JP3463513B2 - AD converter - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an AD converter for converting an analog signal into a digital code, and more particularly to an AD converter for realizing high S / N and wide band conversion by using a plurality of AD conversion elements.

[0002]

2. Description of the Related Art There are various types of AD converters, but a ΔΣ modulation type AD converter that performs high-speed sampling with 1 bit is widely used. This is a method that concentrates quantization noise in a high frequency band by using high-speed sampling and noise shaping, and if there is time-axis accuracy, conversion accuracy in a desired band can be obtained relatively easily. Therefore, it is suitable for use in LSIs and is widely used in conventional digital recorders (DAT, MD, etc.) for the purpose of digital recording of analog signals.

In particular, recently, with the release of DVD (Digital Versatile Disk), there is a tendency toward high sampling and high bit. 96k with this
A delta-sigma modulation type A / D converter, which is labeled as Hz.24 bits, has also appeared.

[0004]

However, the surface format of such a ΔΣ modulation type AD converter is 9
Although it is 6 kHz / 24 bits, sufficient performance matching the specifications has not been obtained so far due to insufficient device operating speed. More specifically, there are the following problems.

(A) It is desired to increase the oversampling ratio, but since the device operating speed cannot keep up, the oversampling ratio is limited to about 32 times at most.

(B) It is desired to increase the order of the feedback filter for higher accuracy, but there is a problem in stability such as oscillation during full swing.

(C) On the contrary, when a low-order feedback filter is used, the band becomes narrow and the dynamic range is reduced.

(D) Since the performance depends on the accuracy of the time axis, even slight jitter is vulnerable to clock jitter, such as deterioration of the SN ratio and dynamic range.

The present invention solves the above-mentioned problems, and an AD converter which obtains a high dynamic range over a wide band while using a low-order feedback filter without excessively increasing the oversampling ratio and is resistant to clock jitter. The purpose is to provide.

[0010]

To achieve this object, an AD converter according to the present invention comprises a plurality of AD conversion elements for converting an input analog signal into a digital code in a predetermined partial band, and each AD conversion. The digital code of the entire band is taken out from the combining means for combining the outputs of the elements.

As a result, it is possible to obtain an AD converter which has a high dynamic range over a wide band and is resistant to clock jitter.

[0012]

BEST MODE FOR CARRYING OUT THE INVENTION The first invention of the present invention is a common input terminal and n (n is an integer of 2 or more) ADs for converting an input analog signal into a digital code in a predetermined partial band. An AD conversion element group having a conversion element and a synthesizing unit for synthesizing n outputs of the AD conversion element group are provided, and the digital code of the entire band is taken out from the synthesizing unit.

A second aspect of the present invention is a common input terminal and n (n is an integer of 2 or more) AD conversion elements for converting an input analog signal into a digital code in a predetermined partial band. And a synthesizing means for synthesizing n outputs of the AD converting element group into an m-bit (m is a positive integer satisfying 2 ^m ≧ n + 1) digital code. The m-bit digital code synthesized from is extracted.

A third aspect of the present invention is provided with i (i is an integer of 2 or more) analog input terminals, and the analog signals input from the analog input terminals are converted into digital codes in predetermined subbands. N to convert (n is 2
An AD conversion element of the above integer), and a synthesis unit that synthesizes n outputs of the AD conversion element group into an m-bit (m is a positive integer satisfying 2 ^m ≧ n + 1) digital code,
The m-bit digital code synthesized from the synthesizing means is taken out.

Further, the AD conversion element of the present invention is a ΔΣ that ΔΣ-modulates an input analog signal with a predetermined band characteristic.
It comprises a modulator and a decimation filter that thins the sampling data by limiting the band of the output of the ΔΣ modulator with a predetermined frequency characteristic.

Further, a fourth aspect of the present invention is provided with i (i is an integer of 2 or more) analog input terminals to which signals of a microphone or a microphone amplifier each having a characteristic suitable for a predetermined partial band are input. And n (n is an integer of 2 or more) AD conversion elements for converting an analog signal input from the analog input terminal into a digital code in each predetermined partial band, and n outputs of the AD conversion elements are combined. And an synthesizing unit for converting an m-bit (m is a positive integer satisfying 2 ^m ≧ n + 1) digital code to extract the synthesized m-bit digital code from the synthesizing unit.

Further, the AD conversion element of the above invention samples the ΔΣ modulator which ΔΣ-modulates the input analog signal with a predetermined band characteristic, and samples the output of the ΔΣ modulator by limiting the band with a predetermined frequency characteristic. And a decimation filter for thinning out data. The decimation filter mainly passes in a predetermined band of the connected ΔΣ modulators and blocks other bands, and the decimation filter is connected between the i input signals. It is characterized by adapting the frequency characteristic, the phase characteristic, the group delay characteristic, and the variation therebetween, and convolving the transfer characteristics which are the inverse characteristics.

Since the present invention is configured as described above, each AD conversion element can increase the dynamic range in a predetermined divided band by using a relatively low clock and low-order feedback filter, and also stabilizes it. To be These AD
The conversion element is combined with a decimation filter that passes a predetermined band and blocks it outside the band, and removes noise outside the predetermined divided band. Furthermore, by combining the outputs of a plurality of AD conversion elements that obtain predetermined performance in different bands, a wider band and a higher dynamic range can be achieved as a whole. Further, when the outputs of the n AD conversion elements are combined, the occurrence probabilities of the rising and falling edges of the ΔΣ modulation waveform are approximately balanced, so that a conversion error due to jitter is geometrically averaged, and noise due to jitter is reduced. A reduction effect occurs secondarily.

Also, by connecting a suitable signal to each input terminal, it becomes possible to convolve optimum characteristics individually.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows the AD according to the first embodiment of the present invention.
It is a schematic block diagram which shows a converter. In consideration of future development, an AD converter for converting an analog signal from DC to 100 kHz into a digital signal is used as an embodiment according to the present invention. Therefore, the output signal is a digital code having a sampling frequency of 192 kHz and 24 bits. In the figure, 101 is an analog input terminal, 1000 is an AD conversion element (a), 2000 is an AD conversion element (b), and 3000 is an AD
The conversion element (c), 7000 is a synthesizing means, and 710 is a digital output terminal. In addition, the number written without a lead line beside the signal line represents the number of bits. The input signal X input from the input terminal 101 is an AD conversion element (a) 1000, AD
Conversion element (b) 2000 and AD conversion element (c) 3000
Commonly supplied to. In each AD conversion element, AD conversion is performed by the ΔΣ modulators 100, 300 and 500 and the decimation filters 200, 400 and 600, respectively. AD conversion element (a) 1000, AD conversion element (b) 20
00 and the output Wa70 of the AD conversion element (c) 3000
1, Wb702 and Wc703 are supplied to the synthesizing means 7000. The combining means 7000 uses these three inputs W
Then, a, Wb, and Wc are added and combined and output as an output signal W from an output terminal 710.

FIG. 2 is an internal block diagram of the AD conversion element (a) 1000. In the figure, 110 is an adder, 120 is a quantizer, 140 is a subtractor, 150 is a feedback filter, 210
Is the first FIR1 (finite impulse response) filter
(a) and 220 are second FIR2 filters (a). The analog input signal X input from the input terminal 101 is supplied to the quantizer 120 via the adder 110. 1-bit quantization is performed at an oversampling frequency of 3072 kHz. The other addition input signal of the adder 110 is the quantization noise Q supplied from the feedback filter 150. The quantizer 120 quantizes the input analog signal into 1 bit. The quantized 1-bit signal is transmitted to the decimation filter 200 and input to the subtractor 140. The subtractor 140 outputs the difference signal between the input signal and the output signal of the quantizer 120, that is, the quantization noise Q. The sampling frequency of 3072 kHz is a value selected in consideration of the current device performance, and the performance such as the obtained dynamic range deteriorates even if the sampling frequency is higher or lower than this. The quantization noise Q is converted in frequency and / or phase characteristics by the feedback filter 150 having the transfer characteristic Ha (z) and fed back to the adder 110. Here, when the signal Ya is expressed by an equation, Ya = X + (1-Ha (z)) * Q, and the signal Ya is a component fed back by the transfer characteristic Ha (z) of the component of the input signal X and the quantization noise Q. Is the sum of Therefore, the spectrum conversion of the quantization noise Q can be performed by the transfer characteristic Ha (z). It is preferable to provide a pass characteristic such that the transfer function becomes 1 in a practical low frequency range and an attenuation characteristic in a high frequency range outside the practical band. For example, LPF corresponds to this. By the ΔΣ modulation by this feedback loop, the quantization noise Q can be driven to a high band outside the practical band. The 1-bit signal thus subjected to the predetermined spectrum conversion is output from the output terminal 710 to the decimation filter 200 of the next stage as the output signal Ya.

In the decimation filter 200,
The 1-bit 3072 kHz signal is converted into the FIR1 filter 21.
Decimation to 1/4 at 0 and 24-bit 768 kHz
Down sampling and multi-biting,
The FIR2 filter 220 further decimates it to 1/4 and converts it into a 24-bit 192 kHz output Wa.

The internal structure and operation of the AD conversion element (b) 2000 and the AD conversion element (c) 3000 in FIG. 1 are similar to those of the AD conversion element (a) 1000 described in FIG. The difference is that the noise shape characteristic of the ΔΣ modulator, more specifically, the transfer function of the feedback filter is different, and the band characteristic of the decimation filter is different. These characteristics will be described below.

FIG. 3 is a diagram for explaining the signal spectrum and noise spectrum of the ΔΣ modulator. Figure 3 (a) shows ΔΣ
Output Ya of the modulator 100, FIG. 3B shows a ΔΣ modulator 300
Output Yb of FIG. 3 and the output Y of the ΔΣ modulator 500 are shown in FIG.
The characteristic of c is shown. The horizontal axis is the frequency axis and is shown in logarithmic form in order to see a wide range.

In FIG. 3 (a), S0 is the signal spectrum of the input signal X, and Na is the noise spectrum of the quantization noise driven to the high band by ΔΣ modulation. Sampling frequency of 192kHz is fs, and it is 3072k of 16fs.
Oversampling is performed at Hz, and ΔΣ modulation is performed by a feedback loop to drive the quantization noise Q to a high band outside the signal band. Therefore, the noise spectrum has a mountain shape having a peak centered on the Nyquist frequency of 1536 kHz (8 fs). The value of oversampling ratio 16 is insufficient to obtain a dynamic range of 24-bit accuracy in the entire signal band, and it is unreasonable that a high-order ΔΣ
Modulation causes a fatal defect such as deterioration in stability and oscillation. Therefore, in this embodiment, a second-order to fourth-order low-order feedback filter is adopted so that the filter always operates stably. Instead, the band for obtaining a dynamic range of 24-bit accuracy is narrowed. In FIG. 3 (a), a dynamic range with 24-bit accuracy is obtained in the band up to 24 kHz, and even within the band, the dynamic range in the range of 24 kHz to 96 kHz is rather bad.

Next, FIG. 3B will be described. Figure 3 (b)
Also controls the quantization noise as in FIG. 3 (a). The difference from FIG. 3A is that the feedback filter is from 24 kHz to 48 kHz.
The point is to use a bandpass filter that passes Hz and blocks the other bands. The bandpass filter can be obtained by, for example, optimally arranging the poles and zeros of the transfer characteristic Hb (z) of the feedback filter at frequencies other than DC. Due to these, the noise spectrum is Nb
Becomes The signal spectrum S0 is flat. In this way, a 24-bit precision dynamic range is obtained at 24-48 kHz.

FIG. 3C is the same as FIGS. 3A and 3B.
In FIG. 3C, a 24-bit precision dynamic range is obtained in the 48 to 96 kHz frequency band.

These three signals, Ya, Yb and Yc
Are flat in the band up to 96 kHz for the input signal X, but are output to the next stage as signals containing different noise spectra. Ya is supplied to the decimation filter 200, Yb is supplied to the decimation filter 400, and Yc is supplied to the decimation filter 600, and thinning and filtering are performed, respectively. Decimation uses an FIR filter with no group delay distortion to reduce the aliasing distortion due to downsampling to 100.
Sufficient characteristics are provided so as to prevent mixing into the band of kHz.

FIG. 4 shows the characteristics of the decimation filter and also shows the noise spectra contained in the signals Ya, Yb and Yc. In the figure, Fa is the frequency characteristic of the decimation filter 200, Fb is the frequency characteristic of the decimation filter 400, and Fc is the frequency characteristic of the decimation filter 600. Further, Na shown together is a noise spectrum included in the signal Ya, and Nb and Nc are noise spectra included in the signal Yb and the signal Yc, respectively. As shown in FIG. 4, the characteristic of the decimation filter does not simply remove the out-of-band component that causes aliasing distortion. Preferably, further, the respective decimation filters are combined so as to pass only the signals in the low noise spectrum region within the respective bands. By doing this, 96 kHz
Even in the band of z or less, the rising portion of the noise spectrum is filtered, and the rise of noise can be suppressed. In the figure, the broken line portion of the curve showing the magnitude of the noise spectrum Na indicates the portion that is attenuated by the decimation filter 200. Similarly, the broken line portions of the curves representing the noise spectrum Nb and the noise spectrum Nc similarly indicate the portions attenuated by the decimation filter 400 and the decimation filter 600, respectively. The 192 kHz / 24-bit signal Wa, the signal Wb, and the signal Wc obtained by down-sampling in this way are supplied to the synthesizing means 7000.

FIG. 5 is an internal block diagram of the synthesizing means 7000. Signals Wb and Wc whose signal levels are relatively low
Are added by the adder 720, and the result and the signal Wa are added by the adder 7
The output signal W is added and synthesized in 10 and output to the output terminal 710 as the output signal W.

FIG. 6 is a diagram showing the overall frequency characteristic of the output signal W with respect to the input signal X. In the figure, S is the characteristic of the signal component, and Na, Nb and Nc are the characteristics of the residual noise. From the figure, it can be seen that a dynamic range and SN ratio of 146 dB or more can be obtained in the band from DC to 96 kHz.

As is clear from the above description, such excellent characteristics are obtained by using the AD conversion elements each having a relatively narrow band in the partial band.
The AD conversion element can be realized by a relatively low-order feedback filter. Therefore, each AD conversion element always operates stably with respect to an input of any intensity. For example, it is possible to prevent the occurrence of instability and oscillation in a substantially full-scale input, or the occurrence of a small-amplitude oscillation phenomenon (peeling sound) or the like due to a rounding error in internal calculation in a DC input, in other words, a DC offset state. . Moreover, regarding the signal transfer characteristics, the AD conversion element, which originally has a flat characteristic over the entire band, is divided into a plurality of partial bands by digital decimation filters, and these are digitally added and combined, so the operation word length should be selected appropriately. Therefore, frequency continuity and phase continuity can be easily achieved.

Further, since the outputs of the plurality of ΔΣ modulators and the decimation filter are added and synthesized by the synthesizing means, the signal components are added and synthesized according to the transfer characteristics, but the noise components are geometrically averaged so that equal level synthesis is performed. For example, since the noise is attenuated by about 3 dB, the noise particularly near the boundary of the partial band is reduced, and the overall dynamic range and the SN ratio are slightly improved, which is a secondary effect.

Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 7 is a schematic block diagram showing an AD conversion apparatus according to the second embodiment of the present invention. In the figure, 11 is a first microphone (simply referred to as a microphone in the figure), 12 is a second microphone, 10
Reference numeral 1 is an input terminal for inputting a signal from the microphone 11, and 102 is an input terminal for inputting a signal from the microphone 12. Others are the same as those in the first embodiment. First
The same parts as those in the embodiment are designated by the same reference numerals, and detailed description thereof will be omitted. The microphone 11 is a microphone used in the band from DC to about 30 kHz. The microphone 12 is a microphone having excellent characteristics in the band of 20 kHz or higher. The microphone 11 has a diaphragm having a relatively large diameter, and the microphone 12 has a relatively small diameter. This is to increase the resonance point frequency and anti-resonance point frequency due to the divided vibration of the diaphragm in the microphone 12 to optimize the characteristics in the ultrasonic band. However, the sensitivity tends to decrease due to the small aperture.

FIG. 8 is a diagram for explaining the signal spectrum and noise spectrum of the ΔΣ modulator. Figure 8 (a) shows ΔΣ
Output Ya of modulator 100, FIG. 8B shows ΔΣ modulator 300
Output of the ΔΣ modulator 500 and the output Yb of FIG.
The characteristic of c is shown. The horizontal axis is the frequency axis and is shown in logarithmic form in order to see a wide range.

In FIG. 8A, S1 is the signal spectrum of the input signal X1, which has a flat spectrum in the low frequency range.

Next, S2 in FIGS. 8B and 8C is a signal component input from the microphone 12, and has a flat characteristic in a high frequency range.

Similar to the first embodiment, these are combined by the ΔΣ modulation and decimation filter 700.
When 0 is added and synthesized.

FIG. 9 is a diagram showing the characteristic of each decimation filter in the second embodiment. In the figure, Fa represents the characteristic of the decimation filter 200, and the reason that the gain is reduced by about 15 dB is to correct that the output of the microphone 11 is larger than the output of the microphone 12. Further, Fb shows the characteristic of the decimation filter 400, but in order to correct the tendency of the high-frequency attenuation of the microphone 12, it is corrected by convolving the transfer characteristic having the opposite characteristic. As a result, the total input / output characteristics are the same as in FIG. Na, Nb, and Nc in the figure are residual noises. A dynamic range of 146 dB or more is obtained in the band from DC to 96 kHz. By providing a plurality of input terminals, the microphone can be made into a multi-way system, and the characteristics of each microphone can be utilized to the maximum, and the optimum equalization characteristics can be obtained, thereby facilitating correction. In addition, the same operational effects as those of the first embodiment are achieved with respect to the basic part.

[0041]

As described above, according to the present invention, n (n is an integer of 2 or more) AD conversion elements (AD conversion element group) for converting an input analog signal into a digital code in a predetermined partial band. And a synthesizing means for synthesizing n outputs of the AD conversion element group, and the digital code of the entire band is taken out from the synthesizing means.

The AD conversion element thins out sampling data by limiting the band of the ΔΣ modulator that performs ΔΣ modulation of the input analog signal with a predetermined band characteristic and the output of the ΔΣ modulator with a predetermined frequency characteristic. The decimation filter is composed of a decimation filter, which mainly passes in a predetermined band of the connected ΔΣ modulator and blocks in other bands, and each decimation filter combines the respective outputs of the AD conversion elements. Since the output frequency characteristics are configured to be substantially flat in the entire band, each AD conversion element can increase the dynamic range in a predetermined divided band by using a relatively low clock and a low-order feedback filter. Stabilization is also achieved. These AD conversion elements are combined with a decimation filter that allows a signal to pass within a predetermined band and block the signal outside the band, and removes noise outside the predetermined divided band. Furthermore, by combining the outputs of a plurality of AD conversion elements that obtain predetermined performance in different bands, a wider band and a higher dynamic range can be achieved as a whole. Also, n AD
When the outputs of the conversion elements are combined, the appearance probabilities of the rising and falling edges of the ΔΣ modulation waveform are approximately balanced, so that the conversion error due to jitter is geometrically averaged, and the noise reduction effect due to jitter is a secondary effect. Occurs in

That is, there are the following concrete operational effects. (A) Since the oversampling ratio can be lowered, a device operation speed can be afforded, and an AD converter which is economically excellent with less variation as designed can be realized.

(B) Since the order of the feedback filter can be lowered, the operation is always stable even during full swing or when there is a DC offset.

(C) Since the microphones or AD conversion elements divided into the partial bands are added and synthesized through the decimation filter, it becomes easy to obtain a desired wide band width. At the same time, a high dynamic range can be obtained.

(D) The effect of geometrically averaging the conversion error due to the jitter occurs, and the noise reducing effect due to the jitter secondarily occurs. As described above, the present invention has been difficult to realize in the ultra-wide band. This is an excellent one that can realize an AD converter with a high dynamic range of an audio signal stably and economically.

[Brief description of drawings]

FIG. 1 is a schematic block diagram showing an AD conversion apparatus according to a first embodiment of the present invention.

FIG. 2 is an internal block diagram of an AD conversion element (a) 1000 according to the same embodiment.

FIG. 3 is a frequency characteristic diagram showing the characteristic of the output of the ΔΣ modulator in the embodiment.

FIG. 4 is a frequency characteristic diagram showing the characteristic and noise spectrum of the decimation filter in the example.

FIG. 5 is an internal block diagram of a synthesizing means 7000 in the same embodiment.

FIG. 6 is a diagram showing an overall frequency characteristic of an output signal W with respect to an input signal X in the example.

FIG. 7 is a schematic block diagram showing an AD conversion device according to a second embodiment of the present invention.

FIG. 8 is a frequency characteristic diagram showing the characteristic of the output of the ΔΣ modulator in the embodiment.

FIG. 9 is a diagram showing characteristics of each decimation filter in the example.

[Explanation of symbols]

100,300,500 ΔΣ modulator 110 adder 120 quantizer 140 subtractor 150 feedback filter 200,400,600 decimation filter 210 FIR1 filter 220 FIR2 filter 1000 AD conversion element (a) 2000 AD conversion element (b) 3000 AD conversion element (c) 7,000 Synthetic means

Claims (17)

(57) [Claims] 1. An AD conversion element group having a common input terminal, and n (n is an integer of 2 or more) AD conversion elements for converting an input analog signal into a digital code in a predetermined partial band respectively, a synthesizing means for synthesizing the n outputs of the AD conversion element group, from said combining means be those they were taken out of the digital code of the entire band, before
The plurality of AD conversion elements have different partial bands.
It should be composed of an AD conversion element that can obtain a predetermined conversion characteristic.
A characteristic AD converter. 2. A common input terminal, an AD conversion element group having n (n is an integer of 2 or more) AD conversion elements for converting an input analog signal into a digital code in a predetermined partial band, respectively. A m-bit (m is a positive integer satisfying 2 ^m ≧ n + 1) digital code that synthesizes the n outputs of the AD conversion element group, and the m-bit digital code synthesized from the synthesis means. The plurality of AD conversion elements are
Predetermined conversion characteristics can be obtained in different subbands A
An AD conversion device comprising a D conversion element . 3. An i (i is an integer of 2 or more) analog input terminal is provided, and n (n is 2 or more) converting an analog signal input from the analog input terminal into a digital code in a predetermined sub-band. An AD conversion element and an synthesizing means for synthesizing n outputs of the AD converting element, and a digital code of the entire band is extracted from the synthesizing means . Transform element
Each child has a certain conversion characteristic in different sub-bands.
AD conversion device comprising an AD conversion element as described above . 4. An n (i is an integer of 2 or more) analog input terminal is provided, and n (n is 2 or more) converting an analog signal input from the analog input terminal into a digital code in a predetermined partial band. Of the AD conversion elements, and a synthesis means for synthesizing n outputs of the AD conversion element group into an m-bit (m is a positive integer satisfying 2 ^m ≧ n + 1) digital code. M synthesized from means
A digital code of bits is taken out, and the plurality of AD conversion elements are different from each other.
It is composed of an AD conversion element capable of obtaining a predetermined conversion characteristic in a band.
An AD conversion device characterized by the above . 5. The AD converter according to claim 3, wherein i is equal to n. 6. The AD conversion element includes a ΔΣ modulator that ΔΣ-modulates an input analog signal with a predetermined band characteristic, and the Δ converter.
From claim 1, characterized in that comprising a decimation filter decimating the sampling data by limiting the bandwidth for each predetermined frequency characteristic output of Σ modulator 5 Noise
AD converter of Rekani described. 7. The AD converter according to claim 6 , wherein the ΔΣ modulator obtains a predetermined band characteristic by a transfer characteristic of a feedback circuit for feeding back quantization noise. 8. decimation filter is mainly passed in a predetermined band of ΔΣ modulator to be connected respectively, characterized in that so as to prevent the band otherwise 請
The AD converter according to claim 6 . 9. decimation filter of claim 6, wherein it has a predetermined transfer characteristic, such as frequency characteristics of an output obtained by combining the respective outputs of the AD conversion element is substantially flat in all bands respectively AD converter. 10. An i-number (i is an integer of 2 or more) analog input terminal to which a signal of a microphone or a microphone amplifier having a characteristic suitable for a predetermined partial band is input, and the analog input terminal is input. The present invention comprises n (n is an integer of 2 or more) AD conversion elements for converting analog signals into digital codes in respective predetermined sub-bands, and combining means for combining n outputs of the AD conversion elements. The digital code of the entire band is taken out from the means, and the plurality of AD conversion elements are
AD that can obtain a predetermined conversion characteristic in different sub-bands
An AD conversion device comprising a conversion element . 11. An i-number (i is an integer of 2 or more) analog input terminal to which a signal of a microphone or a microphone amplifier each having a characteristic suitable for a predetermined partial band is input, and is input from the analog input terminal. N pieces (n is an integer of 2 or more) of AD conversion elements for converting an analog signal into digital codes in predetermined sub-bands and n outputs of the AD conversion elements are combined to generate m bits (m is 2 ^m). ≧ n + 1 and a synthesizing means for a digital code of a positive integer) satisfying, be those they were taken out of the digital code of m bits synthesized from said synthesizing means, a plurality
Of the AD conversion elements of
It is characterized by comprising an AD conversion element capable of obtaining conversion characteristics.
AD converter that. 12. A i is claimed, characterized in that equal to n
Item 10. The AD conversion device according to item 10 or 11 . 13. The AD conversion element thins out sampling data by limiting a band of a ΔΣ modulator that ΔΣ-modulates an input analog signal with a predetermined band characteristic and a band of an output of the ΔΣ modulator with a predetermined frequency characteristic. claims 10, characterized in that comprising a decimation filter 12
AD converter according to any one of 1 . 14. The AD converter according to claim 13 , wherein the ΔΣ modulator obtains a predetermined band characteristic by a transfer characteristic of a feedback circuit for feeding back quantization noise. 15. The decimation filter passes mainly in a predetermined band of each connected ΔΣ modulator,
Characterized by blocking other bands
The AD conversion device according to claim 13 . 16. decimation filter according to claim 13, characterized in that it has a predetermined transfer characteristic, such as frequency characteristics of an output obtained by combining the respective outputs of the AD conversion element is substantially flat in all bands respectively AD
Converter. 17. The decimation filter mainly passes in a predetermined band of each connected ΔΣ modulator,
In addition to blocking other bands, the i
15. The AD conversion apparatus according to claim 14, wherein the transfer characteristics which are inverse characteristics are respectively convoluted by adapting to the frequency characteristics or phase characteristics or group delay characteristics between the individual input signals and variations between them.