JPS63256018A - A/d converter - Google Patents

Abstract

PURPOSE:To simplify circuit constitution by adding a white noise signal at a prescribed band higher than that of an analog input signal thereto, and outputting the result, converting the output into a digital data after applying over-sampling, and eliminating the white noise signal from the digital data. CONSTITUTION:The frequency component below the frequency frmin of the white noise signal is eliminated by a high pass filter HF and the result is fed to an adder circuit AD. After the signal is subject to over sampling by a frequency 2fs at a sample-and-hold circuit SH, the signal is converted into a digital data by an A/D conversion circuit DT1. Then the data of the white noise signal component having a frequency of fdmax or over is eliminated by a low pass filter LF2. The signal is interleaved by a decimation circuit DS, from which a digital data having a frequency spectrum of a sampling rate by the frequency fs is obtained. Since the analog white noise signal is used as a dither signal, no D/A conversion circuit is required to simplify the circuit constitution.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an A/D conversion device used for digital audio and the like.

[Prior Art] For example, when converting original sound into digital data in digital audio equipment such as compact discs,
In order to whiten distortion due to nonlinearity of the A/D conversion circuit, a configuration as shown in FIG. 3 is adopted. In the same figure,
Random data is generated from the random data generation circuit RD as a dither signal at a sampling frequency f, and is sequentially latched by the latch circuit L, and then sent to the D/A conversion circuit DT.
2 into an analog dither signal. This dither signal is supplied to an adder circuit AD and added to the analog input signal, which is the original sound. This addition output is sampled at a frequency fS by a sample hold circuit SH, and then converted into digital data by an A/D conversion circuit DT3. After this digital data is latched by the latch circuit L2, it is supplied to the subtraction circuit SU. Random data is delayed by a delay circuit and supplied to the subtraction circuit SU in synchronization with the data from the latch circuit L2, and the random data is subtracted.

As described above, by adding random data, converting it to digital data, and then subtracting the random data, the distortion caused by the nonlinearity of the A/D conversion circuit is turned into white noise, which reduces the sound quality. It is designed to prevent this.

[Problems to be Solved by the Invention] The above method requires a D/A conversion circuit for converting random data into an analog dither signal, resulting in a complicated structure.

Furthermore, if there is a difference in gain between the D/A conversion circuit and the A/D conversion circuit, a subtraction error remains in the output of the subtraction circuit SU, resulting in an increase in noise.

The present invention eliminates the need for a D/A conversion circuit, simplifies the circuit configuration, and eliminates errors associated with subtraction.

[Means for Solving the Problems] The present invention adds an analog input signal and a white noise signal, thereby adding a white noise signal of a predetermined band in a higher frequency range to the analog input signal and outputting the result. After oversampling this output, it is converted into digital data, and a low-pass filter means is provided for removing the white noise signal from this digital data.

[Embodiment] In FIG. 1, NG is a white noise generation circuit that generates an analog white noise signal over the entire frequency range. HF is a high-pass filter that transmits only noise signals higher in frequency than the analog input signal from terminal C. AD is an adder circuit, and LFl is a low-pass filter, which is used to remove unnecessary high-frequency components of the white noise signal. The high-pass a-pass filter HF, the adder circuit AD, and the low-pass filter LF1 constitute an adding means. S
H is a sample hold circuit, DTl is an A/D conversion circuit,
L is a latch circuit. LF2 is a low pass filter, D
S is a decimation circuit.

Next, the operation will be explained. As shown in FIG. 2a, the white noise generation circuit NG generates a white noise signal having a uniform frequency spectrum over the entire frequency range. This white noise signal is filtered by a high-pass filter HF.
As shown in FIG. 2, frequencies below f are removed and supplied to the adder circuit AD. The adder circuit AD has an address whose upper limit is the frequency f (f ≦ f ) as shown in FIG. 2C.
x dmax rain analog input signals are supplied, and both are added to generate a signal with a frequency spectrum as shown in FIG. 2d. This signal is fed to a low-pass filter LF1, and the frequency f as shown in FIG.
The max signal above is removed. This signal is sample hold circuit S
oversampled by H at frequency 2f,
An aliasing component is generated in the output as shown in FIG. 2f.

Note that the above frequency f has a folded component around rx It is set to be larger than the wave number f.

da+ax In other words, it is set so that the aliasing component of the white noise signal does not overlap with the analog input signal.

The signal sampled and held in the above is the A/D conversion circuit D.
It is converted into digital data by T1. This data includes white noise caused by the nonlinearity of the A/D conversion circuit DTI, as shown by diagonal lines in FIG. 2g.

After this data is latched by a latch circuit, it is processed by a low-pass filter LF2 at a frequency f as shown in the second figure.
The data corresponding to the max white noise signal is removed. This signal is decimated by a decimation circuit DS, and the frequency f
Digital data having a frequency spectrum as shown in FIG. 2i is obtained at a sampling rate of

Since this data contains a frequency spectrum caused by nonlinearity of the A/D conversion circuit that is converted into white noise, deterioration in sound quality can be prevented.

In the above example, the high-pass filter F and the low-pass filter LF1 are used to limit the white noise signal to a predetermined band only, but a band-pass filter may be used instead of the high-pass filter HF. However, in this case,
A low-pass filter is also required on the analog input signal side.

[Effects of the Invention] According to the present invention, since an analog white noise signal is used as a dither signal, a D/A conversion circuit is not required, and the circuit configuration can be simplified.

Moreover, since there is no need to subtract the dither component after A/D conversion, there is no remaining dither component due to a subtraction error, and no deterioration in sound quality occurs.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is an explanatory diagram showing frequency spectra of each part of FIG. 1, and FIG. 3 is a block diagram showing a conventional example. AD...Addition circuit NG...White noise generation circuit HF...High pass filter LFl...Low pass filter SH...Sample hold circuit DTl...A/D conversion circuit LF2... Low-pass filter and above Patent applicant Nippon Presidicin Circuits Co., Ltd. Figure 1

Claims (1)

[Claims] Adding means receives an analog input signal and a white noise signal as input, adds a white noise signal of a predetermined band in a higher frequency range to the analog input signal, and outputs the result, and sample hold means oversamples the output of the adding means. , an A/D converter that converts the output of the sample and hold means into digital data, and a low-pass filter that removes the white noise signal from the digital data from the A/D converter. conversion device.