JPS60202576A - Corrector of audio signal data - Google Patents

Abstract

PURPOSE:To set the input of an A/D converter at a rather high level and to correct the saturated area of output data, by eliminating a discontinuous part that is produced by the sample data corresponding to an input exceeding and A/D conversion enable range. CONSTITUTION:The input signal of an A/D converter 1 is denoted by X, and the digital signal which is A/D-converted and outputted is denoted by (a). Then the output of the converter 1 goes to the maximum value Ma or the minimum value Mi of the converter as long as the output exceeds the conversion enable range of the converter 1. A control circuit 3 compares the digital output signal of the converter 1 with the value Ma and Mi and outputs ''1'' and ''0'' according to the coincidence and dissidence of the comparison result. In other words, the output signal (b) is set at ''1'' to show that the convension enable range of the converter 1 is exceeded and then at ''0'' to show that said range is not exceeded respectively. Thus the signal (b) serves as a control signal and is supplied to a correction value calculating circuit 2 and a correction circuit 4.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to an audio signal data correction device that corrects a saturated portion of audio signal data.

[Technical background of the invention and its problems]

Recently, research on digital audio technology has been actively conducted. This converts the audio signal into a digital signal,
The signal is recorded and reproduced as a so-called PCM signal, or subjected to signal processing. Therefore, an AD converter is required to encode the audio signal, which is a C analog signal, and a DA converter is required to restore the signal.

By the way, the dynamic range of sound handled as an audio signal is about 120 dB, but in order to digitize this signal for recording, playback, or signal processing, an input signal that can be quantized to about 20 bits (b-th) is required. A converter or D-A converter is required.

However, the quantization number of conventional conventional successive transformation type or integral ring type human transformers is about 16 bits at most, which is 69, making it difficult to achieve sufficient quantization. Of course, there are A-D converters with a large number of bits, but the time required for conversion is long, making real-time conversion difficult, and parallel conversion types are expensive.

Therefore, in order to compensate for the narrow dynamic range of conventional human converters and to compensate for the lack of capacity of the medium for transmitting and recording digitized signals, after compressing the amplitude of the analog signal, Various methods have been considered in which the D-person conversion is performed and the amplitude is expanded after the D-person conversion.

However, in order to perform such compression and expansion accurately, it seems difficult to do so due to reasons such as the need for highly accurate analog circuits such as extremely high-speed and highly accurate voltage-controlled amplifiers or extremely accurate non-linear amplifiers. In other words, the input level of the analog signal is A.
It is desirable to set the level to a maximum level that does not exceed the conversion range of the -D converter. However, in this case, the maximum amplitude of the audio signal (for example, a music signal) appears instantaneously, so the frequency of 7 is low and the efficiency is poor.

Therefore, the input level of the C-to-D converter is increased. [Object of the Invention] The present invention has been made in view of the above-mentioned problems of the conventional A-D converter for analog and 4G audio signals. So, A
- By eliminating discontinuities caused by sample data for inputs that exceed the D conversion range, it is possible to set the input of the human converter higher without fear of saturation, i.e. to improve efficiency and to record data already recorded. The above person who is
Analog audio signal □ that can correct the saturated portion of the output data of the converter
An object of the present invention is to provide a conversion data correction device.

[Summary of the invention]

The present invention provides a digital audio signal generating means for obtaining a saturated digital audio signal at a predetermined level or higher, a means for detecting a temporal range of a saturated portion of the digital audio signal obtained by this means, and a time range detected by this means. and a correction signal generating means for generating a digital signal in which a signal in a target range is estimated and corrected from signal states before and after the time range.

〔Effect of the invention〕

According to the present invention, the signal of the saturated portion of the digital audio signal can also be used as an appropriate digital signal, so the influence of saturation can be reduced, and the input level can be set higher. Further, in the present invention, the digital signal for the correction is generated according to the state of the signal before and after the temporal range of the saturated portion of the digital audio signal. , a digital signal that is different from the actual signal but relatively close to the input signal is obtained.

[Embodiments of the invention]

An embodiment of the present invention will be explained below.

Figure ts1 shows the configuration of an embodiment of the present invention. This audio signal data correction device u, the A-D converter ratio, the correction value calculation circuit (2) which receives the power, the control circuit (3) and the compensation circuit (2).
゜#-7'4 km%f-1mTE&WIKmH8h.

:The input analog audio signal of the MuD converter (1) consisting of the positive circuit (4) is designated as X, and after this signal is sampled and held, the human-
When the digital signal converted and output is □□ heat, while the above input signal X exceeds the convertible range of the A-D converter (1), the output of the A-D converter (1) is This is the maximum value Ma or minimum value M1 of this converter. The control circuit (3) is a digital converter of the A-D converter (1).
゛) Compare the output signal with the maximum value M- and the minimum value Mi, and output 11 when both are equal, and 0 when they are not equal.

That is, as shown in FIG. 2, when this output signal is 1, it means that the convertible range of the human converter (1) has been exceeded, and when it is 0, it means that it has not exceeded this range. This control signal is input to the correction circuit (2) and the correction circuit (4).

The above control circuit (3) can be realized by a simple combination of AND gates and the like. The configuration of this example is shown in FIG.

This control circuit (3) consists of seven D-type flip-flops (FFI) to (FF7) connected in series, the Q output of the flip-flop (FF7), and the exclusive input of each of the seven and 70 flip-flops. Seven exclusive OR circuits (F!1) to (17) that take the logical sum, an AND circuit (5) that takes the logical product of the outputs of these exclusive OR circuits, and the output of this AND circuit (5) It consists of a flip-flop (FF8) as an input.

The digital signal a output from the A-D converter (1) is input in series from the MSB to the terminal T, and the four clock signal (CLK) is input to the terminal T. Maximum value Ma of A-D converter (1) output = 01111111. Minimum value Ml=100
00000, each exclusive circuit (B1)
In ~(B7), the signal of each bit is compared with the MSB, so if the input signal is Ma or Mi, the AND circuit (5) output will be 1, and the input to terminal T will be the same as the clock signal. Assuming that the pulse is 1/8 and the rise coincides with the center of 18B, it is determined whether Ma and MI are the same every 8 pits.

The correction circuit (4) is controlled by the control signal A output from the control circuit (3), and when the signal is 0, outputs a signal obtained by delaying the output digital signal a of the A-D converter (1) by a predetermined time. , when the control signal is 1, a signal d obtained by delaying the calculated digital correction signal C by a predetermined time is output to a correction value calculation circuit (2) which will be described in detail later. This correction circuit (4
) is shown in FIG. 4.

As is clear from this figure, the output a of the AD converter (1) is delayed by (N+α) 7 hours, which will be described later, in the delay circuit a0 and amplified by the amplifier (A1). Here, τ indicates the sampling interval.

On the other hand, the correction data C calculated by the correction value calculation circuit (2) is amplified by the amplifier (A3). On the other hand, the obtained control signal is sent to the control circuit (3) by the delay circuit (L terminal), where it is delayed for a predetermined period of time, and the amplifier (AI), (A) is selectively operated. As the output d of the circuit (4), a corrected digital signal obtained by additionally correcting the correction data is obtained by the correction value calculation circuit (2) at the saturated portion of the human displacement converter (1).

By the way, the correction value calculation circuit (2) calculates the range of the first part of the control signal output from the control circuit (3), that is, the range in which the input analog audio signal can be converted by the human converter (1). This is a circuit that predicts the input signal in the exceeded time range based on the shape of the input signals before and after the time range, and generates a digital correction signal.

Here, one method of obtaining the C correction data in the correction value calculation circuit (2) will be explained. Let n-3 be the number of samplings that fall within the time range that exceeds the convertible range of the Mu D converter (1), and α be the time required to obtain °C correction data in the correction value calculation circuit (2). , N+α, and n, the correction method changes.

That is, when n is small, C correction is performed using a cubic curve; when n is large, correction is performed using a quadratic curve and a cubic curve; and when n is even larger, correction is performed using a quadratic curve, a straight line, and a cubic curve.

Figure 5 diagrammatically shows the magnitude of the input signal at each point in time.
The following is a case-by-case explanation. The lock, f(τ), means the waveform of the input signal.

(1) When n≦N0 In this case, before the data of the erroneous part is output from the delay block, the correction value calculation circuit (2) calculates f (o),
Correction data can be calculated and output using the values of f(r), f((ns)r), and f(at).

Therefore, if we use the C cubic curve as the C1 correction curve,
A curve that satisfies the following conditions as shown by the dotted line in Figure 6: y
−128+B, fl +Cz+]) is obtained.

f(zu)=-Aτl+Bτt-Cτ×Df(o)==D f((n-1)r) =A((nl)r) ” +B
((n-1)r)”+C(n-1)r+nf(nf
)=person(nf)"+B(nr)I+Cnr-)-1)(
2) In the case of 3=N+1 (see Figure 7) In this case, the data that the correction value calculation circuit (2) values before the erroneous portion of data is output from the delay block is f((n-1) ) until ). Therefore, a curve satisfying the following condition is fitted to f(τ), and C1f(τ) is inserted.

! (-2τ)=-8Aτ-+4B--20τ+Df(-
τ1 =-Afa +B--Cτ+Df(o)=D f((n-1)f)=:A((n-1)T)" +H(
(n-4)τ) Shigeko〇(n-s)τ0Df(2τ) and onwards, a cubic curve (y=Az”+Br”+
Cz+D) and enter the correction data.

f(o)=n f(τ)=Aτj 10B
((n-1) rl mushroom + C (n-1) clove Df (nT)
= A (nt) Hatake + B (nr) Snow + Cnt + D (3) Lu≧
In the case of N+2, in this case, the correct data value is up to f (o) before the erroneous part of the data is output from the delay block. Therefore, first of all, we need a quadratic curve '3' that satisfies the following conditions.
= At (s-Bl) Add proverb + C1.

f (-2 pieces) = A1 (-iτ-Bs) 100.

f(zu) = person, (-TmBl) Shigeko〇.

f(o)=c.

Next, this case can be divided into two cases.

■When Le<B, +H In this case, the above quadratic curve 3'=At(2-B
, )" + C1 is applied. C correction data is input, but before outputting the apex f (Bl) of the quadratic curve, the correction value calculation circuit (2) is inputted with f ((n-1) y), f ( nt) is input, and new correction data can be calculated.

That is, first, from f(τ) to f((W-N)τ), the above quadratic curve 1 "' At (" B+) +
Calculate correction data by applying C.

Next, from f ((s-N+1) r) to f ((n-z
)r) t The following four points ((ruN-1)τ
, f()) ((3-N)τ, fC)) ((n-t)τ, f(1) (Leτ, fC)) Cubic curve passing through: y; person, xl 10 B℃ snow Gojut” +
Determine DI and apply it to C to calculate C correction data.

When ■≧B1+N+1 (see Figure 8) In this case, apply the above quadratic curve y=^s(:c-B, 11+C1!) to calculate 7(Bt) from (τ), and then C The correction value calculation circuit (2) cannot output correction data applied to the curve calculated by C using f((nt)r) and f(1L).

Therefore, after f(Bs+τ), /'(Bl) is used as a correction value and C is used, and the correction value calculation circuit (2) calculates /((-t)r)
, font). After f((rL-N+1)τ), use the following 4
Cubic curve passing through the point y=4. , z4+13. Me+C,,
The correction value applied to f+D is used.

((Rue N+1)τ1f(BS))l((rL−N)τ
, f(B,)).

((1L-1)T, f((n-t)τ)), (rL
t, f(yn)), that is, in the range of f(τ) to f(B1), y=A, (""nt) Shigeko et al. as f (ns+τ)
~f(C-N)τ), f(Bl)=C, and f((m-N+1)T)~J'((m-1)τ)
In the range of 31 = As3:fi + Bszf +
c, r+n, are used.

As described above, in the correction value calculation circuit (2), A-
Correction data can be output regardless of the length of the temporal range that exceeds the convertible range of the D converter (1).

In the above embodiment, the correction data was mainly obtained using a cubic curve, but it is of course possible to obtain the correction data using other lines. For example, correction data can be obtained using an elliptical curve as shown in FIG. 9, or a tangent line of this temporal range as shown in FIG.

By the way, in the above A-D converter, the A-
Since a digital signal with a larger number of bits is obtained using a D-converter, a D-person converter with a larger number of bits than the human converter described above is required to convert the digital signal into an analog signal. However, it has the effect of giving the impression that the dynamic range has expanded. Furthermore, even if the DA converter has a large number of pits, it is not as demanding as the A-D converter in terms of cost and speed.

The present invention is applicable not only to converting an analog audio signal into a digital signal in an A-D converter but also to a signal having a saturated portion obtained as a digital audio signal from the beginning.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the configuration of an embodiment of the present invention. Figure 2 shows the man-handling converter (1) and control circuit (3) in Figure 1.
FIG. 3 is a diagram showing the relationship between output waveforms. FIG. 3 is a diagram showing an example of the configuration of the control circuit (3) in FIG. 1. FIG. 4 is a diagram showing an example of the configuration of the correction circuit (4) in FIG. 1. 5 to 8 are diagrams for explaining the correction method in the correction value calculation circuit (2) of FIG. 1. FIG. 9 and FIG. 10 are diagrams for explaining another method of correction. 1... A-D converter, 2... Correction value calculation circuit 3.
...Control circuit, 4...Correction circuit. Agent Patent Attorney Noriyuki Chika (and 1 other person) Figure 1 Figure 21! 1 3 wi 1411 f(t) Figure 5 Figure 6 Figure 7 Figure 8 wi Figure 9 10E

Claims (2)

[Claims] (1) Digital audio signal generating means for obtaining a saturated digital audio signal at a predetermined level or higher, means for detecting the temporal range of the saturated portion of the digital audio signal obtained by this means, and a temporal range detected by this means. and a correction signal generating means for generating a digital signal in which a signal in a range is estimated and corrected from signal states before and after the temporal range. (2) The digital audio signal generating means is an A-D converter that converts an analog audio signal from analog to digital.
The audio signal data correction device according to claim 1, wherein the audio signal data correction device is a converter.