the quantization error reduction device according to the present invention, Figure 5 is a signal waveform diagram in the second embodiment shown in -, and Figure 6 is a recording/playback system of a conventional digital audio device. FIG. 7 is a signal waveform diagram in a recording/reproducing system of a conventional digital audio device. DESCRIPTION OF SYMBOLS 1... Input analog signal, 4... Analog-digital converter, 5... Encoder, 6... Recording medium, 7...
...Decoder, 11...Quantization error reduction device, 12...
- Digital-to-analog converter, 13...Fourier transformer (frequency detection means), 15.18... Bandwidth limiting digital filter [filter], 16... Frequency detection digital filter [frequency detection means]. Agent Yoshihiro Morimoto

Claims (1)

[Claims] 1. A filter is provided in the signal processing path for digitally converting an input analog signal and then analog-converting and outputting the signal, and a frequency detection means for detecting the distribution of frequency components of the input signal of the filter. . A quantization error reduction device, wherein the cutoff characteristic of the filter is varied by the output of the frequency detection means.