JPH1079667A - Distortion detecting device, device and method for distortion correction - Google Patents

Abstract

PROBLEM TO BE SOLVED: To correct the distortion of an input digital audio signal by detecting the appearance frequency of each code expressing the quantization level of a digital audio signal and detecting distortion based on the linear error of an analog/digital(A/D) converter from a signal converted by the converter. SOLUTION: A switching means 1 in the distortion correcting device 10 switches a digital audio signal converted into a code by the A/D converter 11 and stores the switched signal in a storage means 2. An appearance frequency detecting means 3 reads out the digital audio signal and detects the appearance frequency of each code expressing the quantization level of each of plural samples and a normalizing means 4 detects distortion based on the linear error of the A/D converter 11 from the detected appearance frequency as normalized appearance frequency. An error bit detecting means 5 detects an error bit from the code of the appearance frequency and a digital signal processing means 6 corrects an error bit detected from the inputted digital audio signal and outputs a corrected signal.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an analog / digital converter which converts a digital audio signal into a digital audio signal by converting the analog / digital signal into a digital audio signal. The present invention relates to a distortion detection device, a distortion correction device, and a distortion correction method for detecting and correcting the distortion included in a digital audio signal.

[0002]

2. Description of the Related Art Conventionally, an analog / digital converter (A /
D converter: Analog / Digital converter) is A / D
There is a linearity error due to the conversion accuracy. When an analog audio signal is converted into a digital audio signal by the A / D converter, the converted digital audio signal is distorted due to a linearity error of the A / D converter (hereinafter, referred to as a conversion error). In this case, when performing A / D conversion using the A / D converter, the resistance is adjusted at the time of A / D conversion by adjusting the resistance value inside the A / D converter or adjusting the bit weight by an external circuit. The distortion based on the error can be reduced to some extent.

[0003]

However, many music software (digital audio signals recorded in the past) which have been A / D converted and digitally recorded in the past,
In the first place, a low bit number (for example, 13 bits or 1 bit)
A / D converter with low conversion accuracy (4 bits)
/ D conversion, there is a conversion error even if the A / D converter itself is adjusted (current A / D converter
(For example, 16 bits), the distortion based on the conversion error due to the poor conversion accuracy is one of the factors that degrade the quality of the recorded music signal. Has become one. Therefore, it is now desired to obtain a digital audio signal with reduced distortion based on the conversion error of the A / D converter used in the past recording.

In this case, the A / D used in the past recording
The characteristics of the converter are measured, and the A / D
Although it is possible to correct the conversion error, since the characteristics of the A / D converter change with time, the characteristics at the time of A / D conversion, that is, the characteristics at the time of A / D conversion from the A / D converter used at the time of A / D conversion itself, ie, Conversion errors cannot be accurately detected. In many cases, the 13-bit or 14-bit A / D converter used for recording does not exist. In this case, the characteristics of the A / D converter cannot be detected from the A / D converter. As described above, conventionally, regardless of whether or not the 13-bit or 14-bit A / D converter used in the past recording exists, the A / D conversion error of the digital audio signal can be reduced by using this. It is difficult to detect accurately, and there is a problem that the A / D conversion error cannot be reduced.

Further, a conversion error at the time of A / D conversion is estimated from a digital audio signal recorded from a digital audio signal having a distortion based on the conversion error (from an already recorded music signal).
(Detection) was also difficult. That is, the conversion error at the time of A / D conversion is estimated (detected) from the digital audio signal.
A measuring device using a sine wave histogram method is known as a device for performing the above. In this measuring device, a sine wave without distortion is input to an A / D converter to be measured, a histogram is obtained, and the histogram is obtained. In the case where a conversion error of a measurement target is measured using a signal source other than a sine wave such as a digital audio signal, the conversion error cannot be estimated (detected).

As described above, conventionally, a digital audio signal (music signal) that has already been subjected to A / D conversion and recorded has been recorded.
Therefore, it is difficult to estimate (detect) a conversion error at the time of A / D conversion, and it has not been possible to obtain a digital audio signal with reduced distortion based on the conversion error at the time of A / D conversion.

The present invention relates to an analog / digital converter having a linearity error.
Estimate (detect) distortion based on the linearity error of the analog / digital converter at the time of analog / digital conversion from the digital audio signal that has been analog / digital converted by the digital converter (from the already recorded music signal), An object of the present invention is to provide a distortion detection device, a distortion correction device, and a distortion correction method that can correct a digital audio signal so as to reduce distortion included in the digital audio signal.

[0008]

According to a first aspect of the present invention, there is provided a storage unit for storing a digital audio signal, and the digital audio signal is read from the storage unit, and each of the plurality of samples is read for a plurality of samples. An appearance frequency detection means for detecting an appearance frequency for each code representing a quantization level; and a distortion based on a linearity error of the analog / digital converter based on the appearance frequency for each code detected by the appearance frequency detection means. It is characterized by comprising a normalizing means for detecting as a normalized appearance frequency, and an error bit detecting means for detecting an error bit from a code of the normalized appearance frequency detected by the normalization means.

The invention according to claim 2 is an analog / digital converter.
Conversion error detection means for detecting, from the appearance frequency of each code of a digital audio signal converted into a code by a digital converter, an error bit having a distortion based on a linearity error of the analog / digital converter, and conversion error detection means And a conversion error correction unit for correcting an error bit in which a distortion based on a linearity error of a digital audio signal analog-to-digital converter has occurred based on a detection result from the digital audio signal.

According to a third aspect of the present invention, in the distortion correction apparatus according to the second aspect, the conversion error detecting means includes:
Storage means for storing the digital audio signal input from the switching means, and occurrence frequency detection means for reading the digital audio signal from the storage means and detecting the frequency of occurrence of each code representing a quantization level for a plurality of samples Normalizing means for detecting a distortion based on a linearity error of the analog / digital converter as a normalized frequency of occurrence from the frequency of occurrence of each code detected by the frequency of occurrence detection means; And an error bit detecting means for detecting an error bit from the code of the normalized frequency of occurrence detected in (1).

According to a fourth aspect of the present invention, in the distortion correction apparatus of the third aspect, the appearance frequency detecting means has a histogram memory for storing a code representing a quantization level of the digital audio signal as an address, Data of the address corresponding to the code of the digital audio signal read from the storage means is read from the histogram memory, 1 is added, the added data is written in the address memory, and a histogram representing the frequency of appearance for each code is created. Features.

According to a fifth aspect of the present invention, in the distortion correction apparatus according to the third aspect, the normalizing means includes a plurality of codes whose quantization levels are close to each other based on the appearance frequency of each code obtained by the appearance frequency detection means. Is a block, all codes are divided into a plurality of blocks, an average appearance frequency is obtained for each block, and an appearance frequency obtained by dividing each code in each block by the average appearance frequency is set as a normalized appearance frequency. And

According to a sixth aspect of the present invention, in the distortion correction apparatus according to the third aspect, the error bit detecting means is configured to determine a quantum for each code based on the normalized frequency of occurrence detected by the normalizing means. It is characterized in that the difference between the conversion levels is detected as an error bit having a distortion based on the linearity error of the analog / digital converter.

According to a seventh aspect of the present invention, in the distortion correction apparatus of the second aspect, the digital signal processing means comprises:
It is characterized in that error bits of an input digital audio signal are corrected to a code estimated to be correct and output.

According to an eighth aspect of the present invention, in the distortion correction apparatus of the seventh aspect, the correct code estimation in the digital signal processing means is performed by calculating the average of the normalized appearance frequency of each code having error bits by X. When the same value of the digital audio signal continues for a plurality of samples K _t from time K ₁ to K _t , the value of the first sample having the same value and 1
Take the first difference D1 from the value of the previous sample, take the second difference D2 between the value of the last sample followed by the same value and the value of the next sample, and take the first difference D1 and If the second difference D2 is a positive value, the value of the K _{t / X} of the plurality of samples to K _t is +1, if the first difference D1 and the second difference D2 is a negative value, a plurality of samples Among them, the value from K ₁ to K _t (1-1 / X) is set to +1.

The invention according to claim 9 is an analog / digital converter.
Storage means for storing a digital audio signal converted into a code by a digital converter, and reading out the digital audio signal from the storage means, and detecting, as a level distribution, an appearance frequency of each code representing a quantization level for a plurality of samples; Low frequency filtering is performed on the frequency distribution detected by the frequency detecting means and the level distribution detected by the frequency detecting means to approximate the ideal level distribution when there is no error in the analog / digital converter. A normalization means for detecting a level distribution obtained by dividing the level distribution before being performed by the result approximating the ideal level distribution as a normalized level distribution, and a digital audio signal from the level distribution normalized by the normalization means. Conversion table calculating means for calculating a conversion table for each level It is characterized in that a data replacing means for outputting the level of said digital audio signal by substituting the level data according to the conversion table.

According to a tenth aspect of the present invention, a conversion error of a used analog / digital converter is detected from a level distribution of a digital audio signal recorded using the analog / digital converter, and the detection is performed. A conversion table is calculated from the obtained conversion error, a table conversion process is performed on the digital audio signal, and distortion generated by a conversion error of the analog / digital converter is corrected.

The invention according to claim 11 is the same as the claim 1.
0, wherein the conversion table is:
The digital audio signal recorded using the analog / digital converter is subjected to a level distribution of a fixed number of samples, and the level distribution is subjected to a low-pass filtering process, which is ideal when there is no error in the analog / digital converter. It is characterized in that it is created by multiplying the result obtained by dividing the level distribution before the low-pass filtering process is performed by the result of approximation to the ideal level distribution.

According to a twelfth aspect of the present invention, a linear count-up signal is input to an analog / digital converter, and an input signal and an output signal of the analog / digital converter at this time are compared with each other. The conversion error of the digital / digital converter is calculated, and the distortion based on the conversion error included in the output signal from the analog / digital converter is corrected by the conversion error calculated as described above.

According to the first to eighth aspects of the present invention, it is possible to detect a bit having a distortion based on a linearity error of an analog / digital converter from a frequency of appearance of each code of a digital audio signal, Further, based on the detection result, in order to correct the bit in which the distortion based on the linearity error of the analog / digital converter of the digital audio signal has occurred, it is based on the linearity error of the analog / digital converter of the input digital audio signal. Distortion can be reduced.

According to the fourth aspect of the present invention, there is provided a histogram memory for storing a code representing a quantization level of a digital audio signal as an address, and an address corresponding to the code of the digital audio signal read from the storage means. Is read from the histogram memory, 1 is added, and the data obtained by adding 1 is written to the address memory, and a histogram representing the frequency of appearance for each code is created. Even if the signal source is not a sinusoidal wave, the frequency of appearance for each code can be detected.

According to the fifth aspect of the present invention, all codes are divided into a plurality of blocks by using a plurality of codes having quantization levels close to each other as one block based on the appearance frequency of each code, and the average appearance frequency is determined for each block. , And the appearance frequency obtained by dividing each code in each block by the average appearance frequency can be obtained as the normalized appearance frequency, so that the distortion based on the linearity error of the digital audio signal is detected as the appearance frequency for each code. can do.

According to the sixth aspect of the present invention, based on the normalized frequency of occurrence detected by the linearity error detecting means, the difference between the quantization levels for each code is determined by the analog / digital converter. It can be detected as the number of bits in which distortion based on the linearity error has occurred.

According to the seventh aspect of the present invention, the error bit of the input digital audio signal detected by the error bit detecting means is replaced with a code estimated to be correct and output, so that the digital audio signal is output. , The distortion based on the linearity error of the analog / digital converter can be reduced.

According to the ninth to eleventh aspects of the present invention, the conversion error of the used analog / digital converter is determined from the level distribution of the digital audio signal recorded using the analog / digital converter. , A conversion table is calculated from the detected conversion error, and a table conversion process is performed on the digital audio signal to correct the distortion caused by the conversion error of the analog / digital converter. Can be corrected, and A / D conversion errors can be significantly reduced.

According to the twelfth aspect of the present invention, a linear count-up signal is input to an analog / digital converter, and an input signal and an output signal of the analog / digital converter at this time are compared. The conversion error of the analog / digital converter is calculated, and the distortion based on the conversion error included in the output signal from the analog / digital converter is corrected by the conversion error calculated as described above.
If the A / D converter used at the time of recording the music source is present without any change in the conversion characteristics, for example, a music signal currently recorded at 16 bits is converted to a music signal having a bit length of 24 bits or more in the future. Can be converted.

[0027]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram illustrating a configuration example of a distortion correction device according to the present invention.

Referring to FIG. 1, this distortion correction device 10
Converts an analog audio signal into a digital audio signal (DAT: Dig: Digital Audio Tape)
italAudio Tape) and compact discs (CD: Compact)
A digital audio signal recorded on a recording medium 12 such as a disk is read from the recording medium 12 to reduce distortion based on a conversion error at the time of A / D conversion from the digital audio signal. Therefore, the distortion correction device 10 includes a switching unit 1 for switching the output of the digital audio signal converted into a code by the A / D converter 11 and a storage unit 2 for storing the digital audio signal input from the switching unit 1. An appearance frequency detection unit 3 for reading the digital audio signal from the storage unit 2 and detecting an appearance frequency of each code representing a quantization level for a plurality of samples; and the code detected by the appearance frequency detection unit 3 From the appearance frequency of each A / D converter 11, the distortion based on the linearity error is normalized with the appearance frequency. Error bit detecting means 5 for detecting an error bit from the code of the normalized appearance frequency detected by the normalizing means 4, and the digital audio signal input from the switching means 1. And a digital signal processing unit 6 for correcting and outputting the error bit detected by the error bit detection unit 5.

More specifically, the switching means 1 is composed of, for example, a switch, reads a digital audio signal from the recording medium 12 via a digital audio interface, and stores the read digital audio signal in a storage means 2 such as a hard disk or the like, which will be described later. The output of the digital audio signal is switched so as to be given to the digital signal processing means 6. For example, the switching unit 1 initially supplies the digital audio signal read from the storage medium 12 to the storage unit 2 as shown in an example described later, and the storage unit 2, the appearance frequency detection unit 3, the normalization unit 4, and the error bit detection. After the error bit included in the digital audio signal is detected by the means 5, the digital audio signal is read out again from the storage medium 12 and supplied to the digital signal processing means 6. The distortion contained in the signal is corrected based on the error bit detected by the error bit detection means 5.

When focusing on the function of the switching means 1, the storage means 2, the appearance frequency detection means 3, the normalization means 4,
The error bit detection means 5 functions as a conversion error detection means, and the digital signal processing means 6 functions as a conversion error correction means.

Here, the function as the conversion error detecting means (storage means 2, appearance frequency detecting means 3, normalizing means 4, error bit detecting means 5) will be described first.

As the storage unit 2, a device such as a hard disk capable of storing data is used, and a digital audio signal provided from the switching unit 1 is stored.

Further, the appearance frequency detecting means 3 comprises a storage means 2
Is read out, and the frequency of occurrence of the code is detected for each code representing the quantization level of the read out digital audio signal for a plurality of samples. That is, for digital audio signals for a plurality of samples, a histogram (level distribution) representing the frequency of appearance for each code is created.

Here, a code representing the quantization level of the digital audio signal will be described. The quantization level of the digital audio signal uses a binary system of 1 and 0,
It is represented by multiple bits. By increasing the number of bits, a finer amplitude level can be expressed. For example, if the amplitude level represented by 13 bits in the binary system is converted to the decimal system, it can be represented in the range of -4096 to 4095. These -4096 to 4096 are represented as codes. A code of 0 indicates that the amplitude level of the analog signal is 0, and indicates that as the absolute value of the code increases, the amplitude level of the analog signal increases.

Next, a histogram (level distribution) of a digital audio signal will be described. FIG. 2 is a diagram illustrating an example of a histogram (level distribution) of a digital audio signal. The appearance frequency in the histogram shown in FIG. 2 is substantially symmetric in the positive direction and the negative direction with the code 0 as the center. When the code is 0, the appearance frequency is high, and as the code increases in the positive direction, Also, as the code becomes smaller in the negative direction, the appearance frequency gradually decreases. Such a feature of the histogram is unique to each input audio signal. In this histogram, a code whose appearance frequency rapidly increases or a code whose appearance frequency rapidly decreases with respect to an adjacent code is a code in which distortion based on a conversion error has occurred.

A method of calculating a histogram in the appearance frequency detecting means 3 will be described. The appearance frequency detection means 3
It has a histogram memory that stores a value corresponding to the code itself as an address using the code itself representing the quantization level. The appearance frequency detection means 3 is a storage means 2
, A digital audio signal is read out, a histogram memory is accessed based on the code, and a value corresponding to an address stored in the histogram memory is read out. 1 is added to this value, and the value obtained by adding 1 is written at the same address. This process is repeated for a plurality of samples. As a result, the cumulative number of times for each code, that is, the frequency of appearance is obtained and stored in the histogram memory.

The number of samples to be processed is not all of the digital audio signals whose distortion based on the conversion error is to be reduced, but if they are converted into digital audio signals by the same A / D converter. It may be a part of the digital audio signal, for example a sample for the first few seconds. However, the greater the number of samples, the better the accuracy of the histogram. For example, if 2,500,000 samples in which a relatively high level portion is continuous are used, a favorable histogram distribution can be obtained. This histogram can be realized by a digital counter circuit or software.

Next, the normalizing means 4 will be described. Since the frequency of appearance of the digital audio signal input as described above represents the frequency of occurrence unique to the digital audio signal, the normalization means 4 more specifically includes, for example,
By dividing the frequency of occurrence of the distorted code based on the conversion error by the average frequency of occurrence in a plurality of adjacent codes including the distorted code based on the conversion error,
The distortion based on the conversion error of the digital audio signal can be extracted as a feature of the frequency of appearance.

More specifically, all codes are divided into a plurality of blocks with a plurality of codes having adjacent quantization levels, for example, four codes as one unit. From the appearance frequency of each code (four codes) constituting this block,
The average frequency of appearance for each block is determined. If the value obtained by dividing the frequency of appearance of each code constituting the block by the average frequency of appearance for each block is the normalized frequency of appearance A of each code, the normalized frequency of appearance A of each code is
It is represented by the following equation.

[0040]

## EQU1 ## Normalized appearance frequency A = (appearance frequency of each code divided into four) / (average appearance frequency of four codes)

The histogram including the normalized appearance frequency A is obtained by extracting a distortion based on the conversion error of the digital audio signal as a feature of the appearance frequency. In other words, the appearance frequency A is obtained by almost eliminating the characteristic of the unique appearance frequency of the digital audio signal.
When there is no distortion based on the conversion error, the normalized appearance frequency A is 1 in all codes. In the above-described example, the normalizing means 4 determines the average of the frequencies of occurrence of the respective codes in the previously divided block as the average frequency of occurrence. The moving average of the frequencies of appearance of the code may be obtained and used as the average frequency of appearance.

As an example of processing in the normalizing means 4,
As a special case, a case where a digital audio signal read from the recording medium 12 includes a sine wave or a digitized music signal close to a sine wave will be described. If such music signals are included,
The normalized appearance frequency can be obtained by the same method as the method of calculating the differential linearity error in the conventional sine wave histogram measuring device. That is, theoretically, a sine wave histogram having no distortion has a smaller frequency of occurrence at the amplitude level 0, and changes the amplitude level 0 as the amplitude level increases in the positive direction or decreases as the amplitude level decreases in the negative direction. The frequency of appearance becomes sharply symmetrical with respect to the center, and the frequency of appearance becomes maximum at the peak levels in the positive and negative directions. The histogram obtained theoretically represents the frequency of appearance unique to the input digital audio signal. Therefore, in this case, by inputting a sine wave or a digital audio signal close to a sine wave without distortion, and dividing the frequency of appearance of each code by the frequency of appearance of the same code obtained from the theoretical value of the sine wave histogram, The normalized frequency of appearance can be obtained.

Next, the error bit detecting means 5 will be described. The error bit detecting means 5 detects error bits from the normalized histogram from the normalizing means 4. Here, the normalized histogram includes a code having an extremely large number of appearances and a code having an extremely small number of appearances. The quantization levels of the code having a large number of occurrences and the code having a small number of occurrences are included in the normalized histogram. The difference represents the weight of the bit in which the distortion based on the conversion degree difference has occurred, in other words, the digit position of the bit.

The cause of the conversion error is that the audio signal is A / A
When digitized by the D converter 11, A /
It depends on the conversion method of the D converter 11. here,
Successive approximation type A / D using a weighted resistance type A / D converter or a binary weighted current source type D / A converter having an R-2R ladder as an internal D / A converter
Assume a converter. That is, it is assumed that a weight is used for setting the amplitude level as the threshold. In other words, paying attention to the fact that the influence of a certain weight spreads to a plurality of quantization levels, it detects a plurality of quantization levels, that is, weights that cause distortion based on a conversion error from the frequency of occurrence of a code. I have to.

[0045] For example, the spacing between the occurrence frequency is extremely large code 2, i.e. when the distance between the quantization level is 2 ¹ times the quantization level of the least significant bit, that there is an error in the least significant bit Recognize. The distance between the occurrence frequency is extremely large code 2 ^3, i.e. when the distance between the quantization level is 2 ³ times the quantization level of the least significant bit, the digit position of the third bit from the least significant bit It can be seen that there is an error in the bit. That is, if the interval between the quantization levels of a plurality of codes having extremely large numbers of appearances is 2 ⁿ , it can be seen that there is an error in the bit at the n-th digit position from the least significant bit.

To determine that the frequency of occurrence of a code is extremely large, a threshold value is required. For example, if the normalized appearance frequency A is 1.2 or more, it is determined that the appearance frequency is extremely large. In addition, the digit position of an erroneous bit can be detected from the interval between codes having extremely small appearance frequencies. Further, a bit having an error may be detected by using both of them.

The error bit detection processing by the error bit detection means 5 will be described. FIG. 3 is a flowchart showing an example of the error bit detection processing of the error bit detection means 5.

In the processing example of FIG. 3, first, the normalized histogram is input from the normalizing means 4 (step S).
11). Next, the initial value of the value of the variable n is a value obtained by subtracting 1 from the number of quantization bits of the digital audio signal.
(Step S12). Next, the normalized appearance frequency of each code is extracted for every 2 ⁿ quantization levels of each code.
(Step S13). Then, it is determined whether or not all occurrence frequencies extracted for a certain value of n are 1.2 or more (step S14). If the frequency of occurrence is 1.2 or more, it is determined that there is an error at the nth bit from the least significant bit, and the value of n is stored in the error bit memory (step S15). If the frequency of appearance is less than 1.2,
The value of n is updated by subtracting 1 from the value of n (step S16). It is determined whether the value of n is 0, and the value of n is 0
If not, the process returns to the process of extracting the normalized appearance frequency of each code for every 2 ⁿ quantization levels of each code (step S17). If the value of n is 0, the error bit stored in the error bit memory is output to the digital signal processing means 6 (step S18). Through the above processing, an error bit can be detected.

FIG. 4 is a diagram showing an example of a level distribution of a music signal using a recent accurate A / D converter.
FIG. 5 is a diagram illustrating an example of the level distribution of the initial PCM music signal. Both use the same music signal,
0000 samples (sampling frequency 44.1KH
z, about 1 minute). The vertical axis is the frequency of appearance,
The horizontal axis is the LSB unit in 13 bits. Music signal
The level distribution (for a long time) is close to a normal distribution centered on 0 (LSB), and FIG. 4 shows this tendency. FIG. 5 shows a shape almost similar to a normal distribution.
There are levels at which the frequency of appearance is extremely high and levels at which the frequency is extremely low. This is due to the error of the successive approximation A / D converter. In an ideal A / D converter, for one output level, the analog input voltage is ±
It has a width of 0.5 (LSB). However, in an actual successive approximation A / D converter, the width is not constant because there is an error in the internal D / A converter and the like. Therefore, when the width of the input analog voltage of the A / D converter is larger than ± 0.5 (LSB), the data contains data that should be converted to an adjacent level. Becomes large, and conversely, when the width of the input analog voltage of the A / D converter is smaller than ± 0.5 (LSB), the appearance frequency of the level becomes small. Therefore, if the level distribution of the music signal is used, the A / D
An error of the converter can be detected.

As described above, the conversion error detecting means (the storage means 2, the appearance frequency detecting means 3, the normalizing means 4, and the error bit detecting means 5) use an analog / digital converter having a linearity error. From the converted digital audio signal (from the already recorded music signal), it is possible to estimate (detect) distortion based on the linearity error of the analog / digital converter at the time of analog / digital conversion.

Next, the digital signal processing means 6 as a conversion error correction means will be described. After the conversion error (error bit) at the time of A / D conversion included in the digital audio signal is detected as described above, the digital signal processing unit 6 receives the same digital audio signal again from the switching unit 1, for example. The digital signal processing unit 6 corrects the distortion based on the conversion error of the digital audio signal based on the conversion error (error bit) at the time of the A / D conversion detected by the error bit detection unit 5, and corrects the distortion. Is to be reduced. That is, among a plurality of samples having a quantization level whose appearance frequency is increased by the distortion based on the conversion error, the quantization level of the sample in which the distortion based on the conversion error occurs is estimated to be the original correct quantization level. Correction to a quantization level is performed to reduce distortion based on a conversion error.

FIG. 6 is a flowchart showing an example of the correction processing in the digital signal processing means 6.

In the processing example of FIG. 6, when the code of each sample of the digital audio signal to be subjected to the error reduction is provided from the switching means 1 (step S21), the digital signal processing means 6 sets the error for each sample. It is determined whether or not the value at the bit position of the bit is 1 (step S2).
2). If the value at the digit position of the error bit is a code of 1 (a code whose quantization level is affected by the error bit), the correct quantization level of the sample, that is, the code is estimated (step S23). If the value at the digit position of the error bit is a code other than 1,
The input sample code or the code estimated in the above processing is written to a file and output (step S24). At this time, data indicating the position on the time axis, that is, the number of the sample from the beginning is also output to the file together with the code. By the above processing, the code in which the distortion has occurred based on the conversion error can be corrected to the original code that is estimated to be correct, and the distortion based on the conversion error can be reduced.

Here, correction of a code affected by distortion based on a conversion error will be described. LSB is the quantization level of the least significant bit, a and b are positive integers, and N is 0 or a positive integer.

For example, when the resistance value in the successive approximation type A / D converter has an error of + b · LSB at the ath bit from the lower order, the code of the quantization level becomes 2 ^a · N + 2 ^a−1 −b occurrence frequency is increased to b + 1 times, the occurrence frequency of each code between quantization levels from 2 ^a · N + 2 ^a -1 to 2 ^a · N + 2 ^a -b are both zero.

Further, the −b · LSB at the ath bit from the lower order
, The quantization level is 2 ^a · N + 2 ^a−1 −1
Increases by b + 1 times and the quantization level is 2 ^a ·
The frequency of appearance of each code from N + 2 ^a-1 -1 to 2 ^a N + 2 ^a-1 + b-1 is zero.

Therefore, when the correction is performed so that such an error is canceled, the distortion based on the conversion error is reduced. However, sample data that must be a code with a low frequency of occurrence is erroneously converted into a code with a high frequency of occurrence. In other words, a plurality of samples having a quantization level with a large number of appearances include a sample that should originally have another quantization level.

In order to correct the erroneous conversion, first, a plurality of codes having an extremely large number of appearances affected by one error bit detected by the error bit detecting means 5 are obtained. Let X be the average of the frequencies of occurrence of the detected codes. A sample having a code whose appearance frequency is extremely increased under the influence of the error bit is converted into one or a plurality of codes whose appearance frequency is extremely reduced by the one error bit in accordance with the value of X. For example, conversion is performed at a rate of one to X pieces. If there are errors in a plurality of bits, the above processing is performed for each bit.

The error bit detection means 5 determines whether the quantization level affected by the detected error bit should be corrected to positive or negative depending on the quantization level difference between adjacent samples on the time axis. ing.

A process for estimating a code without distortion based on a conversion error will be described. FIG. 7 is an A / D conversion characteristic diagram of the analog / digital converter (A / D converter) 11. For a code having a value of 1 at the digit position of the error bit, the internal file is read to check how many samples the code continues on the time axis. For example, when it is determined that there is an error of b = + 1 · LSB at the a = 3th bit from the lower order, the input / output characteristics of the A / D converter are as shown in FIG. That is, at 8n + 3, the digital output does not change with an increase in the analog input corresponding to + 1 · LSB. Therefore, at 8n + 3, the frequency of appearance becomes extremely small. At 8n + 6, the digital output becomes 8n + 8 with respect to the increase of the analog input corresponding to + 1 · LSB, and the appearance frequency increases.

FIG. 8 is a diagram for explaining digital output on the time axis in the distortion correction apparatus of the present invention. FIG.
In order to reduce the distortion based on the conversion error of the digital audio signal, all the digital outputs corresponding to 8n + 4 to 8n + 6 are converted to a value larger by + 1 · LSB. When a sample of the digital output corresponding to 8n + 3 and _{K 1, K 2, K 3} ... K t, if a sample corresponding to 8n + 3 lasted t Sample 2 Sample least, K
The difference between the digital output of ₁ and the sample K ₀ one sample before is taken, and this difference is defined as D ₁ . Further, the difference between the digital output of K _t and the sample K _{t + 1} after one sample is calculated,
This difference and D _2. If the differences D ₁ and D ₂ are both positive,
Using the values of the X, from K _t / _X the value of up to K _t is +1, if the difference D _1, D ₂ are both negative, using the values of the X, from K ₁ to K _t / _X Is set to +1. Where t
If the value of / X is not an integer, the value is rounded down or rounded up so that the value of t / X is an integer.

By doing so, the sample of the digital output corresponding to 8n + 3 at a rate of once every X times is:
The value is converted to a value larger by + 1 · LSB. Which sample is to be transformed is determined by the difference D ₁ , D ₂ between the quantization levels of adjacent samples on the time axis. In the above description, the case where it is determined that there is an error of b = + 1 · LSB has been described. However, even if b is any other value, it is also possible to estimate a code without distortion based on the conversion error. it can.

The linearity error reduction processing of the A / D converter having the above configuration will be described with reference to FIG. When a digital audio signal is read from the recording medium 12 and input to the switching unit 1, the switching unit 1 switches the input digital audio signal to be output to the storage unit 2.

The storage means 2 stores the digital audio signal from the switching means 1.

The frequency-of-appearance detecting means 3 reads the digital audio signal temporarily stored in the storage means 2 for a plurality of samples, detects the frequency of appearance for each code representing each quantization level, and creates and outputs a histogram. .

The normalizing means 4 extracts a distortion based on the conversion error based on the frequency of occurrence of the code from the histogram output from the frequency-of-appearance detecting means 3 and outputs the result as a normalized histogram. In the digital audio signal, when the distortion based on the conversion error is not originally generated, the value of the appearance frequency of each code is the same, but when the distortion based on the conversion error occurs, the difference occurs in the appearance frequency for each code. .

The error bit detecting means 5 detects an error bit of the digital audio signal from the difference in the frequency of occurrence of each code based on the normalized histogram from the normalizing means 4, and outputs the error bit to the digital signal processing means 6. Output.

Next, the switching means 1 switches the digital audio signal read from the recording medium 12 to be output to the digital signal processing means 6 via the digital interface.

The digital signal processing means 6 receives a digital audio signal read from a recording medium and a detection result of detecting an error bit having a linearity error in the digital audio signal.

The digital signal processing means 6 converts the digital audio signal converted into an erroneous code by the conversion error of the input digital audio signal based on the detection result from the error bit detection means 5 into an original conversion error. The output is corrected to a value close to a level without distortion based on the output. The output digital audio signal is a digital audio signal in which distortion based on the conversion error has been reduced.

As described above, a code converted to an incorrect value due to a conversion error of a digital audio signal can be easily corrected, and distortion based on the conversion error can be reduced.

In the above-described processing, as a method of correcting distortion based on a conversion error, an error bit determined by the error bit detecting means 5 is used for a continuous sample having a code affecting the quantization level. ,
Various interpolation techniques used when estimating the quantization level of a certain sample from the quantization levels of adjacent samples on the time axis can be used.

In the above-described distortion correcting apparatus, only the conversion error detecting means (storage means 2, appearance frequency detecting means 3, normalizing means 4, error bit detecting means 5) is constituted as a single apparatus (distortion detecting apparatus). You can also. That is, according to the present invention, the A / D included in the digital audio signal is
It can also be configured as a distortion detection device that only detects a conversion error (distortion) at the time of conversion.

In the above configuration example, the normalizing means 4 and the error bit detecting means 5 obtain a histogram from the frequency of occurrence of a code representing the quantization level of the recorded music signal (digital audio signal), and obtain the histogram. The digital signal processing means 6 corrects the error bit in which the A / D conversion error has occurred, and corrects the error bit code based on the preceding and following codes. (I.e., the unevenness of the occurrence frequency equal to or higher than the threshold value in the histogram is determined as an error bit and corrected based on the codes before and after), so that the unevenness of the occurrence frequency equal to or lower than the threshold value is an error bit. Since it is not corrected without judging that the A / D conversion error occurs, the A / D conversion error may not be significantly reduced.

The present inventor has further devised the following configuration example in order to correct the unevenness of the frequency of occurrence below the threshold value and to significantly reduce the A / D conversion error.

FIG. 9 is a diagram showing another example of the configuration of the distortion correction apparatus according to the present invention. Note that, in FIG. 9, the same portions as those in FIG. 1 are denoted by the same reference numerals. Referring to FIG.
The distortion correction device 20 is a digital audio signal (music signal) converted into a code by the A / D converter 11.
Switching means 1 for switching the output of the digital audio signal, storage means 2 for storing the digital audio signal input from the switching means 1, and reading out the digital audio signal from the storage means 2 to represent each quantization level for a plurality of samples. Histogram of frequency of appearance for each code (level distribution)
And a low-pass filtering process is performed on the level distribution detected by the occurrence frequency detecting means 3 to approximate an ideal level distribution when there is no error in the analog / digital converter. A normalization means 34 for detecting, as a normalized level distribution, a value obtained by dividing the level distribution before the low-pass filtering is performed by the result of approximation to the ideal level distribution, and Conversion table calculating means 35 for calculating a conversion table T (i) for the level of the digital audio signal from the level distribution obtained and storing it in a predetermined memory
And data replacement means 36 for replacing the level of the digital audio signal read from the recording medium 12 with level data according to the conversion table T (i) and outputting the data.

That is, in the configuration example of FIG. 9, the normalizing means 34 has a low-pass filtering function, and the histogram (level distribution) created by the appearance frequency detecting means 3
Is subjected to a low-pass filtering process to approximate this histogram (level distribution) to an ideal histogram (level distribution), that is, an ideal histogram (level distribution) when the A / D converter 11 has no error. Histogram before low-pass filtering
By dividing (level distribution) by the histogram (level distribution) after low-pass filtering, the conversion error of the A / D converter 11 was normalized to an appearance frequency of 1.0 when there was no error. The distribution is detected.

More specifically, a digital audio signal
Assuming that the appearance frequency of the level i of the (music signal) is g (i), the number of quantization bits is n, and the number of samples for averaging the low-pass filter (moving average filter) is m (m: odd number), the normalizing means 34 Calculates the normalized value N (i) as in the following equation.

[0079]

(Equation 2)

Here, c (j) is a window function satisfying the following equation.

[0081]

(Equation 3)

In other words, g (i) is a window function c (j)
Is shown. In addition, since the moving average method cannot calculate both ends of the smoothed waveform, the both ends of the smoothed waveform are portions where the music level is high, and such a large level actually exists. Therefore, in the above equation (Equation 2), it corresponds to samples at both ends of the smoothed waveform.
(m-1) / 2 is excluded from the range of the code i.

As described above, the normalizing means 34 of the distortion correction apparatus 20 shown in FIG. 9 uses the moving average filter for the appearance frequency g (i) of the digital code i of the digital audio signal (music signal), By calculating the average value of the digital code i, the appearance frequency g (i) of the digital code i of the digital audio signal (music signal) is approximated to the ideal appearance frequency,
Normalization is performed, whereby the A / D
Even when the level distribution of the converted digital audio signal (music signal) has an inherent bias depending on the type of musical sound, an ideal level distribution can be calculated. That is, normalization can be performed.

The conversion table calculating means 35 specifically integrates the normalized value N (i) obtained as described above by the normalizing means 34 as in the following equation (Equation 4). , A conversion table T (i) is obtained and stored in, for example, a table conversion memory.

[0085]

(Equation 4)

The conversion characteristics (input / output characteristics) of an ideal A / D converter are as shown in FIG. But,
When there is a conversion error in the A / D converter, the conversion characteristics (input / output characteristics) of the A / D converter are as shown in FIG.

In FIG. 11, the level b has a high frequency of appearance, and the level c has a low frequency of appearance. The normalized level distribution N (i) indicates the input level width (A to F) of each level (a to f). Then, for example,
The level b is the median of the input level width B, and N (b) =
(A + B) / 2 is a representative value. Similarly, level c is N
(c) = (A + B) / 2 + (B + C) / 2, and the level d is N (d) = (A + B) / 2 + (B + C) / 2 + (C +
D) / 2. Equation 4 represents such an integration process of N (i). If the conversion table T (i) = N (i), the correct value of T (i) cannot be obtained because the value of N (i) is not integrated. That is, N (i) corresponds to the width of the stairs as shown in FIG. 12, and T (i) to be obtained is the median value of the stairs shown in FIG. However, T (i) = N
In the case of (i), since N (i) is not integrated, the result is as shown in FIG. In the case of Equation 4, since N (i) is integrated, a correct T (i) can be obtained.

Then, the data replacement means 36 calculates a value T (i) corresponding to the level i of the digital audio signal read from the recording medium 12 from the conversion table obtained as described above, and (i) can be output as a distortion-corrected digital audio signal. Specifically, the conversion table T (i) calculated by the conversion table calculating means 35 stores the value corresponding to the address with the code i itself representing the quantization level in the table conversion memory. In this case, the data replacement means 36
Digital audio signal read from recording medium 12
(Code representing quantization level) i is an address, and a value T (i) corresponding to the address (level i) is read out from the table conversion memory, and this is output as a digital audio signal corrected for distortion. it can.

As described above, in the configuration example of FIG. 9, the conversion error of the used analog / digital converter is detected from the level distribution of the digital audio signal recorded using the analog / digital converter. A conversion table is calculated from the detected conversion error, a table conversion process is performed on the digital audio signal, and distortion generated by the conversion error of the analog / digital converter is corrected. Is also corrected, and A /
D conversion error can be significantly reduced.

Further, in the configuration example of FIG. 9, for example, by making the bit length of the conversion table T (i) larger than the bit length of the input data, the bit length is extended (for example,
Extension from 13 bits to 16 bits) is also possible.

Specifically, assuming that the possible value of the 13-bit signal is −2 ¹³⁻¹ to 2 ¹³⁻¹ −1 (−4096 to +4095), the signal after the table conversion has a value after the decimal point calculated by the operation. Occur. For example, if the level distribution g (i) is g (0) = 100 g (1) = 150 g (2) = 50 g (3) = 100, the level distribution h after passing through the low-pass filter is
Let (i) be h (0) = h (1) = h (2) = h (3) = 100. The level distribution h (i) is obtained from the formula of the low-pass filter as in the following equation.

[0092]

(Equation 5)

From N (i) = g (i) / h (i), N (0) = 100/100 = 1.0 N (1) = 150/100 = 1.5 N (2) = 50/100 = 0.5 N (3) = 100/100 = 1.0

From equation (4), N (0) = 0.0 N (1) = 0.0 + (1.0 + 1.5) /2=1.25 N (2) = 1.25 + (1.5 + 0 .5) /2=2.25 N (3) = 2.25 + (0.5 + 1.0) /2=3.0, and the value after the decimal point is generated by the table conversion. The value after the decimal point is the lower bit, and it can be converted into a signal of an arbitrary bit length depending on the value to be output.

FIG. 15 shows a part of the input / output characteristics before and after the table conversion. In FIG. 15, the horizontal axis indicates input (analog), and the vertical axis indicates output (input / output characteristics of the A / D converter are 13 bits, and input / output characteristics after table conversion are 16-bit LSB units). 100Hz,
Output when a -40 dB sine wave is input to a successive approximation A / D converter (bit length 13 bits) having a large error, and a digital audio signal recorded by the same A / D converter for the output signal Using the conversion table calculated from the (music signal), an F1920 (Fast Fourier Transform) process of a Hanning window is performed on the data obtained by performing the above-described table conversion process, and a plot is shown in FIG. . FIG. 17 shows a case where the above processing according to the present invention is not performed. 16 is compared with FIG. 17, the case where the table conversion processing is performed (FIG. 16) does not perform the table conversion processing (FIG. 1)
It can be seen that the distortion is reduced as compared to 7).

Next, the processing operation of the distortion correction apparatus of FIG. 9 will be described with reference to the flowcharts of FIGS. FIG. 18 shows a conversion table calculation process (conversion error detection process).
FIG. 19 is a flowchart showing an operation example of a digital audio data table conversion process (conversion error correction process).

When this apparatus is used, the bit length of the input and output digital audio signals (music signals) may be arbitrary, but in this example, the bit lengths of the input and output music signals are 13 bits and It is 16 bits. First, switching means (for example, a switch) 1 is switched to the storage means 2 side,
The conversion table is calculated by the storage unit 2, the appearance frequency detection unit 3, the normalization unit 34, and the conversion table calculation unit 35. That is, the number of samples (measured samples) a of the digital audio signal for creating a histogram (level distribution) is set to a predetermined number (step S31).
Next, the digital audio signal (1
(3 bits) is sampled and read out with the number of samples a, and is input to the storage means 2 (step S32).
The music signal having the bit length of 13 bits input and stored in the storage means 2 is subjected to measurement of a histogram (level distribution) by the appearance frequency detection means 3. In this case, the appearance frequency detection means 3 has a histogram memory, as described above, increments the value of the histogram memory at the address (address) corresponding to the input data by one, and increments the value by a predetermined number (the number of samples a). ) Repeat to create a histogram (level distribution) (steps S33 to S35). It should be noted that if the number a of the samples taking the level distribution of the digital audio signal is not large enough, the accuracy will be low as statistical data, and if it is unnecessarily large, the processing will take a long time. Must be set to a value close to That is, the number a of the samples for measuring the level distribution is preferably as large as possible, but it takes more time than necessary.

Next, the normalizing means 34 performs an arithmetic operation of Equations 2 and 3 on the histogram (level distribution) created by the appearance frequency detecting means 3 and outputs a normalized value N (i). (Step S36). The normalizing means 34
In the low-pass filtering processing in the above, the number m of samples of the moving average filter is preferably set to an optimum value as much as possible. 20A shows a case where the number m of samples is "1" (when there is no filter), FIG. 20B shows a case where the number m of samples is small, and FIG. FIG. 20D is a diagram showing a level distribution when the number m of samples is large. When the number m of samples of the moving average filter is smaller than the optimum value, FIG.
In the case where the unevenness of the level distribution remains as shown in (b) and the number of samples m of the moving average filter is larger than the optimum value, the value near 0 (LSB) becomes smaller as shown in FIG. As a result, the distortion does not decrease as in the case where m is an optimum value (as in the case of FIG. 20C).

Next, the conversion table calculating means 35 outputs the output from the normalizing means 34, that is, the normalized data N (i).
, The conversion table T (i) is calculated by performing the operation of Equation 4 (the integration processing of the normalized data N (i) is performed), and the conversion table T (i) is output to the table conversion memory with a bit length of 16 bits ( Step S37).

FIGS. 21A to 21D show the outline of the above processing. Here, FIG. 21 (a) is a diagram showing a level distribution measured by the appearance frequency detecting means 3, and FIG. 21 (b) is a level obtained by performing a low-pass filtering process on the level distribution of FIG. Diagram showing distribution, FIG. 21 (c)
Are the data normalized by the normalizing means 34 (FIG. 21)
FIG. 21D shows data obtained by dividing the level distribution of FIG. 21A by the smoothed level distribution of FIG. 21B), and FIG. 21D shows the normalized data of FIG. Normalized frequency N
It is a figure which shows the conversion table created by integrating (i) by Formula 4.

After the conversion table T (i) is calculated in this way and stored in the table conversion memory, the switching means 2 is switched to the data replacing means 36, and the digital audio signal read from the recording medium 12 is read. Performs data conversion according to the conversion table T (i). That is,
The data replacement means 36 outputs an address corresponding to the digital audio signal (level i) read from the recording medium 12.
The operation of outputting the value T (i) of the (address) table conversion memory is repeated a predetermined number of times.

More specifically, as shown in FIG. 19, for example, a 13-bit digital audio signal read from the recording medium 12 is sequentially input to the data replacing means 36 (step S41). And outputting the value of the table conversion memory corresponding to the level of the input digital audio signal (step S43).
Until EOF (data indicating the end of the data) is detected (step S42), the processing is repeated, and the digital audio signal is converted into a 16-bit digital audio signal with reduced distortion and output. As described above, the music signal having a bit length of 13 bits read from the recording medium 12 is subjected to table conversion by the data replacing means 36, and the bit length in which distortion based on the conversion error of the A / D converter 11 is reduced. 1
It is converted into a 6-bit music signal and output.

As described above, in the distortion correction apparatus of FIG. 9, when creating the conversion table T (i) from the histogram (level distribution) of the digital audio signal (music signal), the old A /
The unevenness of the level distribution due to the error of the D converter is smoothed by the moving average filter, and approximates the ideal level distribution when the A / D converter has no error. It is also possible to correct irregularities having a frequency of occurrence equal to or less than the threshold value, thereby significantly reducing A / D conversion errors.

As described above, in each of the above-described configuration examples, when the A / D converter used for recording the music source does not exist or the conversion characteristics are changed, the A / D converter is switched from the old recording source to the A / D converter. It is possible to estimate the state of the original music source having no D-conversion error, replace the input data with data estimated to be correct, and correct the music source having the A / D conversion error.

In the configuration example of FIG. 9, the A / D conversion error is estimated using a music signal (digital audio signal). However, the A / D converter used at the time of recording the music source does not change the conversion characteristic. In the existing case, a linear count-up signal is input to the A / D converter, an input signal of the A / D converter is compared with an output signal, and an A / D conversion error can be obtained. The correction using such a linear count-up signal can be performed as follows.

That is, for example, as shown in FIG.
Using an old A / D converter measurement system (for example, a system as shown in FIG. 23), a linear count-up signal and a sine wave are fetched into a personal computer, a conversion table is calculated from the level distribution of the linear count-up signal, and a sine wave is calculated. Table transformation and FFT (Fast Fourier Transform) are performed on the data to examine the increase or decrease in distortion. The conversion table calculation processing and the table conversion processing are performed, for example, by creating a program in the C language and processing the data inside the personal computer. For the FFT, commercially available signal processing software can be used.

Here, the process of calculating the conversion table can be performed, for example, by the following method. That is,
The appearance frequency of the digital code i (−2 ⁿ⁻¹ ≦ i ≦ 2 ⁿ⁻¹ −1) when the number of quantization bits is n and the total number of samples is sc is h (i),
Let N (i) be the normalized appearance frequency. To facilitate the calculation of the table, h (i) is normalized such that N (i) = 1.0 when h (i) = sc / 2 ⁿ as in the following equation. .

[0108]

N (i) = (2 ⁿ / sc) · h (i)

Next, N (i) is integrated as shown in Equation 4 to obtain a table T (i).

In actual processing, data before table conversion
(13 bits) and data after table conversion (16 bits)
However, since the relationship is as shown in FIG. 24, the table value T (i) is multiplied by 2 ³ = 8 to 16 bits.

As described above, the correction using the linear count-up signal is effective, for example, when a music signal currently recorded at 16 bits is converted to a music signal having a bit length of 24 bits or more in the future. is there. The linear count-up signal may be recorded on a recording medium or the like at the time of recording the music signal. In such a case, even if the A / D converter does not exist, the linear count-up signal is more accurately based on the conversion error. Distortion can be reduced.

In the above-described processing described with reference to FIG. 1 or FIG. 9, the switching means 1 outputs the digital audio signal from the recording medium 12 to the storage means 2 for processing for detecting a conversion error, and then outputs the same digital signal again. The audio signal is read from the recording medium 12 and the digital signal processing means 6
Alternatively, the data is output to the data replacing means 36, but the present invention is not limited to this processing.
For example, all the digital audio signals whose distortion is to be reduced based on the conversion error are stored in the storage unit 2 via the switching unit 1, and then the output from the storage unit 2 is directly transmitted via the switching unit 1. The configuration may be applied to the digital signal processing means 6 or the data replacement means 36.

Further, for example, a delay means is provided before the digital signal processing means 6 or the data replacing means 36, and the switching means 1 transmits the digital audio signal read from the recording medium 12 to the storage means 2 and the delay means simultaneously. During detection of the conversion error, the digital audio signal to be corrected can be delayed by the delay means so that the digital signal processing means 6 or the data replacing means 36 can perform correction output processing in real time. .

When it is not necessary to output a digital audio signal with reduced distortion based on a conversion error in real time, for example, when storing a digital audio signal with reduced distortion in another storage device A hard disk or the like having a high data transfer rate is used as the storage means 2, and the output of the storage means 2 is directly switched via the switching means 1 to the digital signal processing means 6 or the data replacement means 3.
6 may be input.

In the above example, the digital audio signal is recorded on the recording medium 12 such as a compact disk, and the digital audio signal is read from the recording medium 12. However, the digital audio signal to be corrected is The digital audio signal can be stored in a storage device including a semiconductor memory such as a ROM and a RAM. In this case, a digital audio signal is read from the storage device. In this case, the storage device itself may be controlled so that the storage device itself also has the function of the storage device 2. In this case, the output from the storage device is further directly input to the appearance frequency detection device 3. Can be done.

In other words, in the present invention, the term "storage means" should be understood in a broad sense, and the storage means is not limited to the storage means 2 of FIG. 1 or FIG. It can also refer to a RAM or the like.

[0117]

As described above, according to the first to eighth aspects of the present invention, the distortion based on the linearity error of the analog / digital converter is generated from the appearance frequency of each code of the digital audio signal. Bit, and based on the result of the detection, correct the bit in which the distortion has occurred due to the linearity error of the analog / digital converter of the digital audio signal. The distortion based on the linearity error of the digital converter can be reduced.

According to the fourth aspect of the present invention, there is provided a histogram memory for storing a code representing a quantization level of a digital audio signal as an address, and an address corresponding to the code of the digital audio signal read from the storage means. Is read from the histogram memory, 1 is added, and the data obtained by adding 1 is written to the address memory, and a histogram representing the frequency of appearance for each code is created. Even if the signal source is not a sinusoidal wave, the frequency of appearance for each code can be detected.

According to the fifth aspect of the present invention, all codes are divided into a plurality of blocks by dividing a plurality of codes having quantization levels close to each other into one block based on the appearance frequency of each code. , And the appearance frequency obtained by dividing each code in each block by the average appearance frequency can be obtained as the normalized appearance frequency, so that the distortion based on the linearity error of the digital audio signal is detected as the appearance frequency for each code. can do.

Further, according to the invention of claim 6, based on the normalized frequency of occurrence detected by the linearity error detecting means, the difference of the quantization level for each code is determined by the analog / digital converter. It can be detected as the number of bits in which distortion based on the linearity error has occurred.

According to the seventh aspect of the present invention, the error bit of the input digital audio signal detected by the error bit detecting means is replaced with a code estimated to be correct, and is output. , The distortion based on the linearity error of the analog / digital converter can be reduced.

According to the ninth to eleventh aspects of the present invention, the conversion error of the used analog / digital converter is determined from the level distribution of the digital audio signal recorded using the analog / digital converter. , A conversion table is calculated from the detected conversion error, and a table conversion process is performed on the digital audio signal to correct the distortion caused by the conversion error of the analog / digital converter. Can be corrected, and A / D conversion errors can be significantly reduced.

In other words, according to the first to eleventh aspects of the present invention, even if the characteristics of the used analog / digital converter are not known, or if the used analog / digital converter itself does not exist. However, the error of the analog / digital converter can be detected from the recorded digital audio signal, and the distortion of the digital audio signal is corrected based on the detected error of the analog / digital converter. The distortion of the audio signal can be reduced, and the sound quality can be improved.

According to the twelfth aspect of the present invention, a linear count-up signal is input to an analog / digital converter, and an input signal and an output signal of the analog / digital converter at this time are compared. The conversion error of the analog / digital converter is calculated, and the distortion based on the conversion error included in the output signal from the analog / digital converter is corrected by the conversion error calculated as described above.
If the A / D converter used at the time of recording the music source is present without any change in the conversion characteristics, for example, a music signal currently recorded at 16 bits is converted to a music signal having a bit length of 24 bits or more in the future. Can be converted.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration example of a distortion correction device according to the present invention.

FIG. 2 is a diagram illustrating an example of a histogram (level distribution) of a digital audio signal.

FIG. 3 is a flowchart illustrating an example of an error bit detection process.

FIG. 4 is a diagram showing an example of a level distribution of a music signal using a recent accurate A / D converter.

FIG. 5 is a diagram showing an example of the level distribution of an initial PCM music signal.

FIG. 6 is a flowchart illustrating an example of a correction process in a digital signal processing unit.

FIG. 7 is an analog / digital conversion characteristic diagram of the analog / digital converter.

FIG. 8 is a schematic diagram showing a digital output on a time axis in the distortion correction device of the present invention.

FIG. 9 is a diagram showing another configuration example of the distortion correction device according to the present invention.

FIG. 10 is a diagram showing conversion characteristics of an ideal A / D converter.

FIG. 11 shows A when the A / D converter has a conversion error.
FIG. 4 is a diagram illustrating conversion characteristics of a / D converter.

FIG. 12 is a diagram for explaining how to calculate a conversion table.

FIG. 13 is a diagram for explaining how to calculate a conversion table.

FIG. 14 is a diagram for explaining how to calculate a conversion table.

FIG. 15 is a diagram showing a part of input / output characteristics before and after table conversion.

FIG. 16 is a diagram illustrating a result of performing a table conversion process on a digital audio signal.

FIG. 17 is a diagram illustrating a case where a table conversion process is not performed on a digital audio signal.

FIG. 18 is a flowchart illustrating a processing operation of the distortion correction device in FIG. 9;

FIG. 19 is a flowchart illustrating a processing operation of the distortion correction device in FIG. 9;

FIG. 20 is a diagram showing a level distribution when the number m of samples of the moving average filter is changed.

FIG. 21 is a diagram showing an outline of processing up to table conversion in the distortion correction device of FIG. 9;

FIG. 22 is a diagram illustrating a correction process using a linear count-up signal.

FIG. 23 is a diagram illustrating a correction process using a linear count-up signal.

FIG. 24 is a diagram showing a relationship between data (13 bits) before table conversion and data (16 bits) after table conversion.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Switching means 2 Storage means 3 Occurrence frequency detection means 4,34 Normalization means 5 Error bit detection means 6 Digital signal processing means 10,20 Distortion correction device 11 A / D converter 12 Recording medium 35 Conversion table calculation means 36 Data replacement means

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification number Agency reference number FI Technical indication // G01R 31/00 G01R 31/00

Claims (12)

[Claims] A storage means for storing a digital audio signal; an appearance frequency detection means for reading the digital audio signal from the storage means and detecting an appearance frequency for each code representing a quantization level for a plurality of samples; A normalization means for detecting a distortion based on a linearity error of the analog / digital converter as a normalized frequency of occurrence from the frequency of occurrence of each code detected by the frequency of occurrence detection means; An error bit detecting means for detecting an error bit from the code of the normalized occurrence frequency. 2. A conversion error detecting means for detecting an error bit having a distortion based on a linearity error of an analog / digital converter from a frequency of appearance of a digital audio signal converted into a code by an analog / digital converter. And a conversion error correction unit for correcting an error bit in which the digital audio signal has been distorted based on a linearity error of the analog / digital converter based on a detection result from the conversion error detection unit. Characteristic distortion correction device. 3. The distortion correction device according to claim 2, wherein
The conversion error detection means includes a storage means for storing the digital audio signal input from the switching means, and a frequency of appearance for each code representing a quantization level for a plurality of samples by reading the digital audio signal from the storage means. And a normalization for detecting a distortion based on a linearity error of the analog / digital converter as a normalized frequency of occurrence from the frequency of occurrence of each code detected by the frequency of occurrence detection means. And a means for detecting an error bit from a code of the normalized appearance frequency detected by the normalizing means. 4. The distortion correction device according to claim 3, wherein
The appearance frequency detection unit includes a histogram memory that stores a code representing a quantization level of the digital audio signal as an address, and stores the data of the address corresponding to the code of the digital audio signal read from the storage unit. A distortion correction apparatus which reads out from a histogram memory, adds "1", writes the added data into the histogram memory, and creates a histogram representing the frequency of appearance for each code. 5. The distortion correction device according to claim 3, wherein
The normalizing unit divides all codes into a plurality of blocks by dividing a plurality of codes having quantization levels close to each other into one block from the occurrence frequencies of the codes obtained by the occurrence frequency detection unit, and divides each code into a plurality of blocks. And calculating the appearance frequency obtained by dividing each code in each block by the average appearance frequency as a normalized appearance frequency. 6. The distortion correction device according to claim 3, wherein
The error bit detecting means calculates a difference in quantization level for each code based on the normalized frequency of occurrence detected by the normalizing means, and generates an error having a distortion based on a linearity error of an analog / digital converter. A distortion correction device characterized by detecting as a bit. 7. The distortion correction device according to claim 2, wherein
The distortion correction device, wherein the digital signal processing means corrects the error bit of the input digital audio signal to a code estimated to be correct and outputs the corrected code. 8. The distortion correction device according to claim 7, wherein
The correct code estimation in a digital signal processing means, said average frequency of occurrence normalized for each code having an error bit to X, the digital audio signal a plurality of samples K _t the same value from time K ₁ to K _t of When the same value continues, take the first difference D1 between the value of the first sample followed by the same value and the value of the previous sample, and take the value of the last sample followed by the same value and the next sample Second with the value of
Taking the difference D2, when the first difference D1 and the second difference D2 is a positive value, the value of the K _{t / X} to K _t of the plurality of samples and +1, the first difference D1 And the second difference D2
K If is a negative value, the K ₁ among the plurality of samples
A distortion correction device, wherein a value up to _t (1-1 / X) is set to +1. 9. A storage means for storing a digital audio signal converted into a code by an analog / digital converter, and reading out the digital audio signal from the storage means, for each code representing a quantization level for a plurality of samples. An appearance frequency detection means for detecting the frequency as a level distribution, and applying a low-pass filtering process to the level distribution detected by the appearance frequency detection means to obtain an ideal level distribution when there is no error in the analog / digital converter. Normalizing means for detecting a level distribution obtained by dividing the level distribution before the low-pass filtering process is performed by the approximation to the ideal level distribution as a normalized level distribution, and A conversion table for the level of the digital audio signal from the level distribution A distortion correction device comprising: a conversion table calculating unit that converts the level of the digital audio signal into level data according to the conversion table; 10. The level distribution of a digital audio signal recorded using an analog / digital converter,
A conversion error of the used analog / digital converter is detected, and a conversion table is calculated from the detected conversion error.
A distortion correction method comprising: performing a table conversion process on the digital audio signal to correct a distortion generated by a conversion error of an analog / digital converter. 11. The distortion correction method according to claim 10, wherein said conversion table obtains a level distribution of a fixed number of samples for a digital audio signal recorded by using an analog / digital converter. Is subjected to a low-pass filtering process to approximate the ideal level distribution when there is no error in the analog / digital converter, and the level distribution before the low-pass filtering process is performed is approximated to the ideal level distribution. A distortion correction method, wherein the distortion correction method is created by integrating the divided values. 12. A linear count-up signal is input to an analog / digital converter, and an input signal and an output signal of the analog / digital converter at this time are compared to determine a conversion error of the analog / digital converter. A distortion correction method comprising calculating and correcting a distortion based on a conversion error included in an output signal from the analog / digital converter using the conversion error calculated as described above.