JPH0265407A - Digital signal processing circuit - Google Patents

Abstract

PURPOSE:To reduce waveform distortion attended with quantization especially when a signal level is small by extracting only a specific frequency component in the frequency component included in a digital signal after the quantization and applying inverse Fourier transformation. CONSTITUTION:A Fourier transformation means 13 applies Fourier transformation to N-sample of an m-bit digital signal input 61 to obtain a phase term 16 and an amplitude term 17. Then a comparator means 14 compares the amplitude term 17 with a level of a mean quantization noise and the spectral component having a larger amplitude than the level of the average quantized noise is fed to an inverse Fourier transformation means 15 as it is. Furthermore, the spectral component whose amplitude is less than the level of the quantized noise is regarded as zero and fed to the inverse Fourier transformation means 15. Thus, The resolution of the 1-bit digital signal output 71 in the amplitude direction is increased and a smooth waveform without any distortion component attended with quantization is reproduced.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application This invention relates to digital signal processing in digital audio equipment, and in particular to realizing high-fidelity reproduction of low-level music signals.

Conventional technology In digital audio equipment, there is a recording system that converts the original signal into a digital signal using an analog-to-digital converter, etc., and records it on a recording medium, and a recording system that reads the digital signal from the recording medium and converts it to a digital signal using a digital-to-analog converter, etc. The playback system that plays back analog signals is an important element that determines the performance of digital audio equipment.

An example of the recording/reproducing system of the above-mentioned conventional digital audio equipment will be described below with reference to the drawings. FIG. 6 shows the basic configuration of a recording/playback system of a conventional digital audio device. In Fig. 6, 1 is the input analog signal, and 2 is the input analog signal 1 at a constant period of 1.
7 A sample hold circuit for sampling at Ftx; 3 a quantizer that converts the output signal of the sample hold circuit 2 into an m-bit digital signal 31;

The sample-and-hold circuit 2 and the quantizer 3 constitute an analog-to-digital converter 9. Furthermore, 4 is an encoder for adding a code such as an error correction code to the m-bit digital signal 31 which is the output of the quantizer 3;
6 is a recording medium for recording the signal encoded by the encoder 4. The analog/digital converter 299 and encoder 4 constitute a recording system.

Next, 6 is a decoder for performing processing such as error correction on the signal read from the recording medium 5 and decoding the m-bit digital signal 61, and 7 is a decoder for performing digital/analog conversion on the m-bit digital signal 61. It is a digital to analog converter. Hereinafter, this digital converter 7 will be referred to as DA.
It is called C7. 8 is a low-pass filter for removing unnecessary aliasing components from the output of the DAC 7, and outputs an analog signal 1o.

The operation of the recording and reproducing system of the conventional digital audio equipment configured as described above will be described below with reference to the drawings. Note that FIG. 7 illustrates each signal waveform in FIG. 6, where a is analog signal human power 1, b
is the output waveform of the sample and hold circuit 2, C is the m-bit digital signal 31.61 and the output waveform of the DAC-r, and d is the analog signal output waveform. First, the sample and hold circuit 2 receives the input analog signal 1 at a constant period of 1.
/Fs, and as a result, the input analog signal 1 shown in FIG. 7a changes into the waveform shown in FIG. 7S. Further, by quantizing to m bits in the quantizer 3, the waveform of FIG. 7 becomes as shown in FIG. 7C. Then, in the encoder 4, this waveform is changed to ll'? The data is recorded on the 1.78 recording medium 5 with a code such as a correction code added thereto.

Next, in the reproduction system, the signal read from the recording medium 5 is decoded by the decoder 6, and the output m-bit digital signal 61 is again converted into an analog signal by the DAC 7. And finally, 18 by low pass filter 8.
By removing aliasing components of 72 or more, the analog output 1o shown in d of FIG. 7 is obtained. By quantizing the input analog signal 1 into a finite digital signal of m bits in this way, the analog signal output 10 has a waveform that is far different from the input analog signal 1, especially when the signal level is small.

Problems to be Solved by the Invention As mentioned above, since the recording/playback system of conventional digital audio equipment quantizes analog signals into finite digital signals, the influence of errors during quantization is particularly important when the signal level is small. is large, making it difficult to faithfully reproduce the analog signal before quantization.

Means for Solving the Problems In view of the above-mentioned problems, the digital signal processing circuit according to the present invention includes a first transformation means for 7-lier transformation of an input digital signal and a spectrum obtained by the first transformation means determined in advance. and a second transform means that performs inverse Fourier transform on the output of the comparison means and the output of the first transform means, and the second transform means performs an inverse Fourier transform on the output of the comparison means and the output of the first transform means, and By extracting and reproducing only the components, waveform distortion caused by quantization is reduced.

Operation The digital signal processing circuit according to the present invention extracts only a specific frequency component from among the frequency components included in the quantized digital signal with the above-described configuration, and performs inverse Fourier transform, thereby processing the signal particularly when the signal level is small. , which significantly reduces waveform distortion caused by quantization.

Embodiment Hereinafter, a digital signal processing circuit according to an embodiment of the present invention will be described with reference to the drawings. First, FIG. 1 shows a block diagram of a recording/reproducing system of a digital audio device using a digital signal processing circuit according to a first embodiment of the present invention.

In FIG. 1, the input analog signal 1, analog-to-digital converter 9, encoder 4, recording medium 5, decoder 6, and low-pass filter 8 are the same as those in FIG. 6, respectively. Reference numeral 11 in FIG. 1 is a digital signal processing circuit according to the present invention, and the number of output bits 1f'i is 1f'i.
, 610 input bits m, the number of input bits of the DAC 12 is also different from that of the DAC 7 in FIG.

The operation of the recording and reproducing system of the digital audio equipment according to the present invention constructed as described above will be explained below. First, the input analog signal 1t is quantized by 11 by the analog-to-digital converter 9, encoded by the encoder 4, recorded on the recording medium 5, and read out again from the recording medium 5 until the digital signal is decoded by the decoder 6. This is the same as the conventional example shown in FIG. Then, the m-bit digital signal 61 which is the output of the decoder 6 is input to the digital signal processing circuit 11, and the 1-bit operation result 71 is passed through the DAC 7 and the low-pass filter 8 to obtain the analog signal output 1o. .

Next, FIG. 2 shows the circuit configuration of the digital signal processing circuit 11 according to the first embodiment of the present invention shown in FIG. Second
In the figure, 13 is a Fourier transform means for Fourier transforming an m-bit input digital signal 61 and outputting a phase term 16 and an amplitude ring 17, and 14 is a Fourier transform means for outputting a phase term 16 and an amplitude ring 17; 16 is an inverse Fourier transform means for performing inverse Fourier transform on the phase term output 16 of the Fourier transform means 130 and the output of the comparison means 14, and outputting a 1-bit digital signal 71. and a 1-bit digital signal 7
Outputs 1.

The operation of the digital signal processing circuit according to the first embodiment of the present invention constructed as described above will be described below with reference to FIGS. 2 and 3. FIG. 3 shows the waveform of each signal in FIG. 2, where a is the waveform of the m-bit digital signal 61, b is the amplitude ring output 17 of the Fourier transform means 13, C is the output of the comparison means 14, and d is the 1-bit output signal 71 of the inverse Fourier transform means 15. First, in FIG. 2, the Fourier transform means 13 performs Fourier transform on N samples of the m-bit digital signal 61 to obtain a phase term 16 and an amplitude ring 17. As a result, the spectrum of the input signal 61 shown in FIG. 3a becomes as shown in FIG. The amplitude ring 17 is then compared with the average quantization noise level by the comparing means 14. As a result of the comparison, the spectral components whose amplitude is larger than the average quantization noise level are not waveform distortion components due to quantization, but are frequency components originally included in the analog signal before quantization, so they remain as they are. It is sent to inverse Fourier transform means 16. As a result of the comparison, the spectral components whose amplitude is smaller than the level of the quantization noise are considered to be distortion components due to quantization, and therefore are sent to the inverse Fourier transform means 15 with their values set to zero. Therefore, the spectrum shown in FIG. 3 becomes as shown in FIG. 3C. Then, using the output of this transform means 14 and the phase term output 16 of the Fourier transform means 13, the inverse Fourier transform means 16 performs inverse Fourier transform to remove distortion components due to quantization. Only the frequency components contained in the analog signal can be extracted. Since the number of bits of the output signal 71 of the inverse Fourier transform means 16 at this time is unrelated to the number of bits of the input signal 61 of the Fourier transform means 130, by making the value of 1 larger than m, a 1-bit digital signal can be converted. It becomes possible to increase the resolution of the output cuff 1 in the amplitude direction as much as possible. As a result, a smooth waveform free of distortion components caused by quantization as shown in FIG. 3d can be reproduced.

Further, a circuit configuration according to a second embodiment of the digital signal processing circuit 11 according to the present invention shown in FIG. 1 is shown in FIG.

In FIG. 4, the Fourier transform means 13 and the inverse Fourier transform means 16 are the same as in FIG. Reference numeral 18 is a delay element that delays the m-bit digital signal input 61 by one sample, and 19 is a comparison means for performing a comparison operation using the amplitude ring outputs 17 of the three Fourier transform means 13.

The operation of the digital signal processing circuit according to the second embodiment of the present invention constructed as described above will be explained below with reference to FIGS. 4 and 5. FIG. 6 shows the waveform of each signal in FIG. The amplitude ring output 17 of the Fourier transform means 13 in the middle part of the figure, d is the amplitude ring output 17 of the Fourier transform means 13 in the lower part of FIG. 4, and e is the comparison means 19
The output f is the 1-bit output signal 71 of the inverse Fourier transform means 16. First, the input digital signal 31 shown in FIG.
The results of Fourier transformation using 3 are as shown in Figure 5, c and d. Looking at the 6th [C], the only spectral component whose amplitude is larger than the comparison level is f2. But the fifth
In FIG. 5, which is the result of performing Fourier transform on the input digital signal 31 one sample before that in FIG. C, the spectral components of f3 as well as f2 have amplitudes larger than the comparison level. Further, in FIG. 6d, which is the result of Fourier transform performed on the input digital signal 31 after one sample, the spectral components of fl as well as f2 have amplitudes larger than the comparison level. In general, a music signal has a strong correlation with previous and subsequent signals regarding its temporal fluctuation. Therefore, in Fig. 5C, the spectral components fl and 13, which have smaller amplitudes than the comparison level, have larger amplitudes than the comparison level before and after, so they are components that were originally included in the analog signal before quantization. It can be considered that. Therefore, in such a case,
The comparison means 19 is as shown in Fig. 5e (in addition to f2, fl
The spectral components of and f3 are also output, and all other spectral components are set to zero. And this spectrum component element 1. f2. By performing an inverse Fourier transform using the inverse Fourier transform means 15 using f3 and the phase term component 16, the analog signal before being quantized further than in the first embodiment is obtained. 1 close to
A bit digital signal element 1 can be obtained.

Effects of the Invention As described above, the digital signal processing circuit according to the present invention includes a first transform means that performs Fourier transform on an input digital signal, and a comparison method that compares the spectrum obtained by the first transform means with a predetermined value. and a second transformation means for performing inverse Fourier transform on the output of the comparison means and the output of the first transformation means, and extracts and reproduces only a specific frequency component included in the quantized signal. By doing so, it is possible to significantly reduce waveform distortion caused by quantization, especially when the signal level is small.

[Brief explanation of the drawing]

FIG. 1 is a circuit diagram of a recording/reproducing system of a digital audio device according to the present invention, FIG. 2 is a digital signal processing circuit diagram according to a first embodiment of the present invention, and FIG. 3 is a signal diagram of the above-mentioned first embodiment. 4 is a digital signal processing circuit diagram according to the second embodiment of the present invention, FIG. 6 is a signal waveform diagram in the second embodiment, and FIG. 6 is a recording/playback diagram of conventional digital audio equipment. The circuit diagram of the system, FIG. 7, is a signal waveform diagram in the recording/reproducing system of a conventional digital audio device. 2...Sample hold circuit, 3...
Quantizer, 4...Encoder, 5...Recording medium, 6...Decoder, 7.12...D
AC, s...Low pass filter, 9...
・Analog-digital converter, 11...Digital signal processing circuit, 13...Fourier transform means, 14...Comparison means, 15...Inverse Fourier transform means. Name of agent: Patent attorney Shigetaka Awano and one other person

Claims (1)

[Claims] a first transformation means for Fourier transforming an input digital signal; a comparison means for comparing the spectrum obtained by the first transformation means with a predetermined value; and an output of the comparison means and the first transformation means. A digital signal processing circuit comprising: a second transform means for performing inverse Fourier transform on the output.