JPH10322206A - Device and method for reducing distortion - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce distortion based on a nonlinear error by detecting parts that have consecutive same values for pluralities of sample periods in error data detected at an error data detection means as an interpolation object range, using data at one preceding sample and one succeeding sample to the interpolation object range to calculate interpolation data and synthesizing the interpolation data with the digital audio signal. SOLUTION: A level distribution detection means 103 and an error distribution calculation means 104 calculate an error distribution of a received digital audio signal. Then a changeover means 102 switches an output destination of the digital audio signal in the device, and interpolation processing to reduce nonlinear error distortion by an error data detection means 105, a pattern detection means 106, a delay means 107, a data interpolation means 108 and an interpolation discrimination means is conducted. The data interpolation means 108 calculates interpolation data of an interpolation object range of the digital audio signal from the delay means 107 and synthesizes the data with the digital audio signal.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to reduction of distortion due to non-linear errors in digital audio signals.

[0002]

2. Description of the Related Art When digitally recording an audio signal, analog / digital (A / D: Analog / Digital) is used.
An analog audio signal is converted to a digital audio signal using a converter. When a digital audio signal is reproduced, digital / analog (D / A: Di)
The digital audio signal is converted to an analog audio signal using a gital / Analog) converter. Among digital recordings, PCM (Puls
e Code Modulation) In recording, the time axis direction of an analog audio signal is subdivided (sampling: sampling) at a certain frequency, and the signal level of each subdivided sample is represented by a certain bit length (quantization: (Quantization).

[0003] In the early stage of such PCM recording, 13 bits or 1 bit, which is lower than the present bit rate, is low.
A 4-bit successive approximation type A / D converter has been mainly used. Also, A / D converters in the early days of PCM recording had poor linearity in A / D conversion characteristics and contained many non-linear errors as compared with 16-bit A / D converters that are currently widely used. And the recorded digital audio signal also had many distortions.

[0004] The digital audio signals of the past having such distortions, some of which are excellent in performance and some of which are valuable in the history of music because of their recording sources, Compensate various distortions, PC
16 bits with a higher bit number than the A / D converter at the time of M recording
There is a demand to reproduce a digital audio signal such as a bit.

[0005]

As one of the methods for correcting the non-linear error distortion among the distortions in the past digital audio signal as described above, A
Although a method of measuring the conversion characteristics of the / D converter and performing the inverse correction is conceivable, at present, there is often no A / D converter used at the time of recording, and even if it is present, the A / D converter is present. Is changing over time, and it is impossible to correct the nonlinear error distortion during recording based on the current A / D conversion characteristics, and whether the A / D converter used during recording exists. Regardless of whether or not it is difficult to reduce the nonlinear error distortion of past PCM recording sources (digital audio signals).

Accordingly, an object of the present invention is to reduce distortion due to a nonlinear error of a digital audio signal.

[0007]

According to a first aspect of the present invention, there is provided a distortion reducing apparatus for reducing distortion due to a non-linear error of a digital audio signal, wherein a quantization level is set for each of a plurality of samples of the input digital audio signal. A level distribution detecting means for detecting a frequency of appearance for each code to be represented as a first level distribution, and applying a filtering process to the first level distribution detected by the level distribution detecting means to provide a non-linear error to the analog / digital converter. Error distribution calculating means for calculating a second level distribution approximating an ideal level distribution in the case where there is no error, and calculating an error distribution obtained by dividing the first level distribution by the second level distribution; Data corresponding to a code having an error frequency equal to or greater than a predetermined threshold from the error distribution calculated in And the error data detection means for detecting,
A pattern detecting means for detecting, as an interpolation target range, a portion in which the error data detected by the error data detecting means has the same value and a plurality of continuous samples; and data one sample before and one sample before the interpolation target range detected by the pattern detecting means. Data interpolation means for calculating interpolation data in an interpolation target range using the subsequent data and combining the data with a digital audio signal.

According to a second aspect of the present invention, there is provided a distortion reducing apparatus for reducing distortion based on a non-linear error of a digital audio signal. A level distribution detecting means for detecting the frequency of appearance as a first level distribution, and a filtering process performed on the first level distribution detected by the level distribution detecting means so that the analog / digital converter has no non-linear error Error distribution calculating means for calculating an error distribution obtained by calculating a second level distribution approximating the ideal level distribution of the first level distribution and dividing the first level distribution by the second level distribution; From the error distribution, an error in detecting data corresponding to a code having an error frequency equal to or greater than a predetermined threshold value as error data. A data detection unit, a pattern detection unit that detects a portion where the error data detected by the error data detection unit has the same value and a plurality of consecutive samples is an interpolation target range, and one sample before the interpolation target range detected by the pattern detection unit. Data interpolation means for calculating interpolation data of an interpolation target range using the data of one sample and the data after one sample, and synthesizing the digital audio signal, and all the interpolation data of the interpolation target range from the data interpolation means, If the error data is within a predetermined judgment value, a digital audio signal synthesized with the interpolation data is output, and if the error data is not within a predetermined judgment value, the input digital audio signal is left as it is. And output interpolation determining means.

The present invention described in claim 3 provides the present invention in claim 2
In the distortion reducing apparatus described above, the predetermined determination value is ± 0.5 LSB with respect to the error data.

The present invention described in claim 4 is based on claim 1.
And a distortion reduction device according to claim 2, wherein the data interpolation means calculates the interpolation data by linear interpolation using the data one sample before and one sample after the interpolation target range.

According to a fifth aspect of the present invention, there is provided a distortion reducing method for reducing a distortion based on a non-linear error of a digital audio signal, wherein a plurality of samples of an input digital audio signal are generated for each code representing a quantization level. The frequency is detected as a first level distribution, and a filtering process is performed on the first level distribution to obtain a second level distribution approximating an ideal level distribution when the analog / digital converter has no nonlinear error. Calculate,
An error distribution is calculated by dividing the first level distribution by the second level distribution, and data corresponding to a code having an error frequency equal to or greater than a predetermined threshold is detected from the error distribution as error data, and the error data has the same value. , A portion that is continuous for a plurality of samples is detected as an interpolation target range, and interpolation data of the interpolation target range is calculated using data one sample before and one sample after the interpolation target range, and synthesized into a digital audio signal. It is characterized by outputting.

According to a sixth aspect of the present invention, there is provided a distortion reducing method for reducing a distortion based on a non-linear error of a digital audio signal, wherein a plurality of samples of the input digital audio signal are provided for each code representing a quantization level. Detecting the frequency of appearance as a first level distribution,
A filtering process is performed on the first level distribution to calculate a second level distribution approximating an ideal level distribution in a case where there is no non-linear error in the analog / digital converter. An error distribution divided by the level distribution of 2 is calculated, and data corresponding to a code having an error frequency equal to or greater than a predetermined threshold is detected as error data from the error distribution. The data is detected as the interpolation target range, the interpolation data of the interpolation target range is calculated using the data one sample before and one sample after the interpolation target range, and the interpolation data is synthesized with the digital audio signal. When the interpolation data is within a predetermined judgment value with respect to the error data, a digital audio signal in which the interpolation data is synthesized is output. Is characterized in that directly outputs digital audio signals input if not within a predetermined determination value with respect.

The present invention according to claim 7 provides the present invention according to claim 5.
In the distortion reduction method described above, the predetermined determination value is ± 0.5 LSB with respect to the error data.

According to the present invention, for a plurality of samples of a digital audio signal having a distortion based on a nonlinear error, the frequency of appearance of each code representing a quantization level is detected as a first level distribution, and the first level distribution is detected. A filtering process (low-pass filtering process) is performed on the level distribution to calculate a second level distribution approximating an ideal level distribution when the analog / digital converter has no non-linear error. An error distribution divided by the second level distribution is calculated. Data corresponding to a code having an error frequency equal to or greater than a predetermined threshold is detected as error data from the error distribution, and how many samples of the error data continue with the same value is detected. Interpolation data is calculated by linear interpolation or the like using a portion where the error data is continuous for a plurality of samples as an interpolation target range, and is synthesized with a digital audio signal and output. This makes it possible to detect a non-linear error from the digital audio signal and reduce distortion based on the non-linear error, regardless of whether an A / D converter is present.

According to the present invention, for a plurality of samples of a digital audio signal having a distortion based on a non-linear error, the frequency of appearance of each code representing a quantization level is detected as a first level distribution, and the first level distribution is detected. A filtering process (low-pass filtering process) is performed on the level distribution to calculate a second level distribution approximating an ideal level distribution when the analog / digital converter has no non-linear error. An error distribution divided by the second level distribution is calculated. Data corresponding to a code having an error frequency equal to or greater than a predetermined threshold is detected as error data from the error distribution, and how many samples of the error data continue with the same value is detected. Interpolation data is calculated by linear interpolation or the like using a portion where the error data is continuous for a plurality of samples as an interpolation target range, and is synthesized with a digital audio signal and output. If all the interpolation data in the interpolation target range is within a predetermined determination value for the error data, a digital audio signal in which the interpolation data is synthesized is output. If not, the original digital audio signal is output as it is. This makes it possible to detect a non-linear error from the digital audio signal and reduce distortion based on the non-linear error, regardless of whether an A / D converter is present.

[0016]

FIG. 1 is a schematic diagram showing a schematic configuration of an embodiment of a distortion reduction apparatus according to the present invention. Here, a digital audio signal that has been A / D converted and recorded on a recording medium in the past, that is, a digital audio signal recorded by a low-bit A / D converter having a non-linear error is a 13-bit digital audio signal. It is an audio signal (13-bit digital audio signal), which is converted into a digital audio signal having a bit length of 16 bits while reducing nonlinear error distortion by a distortion reduction device.

In FIG. 1, a distortion reduction device 101 includes a switching unit 102 for switching the output of an input digital audio signal,
A level distribution detecting means 103 for detecting the frequency of appearance of each code representing each quantization level for a plurality of samples as a histogram, that is, a level distribution (first level distribution), and a first level detected by the level distribution detecting means 103. Is subjected to a low-pass filtering process to calculate a second level distribution approximating an ideal level distribution when the A / D converter has no error, and the first level distribution before the low-pass filtering process is performed. Distribution calculation means 104 for calculating an error distribution indicating a non-linear error distribution by dividing the level distribution of
And comparing the error frequency of each code in the error distribution calculated by the error distribution calculation means 104 with a predetermined threshold value,
Error data detection means 105 for detecting data corresponding to a code having an error frequency equal to or greater than a threshold value as error data
And a pattern detecting means 10 for detecting, as an interpolation target range, a portion where error data has the same value and a plurality of samples are continuous.
6, the delay means 107 for delaying the input digital audio signal, and the data of the sample before and after the interpolation target range, the interpolation data of the interpolation target range in the digital audio signal from the delay means 107 is calculated, And a data interpolation means 108 for synthesizing with the signal.

Each means will be described. The switching unit 102 switches the output destination of the digital audio signal input to the distortion reduction device 101. That is, after the error distribution of the input digital audio signal is calculated by the level distribution detecting means 103 and the error distribution calculating means 104 described later, the output destination of the digital audio signal is switched by the switching means 102, and the error data detecting means described later 105, pattern detection means 106, delay means 107, data interpolation means 108, interpolation determination means 1
Interpolation processing for reducing the non-linear error distortion due to X.09 is performed.

The level distribution detecting means 103 detects the frequency of occurrence of each code representing a quantization level for a plurality of samples of the input digital audio signal. That is, a first level distribution (histogram) representing the frequency of appearance for each code is detected for a digital audio signal for a plurality of samples. The level distribution detecting means 103 may include a storage unit such as a hard disk for storing digital audio signals.

Here, a code representing the quantization level of the digital audio signal will be described. The quantization level of the digital audio signal is represented by a plurality of bits using a binary method of 1 and 0. By increasing the number of bits, a finer amplitude level can be expressed.
For example, when 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, and the -4096 to 4095 are represented as codes. A code of 0 indicates that the amplitude level of the analog signal is 0, and indicates that the larger the absolute value of the code, the larger the amplitude level of the analog signal.

The level distribution (histogram) of the digital audio signal will be described. FIG. 2 is a schematic diagram showing a level distribution of a digital audio signal.
(A) shows a level distribution of a digital audio signal that has been A / D converted using an A / D converter having good conversion accuracy,
(B) shows A / D conversion using an A / D converter having a nonlinear error.
4 shows the level distribution of a digital audio signal that has been D-converted. As shown in FIG. 2A, the level distribution of a digital audio signal that has been A / D-converted using an A / D converter with good conversion accuracy has a positive It is almost symmetric in the negative direction, and the appearance frequency is high when the code is 0, and the appearance frequency gradually decreases as the code increases in the positive direction or decreases as the code decreases in the negative direction. Such a characteristic of the level distribution is unique to each input digital audio signal.

However, the level distribution of a digital audio signal A / D converted by an A / D converter having a 13-bit or 14-bit nonlinear error is, as shown in FIG. Therefore, there is a code whose appearance frequency rapidly increases or a code whose appearance frequency rapidly decreases. These codes are codes in which distortion based on non-linear errors has occurred.

A method of detecting a level distribution in the level distribution detecting means 103 will be described. The level distribution detecting means 103 includes a histogram memory which stores a value corresponding to the code using the code itself representing the quantization level as an address. The histogram memory is accessed based on the frequency of appearance, and a value corresponding to the address stored in the histogram memory is read. 1 is added to this value, and the value obtained by adding 1 is written to the same address. By repeating this process for a plurality of samples, 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 for performing the above-described processing is not all of the digital audio signals for reducing the non-linear error distortion, but if the digital audio signals are converted by the same A / D converter into digital audio signals, It may be a part of the signal, for example a sample for the first few seconds. However, as the number of samples increases, the accuracy of the level distribution improves. For example, when 2,500,000 samples in which a relatively high level portion is continuous are used, a good level distribution can be obtained. This level distribution can be detected by a digital counter circuit or software.

Next, the error distribution calculating means 104 will be described. FIG. 3 is a schematic diagram for explaining the processing of the error distribution calculation means in the distortion reduction device of the present embodiment.
(A) shows a level distribution of a digital audio signal having a non-linear error distortion, (b) shows a level distribution of a filtered digital audio signal,
(C) shows an error distribution. The error distribution calculation means 104
It has a low-pass filtering function, and has the first level detection means 103 detected by the level distribution detecting means 103 as shown in FIG.
Is subjected to low-pass filtering processing to calculate a second level distribution approximating an ideal level distribution when the A / D converter has no nonlinear error as shown in FIG. . Then, by dividing the first level distribution before the low-pass filtering processing is performed by the second level distribution after the low-pass filtering processing is performed, a 13-bit A / A signal as shown in FIG. The nonlinear error due to the D converter is calculated as an error distribution in which the appearance frequency in the case where there is no error is normalized to 1.0.

More specifically, a digital audio signal
The level distribution of the level i of the (music signal) is g (i), the number of quantization bits is n, the number of samples for averaging the low-pass filter (moving average filter) is m (m: odd number), and the level distribution is smoothed. h (i) is expressed by the following equation (1).

[0027]

(Equation 1)

The error distribution calculating means 104 calculates the normalized value
N (i) is obtained from the following equation (2).

[0029]

(Equation 2)

Here, c (j) is a window function satisfying the following equation (3).

[0031]

(Equation 3)

In other words, g (i) is the window function c
The moving average filter of (j) is shown. In the moving average method, it is impossible to calculate both ends of the smoothed waveform. In addition, since both ends of the smoothed waveform are portions where the music level is large and such a large level does not actually exist, it can be ignored. (M-1) / 2 corresponding to samples at both ends are excluded from the range of the code i.

As described above, the error distribution calculating means 104 calculates the appearance frequency g of the digital code i of the digital audio signal.
For (i), by using a moving average filter and calculating an average value with neighboring values, the appearance frequency g (i) of the digital code i of the digital audio signal is approximated to the ideal appearance frequency. 2 are calculated. Thus, even when the level distribution of the digital audio signal after the A / D conversion has an inherent bias depending on the type of musical sound, the second level distribution approximate to the ideal level distribution can be calculated.

Next, the error data detecting means 105 calculates, from the error distribution shown in FIG. 3C calculated by the error distribution calculating means 104, data corresponding to a code whose error frequency is equal to or greater than a predetermined threshold value. (Code to be subjected to distortion reduction). That is, the error distribution shown in FIG. 3C represents the error distribution for each code, and the error frequency when there is no error is 1.0. In this error distribution, for example, an error frequency of 1.2 is set as a threshold, and data corresponding to a code having an error frequency of 1.2 or more is detected as error data to be subjected to distortion reduction. .

Here, a method of detecting error data by the error data detecting means 105 will be described. The level distribution detecting means 103 and the error distribution calculating means 104 calculate a code having a nonlinear error from the frequency of appearance of each code of the digital audio signal. The error data detecting means 105 first detects a code having an error frequency equal to or larger than a threshold value from the error distribution calculated by the error distribution calculating means 104, and stores the code in a memory provided. Then, for the digital audio signal switched and input by the switching unit 102, data corresponding to a code having an error frequency equal to or greater than a threshold stored in the memory is detected as error data.

Next, the pattern detecting means 106 detects how many samples of the error data detected by the error data detecting means 105 are continuous with the same value. A portion where the error data has the same value and is continuous for a plurality of samples (two or more samples) is detected as a range (interpolation target range) for which the interpolation data is calculated.

Next, the delay means 107 is connected to the switching means 102.
For delaying a digital audio signal to be input to a data interpolation unit 108, which will be described later, for a time corresponding to a processing time in which the error data detection unit 105 and the pattern detection unit 106 perform respective processes on the digital audio signal from Things. That is, the digital audio signal input to the data interpolation means 108 is delayed for a time corresponding to a process of detecting error data from the digital audio signal and detecting whether or not the error data corresponds to the interpolation target range.

Next, the data interpolation means 108 will be described. The data interpolation means 108 is used for the pattern detection means 10.
6. Data of samples before and after the interpolation target range detected in 6,
That is, interpolation data is calculated using data before and after the error data which is different from the value of the detected error data, and the interpolation data and the digital audio signal input through the delay means 107 are combined and output. Is what you do.

The data interpolation process will be specifically described. In the data interpolation processing in this embodiment, data interpolation by linear interpolation is performed in consideration of speeding up of arithmetic processing and the like, but spline interpolation or the like may be used. FIG. 4 is a schematic diagram for explaining linear interpolation in the distortion reduction device of the present embodiment. As shown in FIG. 4, when the error data detected by the error data detecting means 105 has the same value and continues for five samples, the error data and the data of two samples before and after the succession are connected by a straight line, and The values are combined with the input digital audio signal as the respective interpolation data and output.

Since the calculated interpolation data has a value smaller than 1 LSB in the number of bits of the input digital audio signal, the bit length of the digital audio signal output from the distortion reduction device 101 is extended. It will be. That is, in FIG.
Assuming that the level (LSB) at the time of x is y (tx) and each level after the linear interpolation is z (tx), the linear interpolation is a temporary function of the slope d as shown in the following equation (4).

[0041]

(Equation 4)

Of the interpolated data of a plurality of samples (data from time t1 to time t5), the value K (n) of the m-th sampled interpolated data is set such that the starting point of the interpolated data (time t0) is a. , And the end point (time t5) is b, the following equation (5) is obtained.

[0043]

(Equation 5)

For example, in the case shown in FIG. 4, each level of the interpolation data is given by the following equation (6).

[0045]

(Equation 6)

As described above, the interpolation data of a plurality of samples corresponds to 1 LSB in the input digital audio signal.
This means that it takes a smaller value, which means that the bit length of the digital audio signal has been extended.

With the above arrangement, the frequency of occurrence of each code representing the quantization level is detected as a first level distribution for a plurality of samples of the input digital audio signal. A filtering process is performed on the first level distribution to calculate a second level distribution approximating an ideal level distribution when there is no non-linear error in the analog / digital converter. Is calculated by dividing the error distribution by the level distribution. Data corresponding to a code having an error frequency equal to or greater than a predetermined threshold value is detected as error data from the error distribution, and is detected as an interpolation target range in which the error data has the same value and a plurality of samples continue. Then, by using the data one sample before and one sample after the interpolation target range to calculate the interpolation data of the interpolation target range, synthesize it with a digital audio signal, and output the digital audio signal, the data is used in the past recording. Irrespective of whether an A / D converter is present or not, a non-linear error can be detected from a digital audio signal, and distortion based on the non-linear error can be reduced.

Next, another embodiment of the distortion reducing apparatus of the present invention will be described. FIG. 5 is a schematic diagram showing a schematic configuration of another embodiment of the distortion reduction device of the present invention. 5, the same components as those in FIG. 1 are denoted by the same reference numerals, and detailed description of each component will be omitted. Further, a digital audio signal having a bit length of 13 bits (13-bit digital audio signal) recorded by a low-bit A / D converter having a non-linear error is input to the distortion reduction device 101, and the non-linear error distortion is reduced. And a digital audio signal having a bit length of 16 bits.

In FIG. 5, a distortion reducing device 101 includes a switching means 102 for switching the output of an input digital audio signal,
A level distribution detecting means 103 for detecting the frequency of appearance of each code representing each quantization level for a plurality of samples as a histogram, that is, a level distribution (first level distribution), and a first level detected by the level distribution detecting means 103. Is subjected to a low-pass filtering process to calculate a second level distribution approximating an ideal level distribution when the A / D converter has no error, and the first level distribution before the low-pass filtering process is performed. Distribution calculation means 104 for calculating an error distribution indicating a non-linear error distribution by dividing the level distribution of
And comparing the error frequency of each code in the error distribution calculated by the error distribution calculation means 104 with a predetermined threshold value,
Error data detection means 105 for detecting data corresponding to a code having an error frequency equal to or greater than a threshold value as error data
And a pattern detecting means 10 for detecting, as an interpolation target range, a portion where error data has the same value and a plurality of samples are continuous.
6, the delay means 107 for delaying the input digital audio signal, and the data of the sample before and after the interpolation target range, the interpolation data of the interpolation target range in the digital audio signal from the delay means 107 is calculated, Data interpolation means 108 for synthesizing with signal
When all the interpolation data in the interpolation target range calculated by the data interpolation means 108 are within the predetermined range, a digital audio signal obtained by synthesizing the interpolation data is output, and the digital audio signal exists within the predetermined range. If not, an interpolation determining means 109 is provided for outputting the input original digital audio signal as it is.

The switching means 102 includes the level distribution detecting means 1
03. After calculating the error distribution of the input digital audio signal by the error distribution calculation means 104, the output destination of the digital audio signal is switched, and the error data detection means 105, the pattern detection means 106, the delay means 107, and the data interpolation means 108 , And outputs a digital audio signal to the interpolation determination means 109.

The level distribution detecting means 103 detects the frequency of occurrence of each code representing the quantization level of a plurality of samples of the input digital audio signal, and detects the first level distribution (histogram). .

The error distribution calculating means 104 has a low-pass filtering function, and the level distribution detecting means 103
A low-pass filtering process is performed on the first level distribution detected in step (1) to calculate a second level distribution that approximates an ideal level distribution when the A / D converter has no nonlinear error. Then, the first level distribution is divided by the second level distribution to calculate a non-linear error due to the 13-bit A / D converter as an error distribution in which the appearance frequency in the case where there is no error is normalized to 1.0. .

The error data detecting means 105 detects, from the error distribution calculated by the error distribution calculating means 104, data corresponding to a code whose error frequency is equal to or greater than a predetermined threshold value as error data. As the predetermined threshold, for example, an error frequency of 1.2 is set as the threshold, and data corresponding to a code having an error frequency of 1.2 or more is detected as error data to be subjected to distortion reduction.

The pattern detecting means 106 detects how many consecutive samples of the error data detected by the error data detecting means 105 have the same value. A portion where the error data has the same value and is continuous for a plurality of samples (two or more samples) is detected as a range (interpolation target range) for which the interpolation data is calculated.

The delay means 107 has a time corresponding to the processing time in which each of the error data detection means 105 and the pattern detection means 106 performs the processing on the digital audio signal from the switching means 102, and a data interpolation means 108 to be described later. Delays the digital audio signal input to.

The data interpolation means 108 uses the data of the sample before and after the interpolation target range detected by the pattern detection means 106, that is, the data before and after the error data which is different from the value of the detected error data. The interpolation data is calculated, and the interpolation data and the digital audio signal input via the delay unit 107 are synthesized and output. In the data interpolation processing, data interpolation by linear interpolation is performed in consideration of speeding up of arithmetic processing and the like, but spline interpolation or the like may be used.

Next, the interpolation determining means 109 will be described. The configuration of the distortion reducing apparatus of the present embodiment is different from that of the above-described embodiment in that an interpolation determining means 109 is provided at a stage subsequent to the data interpolating means 108. Interpolation determination means 10
Reference numeral 9 indicates that all the interpolation data in the interpolation target range calculated by the data interpolation unit 108 are within a predetermined determination value with respect to the data before interpolation (error data detected by the error data detection unit 105). It is determined whether or not to output a digital audio signal obtained by synthesizing the interpolation data or to output the original digital audio signal. By determining whether or not to perform the data interpolation, the non-linear error of the digital audio signal can be reduced more accurately.

That is, when all the interpolation data in the range to be interpolated with respect to the error data are within a predetermined judgment value, all the interpolation data are set as the error range of the original value of the error data. That is, the interpolation data and the input digital audio signal are synthesized and output. However, if all the interpolation data in the interpolation target range for the error data is not within the predetermined determination value,
Data interpolation of error data is not performed because interpolation data other than the determination value is not a correct value. The judgment value is set to, for example, an error range of ± 0.5 LSB which is also possessed by an ideal A / D converter, and by setting the judgment value to ± 0.5 LSB, more accurate data interpolation processing can be performed.

FIG. 6 is a schematic diagram for explaining the interpolation judgment in the distortion reduction apparatus of the present embodiment. (A) shows a case where all the interpolated data is within ± 0.5 LSB with respect to the error data, and (b) shows a case where all the interpolated data is not within ± 0.5 LSB with respect to the error data. For example, as shown in FIG. 6A, when all the interpolation data in the interpolation target range is within ± 0.5 LSB of the error data, a digital audio signal obtained by synthesizing the interpolation data is output, and FIG. As shown in (1), if even one piece of interpolation data within the interpolation target range is not within ± 0.5 LSB of the error data, the original digital audio signal is output as it is.

That is, as a result of performing linear interpolation on the data of level a, the error frequency becomes N (a), 0.5 LSB
If (1 / 2LSB) is 4, all the interpolation data is
It is detected whether the data exists within the range shown by the following equation (7). If the data exists, a digital audio signal obtained by synthesizing the interpolation data is output. If at least one interpolation data out of the range exists, the data is output. Output the original digital audio signal without interpolation.

[0061]

(Equation 7)

In equation (7), if the bit length of the input digital audio signal is Ni and the bit length of the output digital audio signal is No, the equation (8) may be used.

[0063]

(Equation 8)

The processing operation of the distortion reduction device described above will be described. FIG. 7 is a flowchart illustrating a processing operation of the distortion reduction device according to the present embodiment. First, an error distribution of a digital audio signal is calculated by the level distribution detecting means 103 and the error distribution calculating means 104. The digital audio signal sample number (measurement sample number) a for generating the first level distribution is set to a predetermined number (S11).

Next, the digital audio signal (13 bits) is sampled and read from the recording medium with the number of samples a and stored in the storage unit (S12). The 13-bit digital audio signal stored in the storage unit is
The level distribution detecting means 103 detects the level distribution. In this case, the level distribution detecting means 103 has a histogram memory as described above, increments the value of the histogram memory at the address (address) corresponding to the input data by +1 (S13), and repeats this by a predetermined number (the number of samples). a) It repeatedly creates a histogram (first level distribution) (S14, S15).

Next, the error distribution calculating means 104 performs a low-pass filtering process on the first level distribution detected by the level distribution detecting means 103, so that the ideal level distribution is obtained when the A / D converter has no nonlinear error. A second level distribution approximating a typical level distribution is calculated. Then, the first level distribution before the low-pass filtering processing is performed,
By dividing by the second level distribution after the low-pass filtering processing is performed, the non-linear error due to the 13-bit A / D converter is calculated as an error distribution with an error frequency of 1.0 when there is no error ( S16).

Next, the error data detecting means 104 detects a code whose error frequency is equal to or larger than the threshold value from the predetermined threshold value and the error distribution calculated by the error distribution calculating means 104, and stores the code in the memory (S17). Then, when the digital audio signal is input to the error data detecting means 104 by the switching means 102, the error data detecting means 104 detects data corresponding to the code stored in the memory as error data (a code to be subjected to distortion reduction). (S18).

Next, the pattern detecting means 106 detects how many samples of the error data detected by the error data detecting means 105 are continuous with the same value. It is detected as an interpolation target range for calculating interpolation data (S19).

Next, the data interpolation means 108 calculates interpolation data using the data before and after the interpolation target range detected by the pattern detection means 106, and calculates the digital audio signal input via the delay means 107 and the interpolation data. Are combined and output (S20).

The interpolation determining means 109 determines that all the interpolation data in the interpolation target range calculated by the data interpolation means 108 is ± 0.5% of the data (error data) before interpolation.
It is determined whether it is within LSB (S21).

If all the interpolation data in the interpolation target range is within ± 0.5 LSB with respect to the error data, a digital audio signal obtained by synthesizing the interpolation data is output (S2).
2) If any one of the interpolation data in the interpolation target range is not within ± 0.5 LSB with respect to the error data, the original digital audio signal is output as it is (S23).

As described above, a level distribution in a plurality of samples is detected from a digital audio signal, and an error distribution having a non-linear error is calculated from the level distribution. Error data corresponding to a code having an error frequency equal to or greater than a predetermined threshold is detected from the error distribution, and how many samples of the error data continue with the same value is detected. A portion where the error data has the same value and a plurality of samples are continuous is set as an interpolation target range, and interpolation data is calculated by linear interpolation or the like, and is combined with a digital audio signal. Then, all the interpolation data in the interpolation target range is ± 0.5 L with respect to the error data.
If it is within the range of SB, a digital audio signal in which the interpolation data is synthesized is output, and the error data is ± 0.5%.
If it is not within the LSB range, the original digital audio signal is output as it is.

Thus, regardless of whether or not an A / D converter used during recording in the past exists, a non-linear error is detected from the digital audio signal and distortion based on the non-linear error is reduced. Can be.

[0074]

According to the present invention, it is possible to reduce distortion due to a nonlinear error of a digital audio signal.

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing a schematic configuration of one embodiment of a distortion reduction device of the present invention.

FIG. 2 is a schematic diagram showing a level distribution of a digital audio signal. (A) shows the level distribution of a digital audio signal that has been A / D converted using an A / D converter with good conversion accuracy, and (b) shows the A / D converter using an A / D converter having a non-linear error. 5 shows a level distribution of a converted digital audio signal.

FIG. 3 is a schematic diagram for explaining processing of an error distribution calculation unit in the distortion reduction device of the present embodiment. (A) shows the level distribution of a digital audio signal having a non-linear error distortion, (b) shows the level distribution of a filtered digital audio signal, and (c) shows the error distribution.

FIG. 4 is a schematic diagram for explaining linear interpolation in the distortion reduction device of the present embodiment.

FIG. 5 is a schematic diagram showing a schematic configuration of another embodiment of the distortion reduction device of the present invention.

FIG. 6 is a schematic diagram for explaining interpolation determination in the distortion reduction device of the present embodiment. (A) shows a case where all the interpolated data is within ± 0.5 LSB with respect to the error data, and (b) shows a case where all the interpolated data is within ± 0.5 LSB of the error data.
The case where the difference is not within 0.5 LSB is shown.

FIG. 7 is a flowchart illustrating a processing operation of the distortion reduction device in the present embodiment.

[Explanation of symbols]

101: a distortion reduction device, 102: switching means, 103
.. Level distribution detection means, 104 error distribution calculation means, 105 error data detection means, 106 pattern detection means, 107 delay means, 108 data interpolation means, 109 interpolation determination means

Claims (7)

[Claims] 1. A distortion reduction apparatus for reducing distortion based on a non-linear error of a digital audio signal, wherein a frequency of appearance of each code representing a quantization level for a plurality of samples of the input digital audio signal is defined as a first level distribution. The level distribution detecting means to be detected and the first level distribution detected by the level distribution detecting means are subjected to a filtering process to approximate an ideal level distribution when the analog / digital converter has no nonlinear error. Error distribution calculation means for calculating a second level distribution and calculating an error distribution obtained by dividing the first level distribution by the second level distribution; and a predetermined error distribution calculated from the error distribution calculated by the error distribution calculation means. Error data detection means for detecting data corresponding to a code having an error frequency equal to or greater than a threshold as error data Pattern detection means for detecting, as an interpolation target range, a portion in which the error data detected by the error data detection means has the same value and continuous for a plurality of samples; and one pattern before the interpolation target range detected by the pattern detection means. A data interpolation unit that calculates interpolation data in the interpolation target range using the data and the data after one sample, and combines the data with a digital audio signal. 2. A distortion reduction apparatus for reducing distortion based on a non-linear error in a digital audio signal, wherein a frequency of appearance of each code representing a quantization level is provided for a plurality of samples of the input digital audio signal as a first level distribution. The level distribution detecting means to be detected and the first level distribution detected by the level distribution detecting means are subjected to a filtering process to approximate an ideal level distribution when the analog / digital converter has no nonlinear error. Error distribution calculation means for calculating a second level distribution and calculating an error distribution obtained by dividing the first level distribution by the second level distribution; and a predetermined error distribution calculated from the error distribution calculated by the error distribution calculation means. Error data detection means for detecting data corresponding to a code having an error frequency equal to or greater than a threshold as error data Pattern detection means for detecting, as an interpolation target range, a portion in which the error data detected by the error data detection means has the same value and continuous for a plurality of samples; and one pattern before the interpolation target range detected by the pattern detection means. Data interpolation means for calculating interpolation data of the interpolation target range using the data and the data after one sample and synthesizing the data with a digital audio signal; and all the interpolation data of the interpolation target range of the data interpolation means are Outputting a digital audio signal in which the interpolation data is synthesized when the error data is within a predetermined determination value,
An interpolation determining unit that outputs an input digital audio signal as it is when the error data is not within the predetermined determination value. 3. The distortion reducing apparatus according to claim 2, wherein the predetermined judgment value is ± 0.
A distortion reduction device characterized by being 5LSB. 4. The distortion reducing apparatus according to claim 1, wherein said data interpolation means performs linear interpolation using data one sample before and one sample after the interpolation target range. Is calculated. 5. A distortion reduction method for reducing distortion based on a non-linear error of a digital audio signal, wherein a frequency of appearance of each code representing a quantization level for a plurality of samples of the input digital audio signal is defined as a first level distribution. Detecting the first level distribution, performing a filtering process on the first level distribution, and calculating a second level distribution approximating an ideal level distribution when the analog / digital converter has no nonlinear error; An error distribution obtained by dividing the level distribution by the second level distribution is calculated, and data corresponding to a code having an error frequency equal to or greater than a predetermined threshold is detected as error data from the error distribution, and the error data has the same value. , A portion that is continuous for a plurality of samples is detected as an interpolation target range. Distortion reduction method and outputting synthesized digital audio signal to calculate the interpolation data of the interpolation target range by using the data after the sample. 6. A distortion reduction method for reducing distortion based on a non-linear error in a digital audio signal, wherein a frequency of appearance of each code representing a quantization level for a plurality of samples of the input digital audio signal is defined as a first level distribution. Detecting and filtering the first level distribution to calculate a second level distribution approximating an ideal level distribution when there is no non-linear error in the analog / digital converter; An error distribution obtained by dividing the level distribution by the second level distribution is calculated, and data corresponding to a code having an error frequency equal to or greater than a predetermined threshold is detected as error data from the error distribution, and the error data has the same value. , A portion that is continuous for a plurality of samples is detected as an interpolation target range. Using the sampled data, the interpolation data of the interpolation target range is calculated and synthesized into a digital audio signal, and all the interpolation data of the interpolation target range are within a predetermined determination value for the error data. Outputting a digital audio signal synthesized with the interpolation data in the case, and outputting the input digital audio signal as it is when the error data is not within the predetermined judgment value. . 7. The distortion reducing method according to claim 6, wherein the predetermined judgment value is ± 0.
A distortion reduction method characterized by being 5LSB.