JPH1083198A - Digital signal processing method and device therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To record and reproduce an important subband signal while preventing the degradation of a main band signal. SOLUTION: This device is provided with a first-order recorder 7 in which the main band signal and the subband signal obtained by dividing entire band signals of a digital signal are divided and information on a required code length are preliminarily recorded by a 1st pass constitution, an embedding position retrieving circuit 11 retrieving the main band signal whose required code length is less than 16 bits from the signals read out from the first-order recorder 7 and retrieving position where the subband signal is embedded, an LPF(low-pass filter) 15 filtering the subband signal and a data packing circuit 17 adding the subband signal filtered in the LPF 15 to the main band signal retrieved in the embedding position retrieving circuit 11.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a wide band signal such as a digital audio signal (the band here includes not only a frequency band but also a quantization noise band in S / N).
For example, the present invention relates to a digital signal processing method and apparatus suitable for, for example, recording on a recording medium or transmitting to a transmission medium for reproduction.

[0002]

2. Description of the Related Art A conventional so-called CD (Compact Disc: trademark) has a sampling frequency of 44.1 kHz, so that the maximum reproduction frequency is 22.05 kHz.
z. Each sample is represented by straight PCM having 16 quantization bits, and the S / N ratio is about 98d.
B.

[0003]

By the way, it has been found that a sound source such as gamelan music or yodel voice contains a lot of frequency components exceeding the maximum reproduction frequency of a CD. In addition, due to the improvement in the capacity of the recorder, recording with a quantization bit number of 16 bits or more, which is called high bit recording, is performed even in a normal recording studio. From the above, it is considered that sound quality beyond the conventional CD standard has been desired.

[0004] As a method of achieving sound quality exceeding the conventional CD standard, it is conceivable to increase the sampling frequency or increase the number of quantization bits. In order to fit the data obtained by increasing the sampling frequency or the number of quantization bits into a conventional CD-sized disk, the track pitch must be narrower than that of the CD standard. It is technically possible to increase the recording capacity of the disk by shortening the wavelength of the laser light generated by the pickup device, for example, to enable smaller pits to be formed on the disk, or the like.

However, if the recording capacity of the disk is increased as described above, compatibility with the conventional CD standard cannot be maintained, and the software market will be confused.

Therefore, without changing the conventional pickup device or disc format, etc., while maintaining the compatibility with the conventional CD standard, the band conventionally used (hereinafter referred to as the band is referred to as the frequency band). Not only S
/ N), the signal exceeding the conventional band is divided and divided to obtain a sound signal exceeding the conventional noise band (including the quantization noise band in / N). The band corresponding to the conventional CD standard is recorded as a main band in the form of straight PCM having 16 quantization bits of the CD standard, while the remaining band obtained by the above division (the band of the CD standard) For example, a signal processing method, a recording / reproducing apparatus, and a medium that compress and record using various coding methods such as high-efficiency coding as sub bands can be considered.

However, the problem with this method is that the straight band P of the main band depends on the state of the signal.
That is, the CM signal cannot keep the value of 16 bits which is the quantization bit number in the conventional CD standard. In other words, depending on the content of the signal, it may not be possible to obtain a signal of sufficient quality with the above 16 bits. For this reason, in practice, the above-described division is performed by suppressing the quantization noise to a level at which it is determined that the sound cannot be heard aurally based on a so-called psychoacoustic model. It cannot be denied that the psychoacoustic model is not absolute, and there is no guarantee that the sound quality of the signal in the main band is necessarily sufficient.

Accordingly, the present invention has been made in view of such circumstances, and enables recording and reproduction of a wideband signal using a conventional medium (disc). It is an object of the present invention to provide a digital signal processing method capable of recording and reproducing signals in important sub-bands while preventing signal deterioration in the sub-bands.

[0009]

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and a digital signal processing method and apparatus according to the present invention is capable of converting a full-band digital signal into a sub-band signal and a sub-band signal. And a signal of a main band having a required word length less than a predetermined value, filtering a signal of a sub-band, and filtering the signal of the sub-band having a required word length of a predetermined value. The above-described problem is solved by adding to a signal of a main band that is less than the value.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below with reference to the drawings.

According to the embodiment of the present invention, when a wideband signal such as a digital audio signal is recorded on a package medium such as a CD and reproduced, a signal in the entire band of the input digital audio signal is used as a signal in a main band, for example. The signal is divided into sub-band signals, straight PCM is used for the main band signal, and high-efficiency coding such as entropy coding and nonlinear quantization is used for the sub-band signal and filtering is performed. By determining the number of bits that are auditory redundant from the straight PCM word of the band signal and assigning the filtered sub-band signal to the redundant bits, signal deterioration in the main band is reduced, and Efficiently record or transmit sub-band signals and reproduce them It is intended to function. The input digital audio signal at this time is, for example, a signal reproduced from a recording medium on which recording is performed with a quantization bit number of 16 bits or more, which is referred to as the high bit recording, that is, a quantization bit number is, for example. A signal exceeding 16 bits of the CD standard can be given as an example. In this embodiment, the input digital audio signal has a sampling frequency of 44.1 kHz and a quantization bit number of 20 bits.

The above can be realized for the following reasons.

First, in the manufacturing process of a package medium (especially a CD), it is not necessary to give priority to real-time performance, so it is necessary to check the state of a signal to be recorded or transmitted as preprocessing. It is possible.

Here, in order to check the state of the signal, first, in a certain short section (a unit to be described later), a word length actually required by the input digital audio signal is determined. The word length actually required by the input digital audio signal is the word length that is required aurally, and if the word length has accuracy, when the signal is reproduced later,
The word length is such that the quantization noise included in this signal is not perceptually heard as deterioration.

Therefore, if the word length actually required by the 20-bit input digital audio signal is shorter than 16 bits, which is the CD standard, for example, if the word length is sufficiently high in auditory sense, then the input digital audio signal is required. When a digital audio signal is recorded, for example, in accordance with the CD standard, the above-mentioned C is converted from the 20-bit input digital audio signal.
Even if only the 16 bits corresponding to the D standard are taken out and the remaining bits are simply discarded, a good audio signal can be reproduced without any auditory problem during the subsequent reproduction. If the word length actually required for the 20-bit input digital audio signal is shorter than the 16 bits, only 16 bits corresponding to the CD standard are extracted from the input digital audio signal as described above. Then, the remaining bits are discarded, and some information is embedded in redundant bits other than the actually necessary bits in the remaining 16-bit word.
Even if a D-standard 16-bit word is formed, it will not appear as a problematic degradation if the embedded information is audibly extracted during later reproduction. When the information is embedded, the information is embedded from the LSB side (least significant bit) side where there is no problem in hearing.

On the other hand, when the word length required by the 20-bit input digital audio signal actually exceeds, for example, 16 bits, which is the CD standard, the input digital audio signal is recorded according to the CD standard, for example. For this reason, if only 16 bits corresponding to the CD standard are extracted from the input digital audio signal and the remaining bits are simply discarded, only the audio signal that has deteriorated during later reproduction can be reproduced. For this reason, the remaining bits must be included in the CD standard by some method.

Therefore, in this embodiment, when the word length required by the 20-bit input digital audio signal actually exceeds the 16-bit value, the input digital audio signal is converted to, for example, a CD standard. When recording by the, the signals of the entire band of the input digital audio signal are converted into ones corresponding to the CD standard.
The signal is divided into a signal of 6 bits and a signal of more than 16 bits, and the signal of 16 bits is recorded according to the CD standard. The redundant bit portion of another audio signal whose actual required word length is less than 16 bits (a redundant bit portion when another 20-bit input digital audio signal is converted to 16 bits)
Is embedded as described above. Therefore, at the time of subsequent reproduction, the 16-bit signal is reproduced, a signal exceeding the 16 bits embedded in the redundant bit portion is extracted from the another audio signal, and the extracted signal is By combining with a signal of 16 bits, an audio signal of good sound quality is reproduced.

In this embodiment, as described above, the 16-bit audio signal of the CD standard separated from the 20-bit input digital audio signal whose actual required word length exceeds the 16-bit. When,
The 16-bit audio signal of the CD standard extracted from the 20-bit input digital audio signal whose actual required word length is shorter than 16 bits, and the actual required word length is shorter than 16 bits Also, the 16-bit audio signal of the CD standard formed by embedding some information in the 16-bit redundant bits extracted from the short input digital audio signal is called the main band signal. Further, in this embodiment, the remaining signal (exceeding 16 bits) obtained by separating 16 bits of the CD standard from the 20-bit input digital audio signal whose actual required word length exceeds 16 bits. Minute bits)
Is called the signal of the sub-band. More specifically, the 20-bit input digital audio signal whose actual required word length exceeds 16 bits and the 16-bit CD standard obtained by dividing the 20-bit input disk audio. The difference signal between the bit and the signal in the main band is the signal in the sub-band. The sub-band signal is embedded in the redundant bits of the main-band signal extracted from another input digital audio signal whose actual required word length is shorter than 16 bits. In the case of a 20-bit input digital audio signal in which the actually required word length is shorter than 16 bits, it is not necessary to improve the sound quality by the sub-band. The remaining signal (sub-band signal) from which the 16-bit main band signal is extracted is discarded so as not to be recorded or transmitted. Further, a signal at a position different in time from the another audio signal in which the signal of the sub-band is embedded,
These are signals of different channels.

Here, as described above, the signal of the sub-band is extracted from the signal of the main band in which the signal of the sub-band is embedded, and the signal of the main band corresponding to the signal of the sub-band (the sub-band signal). In order to realize the reproduction of the signal by combining the main band signal and the divided main band signal), the information of the actually required word length of the main band and the signal of the sub band are embedded. At least information indicating in which position of the main band signal the sub-band signal is embedded is required. Therefore, when recording or transmitting the signal of the main band and the signal of the sub-band, the information other than the area in which the signal of the main band and the signal of the sub-band are arranged (recorded or transmitted area). An arrangement area (area for recording or transmission) for arranging is required. For example, in the case of recording on a CD standard disk, the above-mentioned arrangement area may be a so-called subcode information area. When reproducing a CD standard disc on which the information of the main band and the sub-band and the information of the embedding position are recorded, the embedding information and the like are read from the subcode information in the signal reproduced from the disc, and based on the information. Then, the signal in the sub-band is extracted and appropriately combined with the signal in the main band. A detailed description of the subcode information will be described later.

Regardless of whether the actually required word length exceeds the above 16 bits or is shorter than the above 16 bits, the 16 bit of the CD standard is converted from the above 20 bit input digital audio signal. When extracting the extracted signal of the main band, it is desirable to control the quantization noise such that sufficient sound quality can be secured when the extracted 16-bit main band signal is reproduced later. This means that a disc conforming to the CD standard, in which signals processed by the digital signal processing method and apparatus of the present invention are recorded, is reproduced using a conventional reproducing apparatus which does not support the method and apparatus of the present invention. This is to ensure that the sound quality can be sufficiently ensured even in the case where the sound is reproduced.

In this embodiment, the input digital audio signal of 20 bits is converted into 16 bits, and so-called noise shaping is adopted as a quantization noise control method capable of securing sufficient sound quality even with the 16 bits. In this embodiment, a technique called SMB (Super Bit Mapping: trademark) is used as the noise shaping. The above SMB is described in, for example, JP-A-2-020812 and JP-A-2-185.
This is one of the noise shaping methods described in detail in JP-A-52-182 and JP-A-2-18556, and is a well-known technique. Therefore, its detailed contents will not be described here.

However, when the noise shaping method such as the above SMB is adopted, the signal is compared with a case where the rounding process by simple rounding is used as a method for converting the 20-bit input digital audio signal into 16 bits. There is a problem that the amplitude increases. This is because noise in the high frequency range increases due to the noise shaping of the SMB. An increase in the signal amplitude in this way means that the area to be reserved for the sub-band increases. As described above, when the area to be reserved for the signal of the sub-band increases, the signal of the sub-band may not be recorded on a medium such as a CD.

Therefore, in the embodiment of the present invention, the signal of the sub-band is subjected to a filtering process by a low-pass filter (LPF) so that the amplitude of the signal of the sub-band is rounded by the simple rounding. In this case, the amplitude is set to be substantially the same as that in the case of (1), and an increase in the area to be secured for the sub-band is suppressed.

A recording / reproducing apparatus as an embodiment of the digital signal processing method and apparatus according to the present invention for realizing the above-mentioned features will be described in detail below.

FIGS. 1 and 2 show the configuration of the recording system of the recording / reproducing apparatus. Here, a configuration for recording a signal on a CD will be described. In the configuration shown in FIGS. 1 and 2, when manufacturing a package medium such as a CD that does not require much real-time properties, the state of all signals is first checked, and then the overall balance is determined according to the state. It employs a multi-pass approach to take. 1 and 2 show examples of the configuration of two paths. FIG. 1 shows a configuration for performing signal processing in the first pass in the recording system, and FIG. 2 shows a configuration for performing signal processing in the second pass. Here, the real-time property is the real-time property at the stage of manufacturing a final master disk after a process such as editing.

First, the signal processing of the first pass shown in FIG. 1 will be described.

In FIG. 1, the audio signal input to the input terminal 50 is a 20-bit wideband digital audio signal having a sampling frequency of 44.1 kHz and a quantization bit number of more than 16, as described above.

The digital audio signal input to the input terminal 50 is input to the unit extracting circuit 1. The unit cutout circuit 1 generates a unit in which a plurality of samples of the input digital audio signal are collected as a predetermined unit for handling in the subsequent processes. Although the size of the unit is arbitrary, in the CD audio standard, the left channel (Lch) and the right channel (Rch) are called 2352 bytes in total, that is, 588 samples per channel are called in units of blocks. It is desirable to adjust the size to that. Further, in the subsequent processing, Lch and Rch may be handled independently, or Lch and Rch may be handled collectively. Here, the size of one unit is 294 samples, which is half of the above-mentioned 588, independent of the channel.

The samples compiled for each unit in the unit extracting circuit 1 are input to a required word length calculating circuit 2. The required word length calculation circuit 2 calculates the minimum required word length of each unit. That is, the word length is a word length that is required auditorily as described above, and if the word length has accuracy, the word length is included in the signal when the signal is reproduced later. The word length is such that quantization noise is not audibly perceived as deterioration. FIG. 3 shows the required word length calculation circuit 2
The following shows a specific configuration example.

In FIG. 3, the signal flow is divided into two parts. One is a path for obtaining an allowable noise amount based on auditory characteristics, and the other is a quantization noise amount obtained by shortening a word. Is a path that changes Note that the allowable noise amount is an auditory threshold value that, when a certain sound is sounding, if the noise amount is equal to or less than that amount, the noise cannot be heard by human ears.

The unitized audio signal supplied to the input terminal 51 of the required word length calculating circuit 2 shown in FIG. 3 is sent to the allowable noise amount calculating circuit 21 of the path for obtaining the allowable noise amount from the auditory characteristics. Sent. In the permissible noise amount calculation circuit 21, from the input audio signal,
For each critical band (a frequency band divided in consideration of human auditory characteristics and divided into, for example, 25 bands), an allowable noise amount Ti as an allowable noise amount is obtained from the above auditory characteristics. This permissible noise amount Ti is actually a so-called minimum audible curve (a threshold value of the sensitivity of the human ear).
And so-called masking (a phenomenon that a certain signal cannot be heard by another signal) spectrum is synthesized. For details on how to find it,
For example, Japanese Patent Application No. Hei 7-14774 previously proposed by the present applicant.
This is described in detail in the specification, drawings, and the like of No. 2, and since these technologies are already known, detailed description thereof will be omitted here. The signal of the allowable noise amount Ti is sent to the judgment circuit 24.

The unitized audio signal supplied to the input terminal 51 of the necessary word length calculation circuit 2 shown in FIG. 3 is obtained by rounding n bits in a path for shortening the word and changing the amount of quantization noise. It is also sent to the circuit 19. Here, the audio signal supplied to the input terminal 51 is a signal of a main band, and is a full word (here, 2
0 bit). The n-bit rounding circuit 19 rounds (ie, rounds) the 20-bit signal supplied from the input terminal 51 to n bits. The n-bit rounding circuit 19 has an n +
The value of n is controlled by one circuit 25. At this time, the first value of n is preferably set to a value which should be the shortest word, for example, about 12 bits. It has been done.

The signal rounded by the n-bit rounding circuit 19 is sent to an adder 20. This adder 2
To 0, the audio signal from the input terminal 51 is also supplied, the signal from the n-bit rounding circuit 19 is input as a subtraction signal, and the audio signal from the input terminal 51 is input as an addition signal. Therefore, the adder 2
0, the audio signal and the n-bit rounding circuit 1
9 is obtained. This adder 20
Is the rounding error in the n-bit rounding circuit 19, that is, the quantization noise. The error signal from the adder 20 is input to an FFT (Fast Fourier Transform) circuit 22.

In the FFT circuit 22, the spectral component of the error signal, that is, the spectral component of the quantization noise is obtained, and the signal of the spectral component of the quantization error is sent to the maximum value detecting circuit 23.

The maximum value detection circuit 23 obtains the maximum value Ni of the spectral component for each critical band from the signal of the spectral component of the quantization error. The signal of the maximum value Ni of the spectrum component obtained by the maximum value detection circuit 23 is sent to the judgment circuit 24.

In the decision circuit 24, an allowable noise amount Ti obtained from a path for obtaining an allowable noise amount from the above-described auditory characteristics and a spectrum obtained from a path for shortening a word to change the quantization noise amount are obtained. Is compared with each other for each critical band, and in any critical band, Ti>
Check whether it becomes Ni. Here, if Ti> Ni in all the critical bands, the determination circuit 24 outputs the value of n as the necessary word length (number of bits). The rounded signal at that time is also output. On the other hand, if at least one critical band in which Ti <ni exists, the determination circuit 24 outputs a signal indicating that to the n + 1 circuit 25.

In the n + 1 circuit 25, the judgment circuit 2
4, the n-bit rounding circuit 19
The value of n in is increased by one bit (the word length is increased)
Is performed. As a result, the error signal is obtained again by the n-bit rounding circuit 19 and the adder 20.

In the example shown in FIG. 3, the error signal is obtained by rounding the signal supplied to the input terminal 51 by n bits. However, when the word becomes shorter by one bit, the quantization noise level increases by about 6 dB. From the allowable noise amount Ti to the minimum value Tm
in may be obtained, and the word length may be calculated backward from the minimum value Tmin.

The required word length obtained as described above is the actually required word length described above. The information on the required word length and the signal rounded to the required word length are output from the output terminal. 52 and is input to the word length check circuit 4 in FIG.

Returning to FIG. 1, the word length check circuit 4 checks whether the required word length obtained by the required word length calculation circuit 2 in FIG. 3 is 16 bits or more.

Here, if the required word length supplied from the required word length calculation circuit 2 is 16 bits or more, the CD standard will be exceeded if sufficient sound quality is to be maintained, and recording will not be possible. . When the required word length is 16 bits or more, the signal supplied from the required word length calculation circuit 2 and rounded to the required word length is converted into a 16-bit word length signal (that is, a main band signal). Signal) and a difference signal obtained by subtracting the signal in the main band from the signal rounded to the required word length, that is, a signal in a sub-band.

For this reason, when the word length check circuit 4 determines that the required word length is 16 bits or more, the word length check circuit 4 converts the audio signal (the signal rounded to the required word length). It is sent to the rounding circuit 3 and the adder 5.

At this time, the rounding circuit 3 further converts the signal rounded to the required word length of 16 bits or more into 16 bits.
Round to bit length. The signal rounded to a 16-bit length by the rounding circuit 3 is a signal in the main band. Here, as a method of the rounding processing in the rounding circuit 3, there is a method of rounding off, a method of rounding down, and the like. Deterioration is inevitable. Therefore, the rounding circuit 3 controls the quantization noise by using the above-described noise shaping method as a method of rounding to a 16-bit length while preventing sound quality deterioration as much as possible. This noise shaping prevents the deterioration of auditory sound quality by pumping the quantization noise to the high frequency side, and uses the so-called SBM (Super Bit Mapping) method as one method for realizing this. ing.

The signal rounded to 16 bits by the rounding circuit 3 using noise shaping as described above is sent to the adder 5. The adder 5 is also supplied with a signal rounded to the required word length of 16 bits or more from the word length check circuit 4, and uses the output of the round circuit 3 as a subtraction signal and the word length check circuit. An addition operation is performed using the signal from the circuit 4 as an addition signal. As a result, the adder 5 calculates the difference between the signal before being input to the rounding circuit 3 and the signal after the rounding processing in the rounding circuit 3. The output of the adder 5 is a sub-band signal.

The 16-bit main band signal and the sub-band signal obtained as described above are combined with the necessary word length information indicating 16 bits or more together with the primary recording device 6.
Sent to

On the other hand, when the required word length supplied from the required word length calculation circuit 2 is less than 16 bits, the word length check circuit 4 rounds the required word length from the required word length calculation circuit 2 to the required word length. The resulting signal is output as a signal in the main band, and information on the required word length at this time is also output. These signals are sent to the primary storage device 6. That is, in this case, the sub-band is omitted.
In addition, for a signal of a unit whose required word length may be less than 16 bits, a signal sample in the unit is placed on the LSB side in a second pass configuration described later.
The sub-band signal formed when the required word length becomes 16 bits or more is embedded. Also, for a signal of a unit whose required word length may be less than 16 bits, it is necessary to adjust the word length to the specified word length, that is, 16 bits of the CD standard later. It is desirable to ensure sufficient sound quality by performing shaping processing.

The signals of the main band and the sub-band and the word length information (information of the required word length) required in each unit obtained by the above processing are recorded on a recording medium by the primary recording device 6. . This primary recording device 6 uses a computer-based system as a recording medium.
(A device using a hard disk or MO (magneto-optical disk)), or a device using a tape such as a so-called DAT (digital tape recorder). The information recorded by the primary recording device 6 is passed to the configuration of the second pass shown in FIG.

Next, the configuration of the second pass in FIG. 2 will be described.

From the primary recording device 7 (corresponding to the primary recording device 6 in FIG. 1), the signal recorded on the recording medium in the first pass configuration is read.

In the configuration shown in FIG. 2, if the required word length information read from the primary recording device 7 is 16 bits or more, the required word length may be less than 16 bits. On the LSB side, a sub-band signal formed when the 16 bits or more is required is embedded.

For this reason, in the embedding position search circuit 11, the necessary word length information of the unit having the required word length of 16 bits or more (for example, the required word length information of the i-th unit) is used for the sub-band. The required word length is determined, and the required word length is 16
From the required word length information of each unit less than bits (for example, the required word length information of the k-th unit),
A unit that can afford to embed the sub-band signal is searched.

In this case, the unit whose required word length is 16 bits or more (the i-th unit)
It is necessary to determine how much the interval between the unit and the unit (the k-th unit) in which the signal of the sub-band of the unit is to be embedded is allowed. That is,
Recording a signal of a sub-band in a unit that is separated in time or in a different channel means that a load is imposed on data reading speed and buffering. Therefore, if a plurality of embeddable units are found when searching for a unit to embed the sub-band, it is better to select a unit that is as close in time as possible. Also, the search range is determined in consideration of the size of the buffering memory of the CD reproducing apparatus, the data reading speed, and the like.

In consideration of the above, the embedding position search circuit 11 sequentially reads out necessary word length information from the primary recording device 7 and performs a search. When the embedding position is determined, the embedding position search circuit 11 returns information indicating the position to the primary recording device 7, and the primary recording device 7 outputs the main band signal of the unit (the k-th unit) corresponding to the embedding position. Is output. Note that the signal of the sub-band and the signal of the main band (the k-th unit) in which the signal of the sub-band is embedded need not be one-to-one, and the signal of one sub-band is converted into a signal of a plurality of main bands. (Unit), or a plurality of sub-band signals may be embedded in one main band signal (one unit). This embedding method is described in Japanese Patent Application No.
Although it is described in detail in the specification and the drawings of 147742, since it is different from the point of the present invention, the details are omitted here.

The sub-band signal embedded and recorded in the main-band signal is composed of the main-band signal component noise-shaped by the rounding circuit 3 and the signal before being input to the rounding circuit 3 as described above. As shown in FIGS. 4 and 5, the signal component at this time also has a large signal amplitude because the quantization noise in the high frequency band is increased as compared with the case where rounding is performed. Has become. FIG. 4A shows a signal component of a main band when noise shaping is performed, and FIG. 4B shows a quantization noise spectrum when noise shaping is performed.
FIG. 5A shows the signal component of the main band when rounded off, and FIG. 5B shows the quantization noise spectrum when rounded off. Here, when a noise shaping method such as the SBM is adopted, the word length is increased by about 23 bits, and a large recording area that requires a sub-band to be recorded is required. That is, for example, when the required word length is, for example, 18 bits in the word length check circuit 4 having the configuration of the first pass, if the signal is subjected to a simple rounding-off process, The magnitude of the band signal can be represented by a 2-bit word length,
The size of the signal in the sub-band when the noise shaping processing is performed has a word length of, for example, 45 bits.

Therefore, in the configuration of the second pass, the high-frequency component of the sub-band is filtered and discarded to reduce the amplitude of the signal of the sub-band when the above-mentioned noise shaping processing is performed. The amplitude is made equal to the amplitude of the signal in the sub-band when the rounding process is performed.

Therefore, in the configuration of the second pass, L
A low-pass filter such as PF15 is used. However, when a low-pass filter such as the LPF 15 is used, when a CD on which a signal is recorded in the configuration of the recording system of the present embodiment is reproduced later, the high-frequency component is deteriorated. However, in the present embodiment, the noise component is moved to a part that is hardly perceivable by the noise shaping, so that there is no significant effect on the perception.

Although various types of the low-pass filter can be considered, a simple FIR (Finite Impulse Response d
It is sufficient to use a digital filter of the digital filter type. However, it is necessary to pay attention to the sample delay (delayed by the order of the filter) caused by filtering.

Although the filter coefficient here may be fixed, it is effective to use a minimum number of filters in order to make the waveform reproducibility during reproduction as faithful as possible. For this reason, in the configuration of FIG. 2, after the signal state before filtering is checked by the power calculation circuit 9, RO patterns in which several patterns of filter coefficients are stored in advance are stored.
The most appropriate filter coefficient is read from M13 in accordance with the output of the power calculation circuit 9, and the LPF 15 is filtered based on this filter coefficient.

The power calculating circuit 9 for obtaining information for selecting the filter coefficients obtains the maximum value of the signal and the power of the unit (sum of squares of each sample) as information for controlling the filter coefficients. It can be. Further, in the power calculation circuit 9, FF is added to the signal.
The frequency component (spectral component) of the signal sample may be obtained by performing T, and the maximum value or the power thereof may be obtained as information for controlling the filter coefficient.

The filter coefficient need not be obtained before the filter processing by the LPF 15 is performed. For example, as shown in FIG. 6, the maximum value of the signal amplitude in the power calculation circuit 10 is calculated from the signal after the filter processing in the LPF 15. Is checked, the filter coefficient is read out from the ROM 13 in which several patterns of filter coefficients are stored in advance in accordance with the output of the power calculation circuit 10, and the LPF 15 is filtered based on the filter coefficient. It is also possible to adopt a method of repeating the process until the process is performed. In addition,
6, components corresponding to those in FIG. 2 are denoted by the same reference symbols, and detailed operations thereof will be omitted.

FIG. 7 shows the signal waveform of the sub-band signal obtained by filtering as described above ((a) of FIG. 7).
And an example of the spectrum component ((b) of FIG. 7). In the example of FIG. 7, a low-pass filter having a cutoff frequency around 15 kHz is used.

The sub-band signal obtained as described above, the embedding position information obtained by the embedding position search circuit 11 (embedding position information of the k-th unit), and the information read from the primary recording device 7 are read. The main band signal and the word length information are stored in a bit packing circuit 17.
Is input to

The bit packing circuit 17 generates data according to the CD format from the supplied information. That is, in the bit packing circuit 17, the LSB of the signal sample of the main band of the k-th unit determined by the embedding position search circuit 11 is obtained.
The sub-band signal of the ith unit is embedded on the side to complete audio data recorded on the medium. Here, if the medium is a CD, the required word length information can be recorded as so-called subcode information.

The sub-code information has 96 bytes for 2352 bytes of the audio signal. As shown in FIG. , R, S, T, U, V,
W, and as shown in FIG.
The portion of S, T, U, V, W is 96 × 6/8 bytes, that is, 576 bits can be used as user data. The user data is also divided into four equal parts as shown in FIG. 8B, each of which is called in a unit called a pack, and these packs have a format as shown in FIG. 8C. In the above-described example, since 294 samples (294 samples / ch) are searched as one unit per channel, the required word length information is recorded in one pack of the subcode information. 1
Packs are more detailed, with instructions (IN, IN, etc.) indicating the mode, item, and type of data.
STRUCTION), a parity bit for data error correction, and a data field. Of those, 96 bits of the data field are actually usable.

In the embodiment of the present invention, one unit has 294 samples per channel.
An available data field, ie, 96 bits, in the pack is allocated for the required word length information. In addition, the data field includes the necessary word length information itself and: Information indicating whether to add a sub-band is also recorded. 2. Information such as embedding position information is also recorded. If a signal of one sub-band is recorded in a plurality of units, or a signal of a plurality of sub-bands is recorded in one unit, it is necessary to record a plurality of these sets.

In the above-described example, an example has been described in which the difference signal is directly recorded as the sub-band signal. However, in order to record the signal more efficiently, it is desirable to compress the signal. Further, since the amount of information to be recorded is reduced when the signal is compressed, it is possible to make the characteristics of the low-pass filter gentler and to reduce the deterioration occurring in the high-frequency component. In other words, if the characteristics of the low-pass filter are made gentler, the amount of information will increase,
By compressing the signal, the amount of information that is actually recorded does not increase, and it is also possible to reduce deterioration that occurs in the high frequency component.

Note that signal compression includes reversible coding that can completely restore the original signal upon decoding, and irreversible coding in which some information is lost. A typical example of lossless coding is entropy coding such as Huffman coding. Regarding Huffman coding, see the document "A Method for Constructionof
Minimum Redundancy Codes ", DA Huffman, Proc.IR
E., 40, p. 1098 (1952)). As for entropy coding, besides Huffman coding, the literature "A Universal Algorithm for Sequent Data Compression"
ial Data Compression ", J. Ziv, A. Lempel, IEEE Trans. o
n Inform. Theory, Vol.IT-23, No.3, pp.337-343, 197
7) and the literature "Buffer Overflow in Variable Length Coding o"
f Fixed Rate Sources ", F.Telinek, IEEE Trans. Info
rm. Theory, Vol. IT-14, No. 3, pp. 490-501, 1968) can be used. As the irreversible coding, there is a method of compressing the amount of information by reducing auditory redundant signals by utilizing the characteristics of hearing. This has been proposed by the present applicant in the specification and drawings of Japanese Patent Application No. 7-147742. Also, LPC (Linear Predictive Coding), ADPCM
It is also possible to perform coding by non-linear quantization such as, or by vector quantization, and to further encode these codes by entropy coding. The above LPC is described by Itakura and Saito in "Speech Decomposition / Synthesis Transmission Method Using Maximum Likelihood Spectrum Estimation", Proc. Of the Acoustical Society of Japan, pp.23.
1,1967 or the document "Predictive coding of audio signals"("Pred
ictive Coding of Speech Signals ", BS, Atal, MRSch
roeder, Reports of 6th Int.Conf.Acoust., C-5-4,196
8) is described in detail. In addition, there are many documents on calculation algorithms, and a description thereof will be omitted.

The signal encoded as described above is thereafter recorded on a CD standard disc.

Next, FIG. 9 shows the configuration of the reproducing system of the recording / reproducing apparatus according to the embodiment of the present invention. In this reproduction system, basically, the processing of the recording system is followed in reverse. Note that FIG.
5 shows an example in which a signal in the sub-band is recorded as a signal in the main band of the k-th unit separated in time. When a sub-band signal is extracted from a main-band signal of a temporally separated unit, it is necessary to read data from a medium (CD) faster than usual.

In FIG. 9, a signal read from a medium (CD) is divided into an audio signal and subcode information by a data read circuit 26.
The audio signal is stored in the buffer memory 28.
This is because it is necessary to extract the signal of the sub-band divided from the main band of the i-th unit from the signal of the main band of the k-th unit which is temporally separated. Although the storage capacity of the buffer memory 28 cannot be unconditionally determined because of the cost of the memory, for example, a so-called shock-proof memory, which is a so-called shock-proof memory mounted on a reproducing apparatus such as a so-called portable CD player. Has a memory capacity enough to buffer sample data for about 10 seconds, and a storage capacity equivalent to or about twice that amount is acceptable in terms of cost. Since the above-described information such as the position where the signal of the sub-band is embedded and the required word length are recorded in the sub-code information, the information is extracted from the sub-code information.

The required word length information recorded in the subcode information is sent to the status check circuit 27. In this state check circuit 27, the necessary word length information is 1
If the length is longer than 6 bits, it is determined that the unit requires a sub-band, and the shorter unit is only the main band.

When the word length is longer than 16 bits,
The sub-band extracting circuit 31 is connected to the buffer memory 28
The sub-band signal is extracted from the signal (the signal of the main band of the k-th unit) stored in. That is, in the sub-band extracting circuit 31, the data reading circuit 26
The embedding position information is obtained and sent to the buffer memory 28, and the main band signal of the k-th unit in which the sub-band signal is embedded is read from the buffer memory 28 based on the embedding position information. At this time, since the sub-band signal is embedded in the LSB side of the main band signal of the k-th unit as described above, the sub-band signal is read from the LSB side of the main band signal. When the sub-band signal is subjected to high-efficiency compression such as entropy coding, the sub-band extracting circuit 31 also performs decompression corresponding to the high-efficiency compression, and converts the time-sequence sub-band signal. Restore.
At this time, the main band extracting circuit 29 also extracts from the buffer memory 28 the main band signal of the i-th unit divided from the sub-band in the recording system.

The sub-band signal extracted from the k-th unit and the main-band signal of the i-th unit as described above are added by the adder 32 and sent to the D / A converter 33, where the analog audio signal is output. After being converted into a signal, the signal is output from the output terminal 60. Here, since the signal obtained by the adder 32 has been subjected to a low-pass filter on the sub-band, the portion below the cut-off frequency of the filter as shown in FIG. , And beyond that the signal has a quantization noise level due to noise shaping.

On the other hand, in the case of a unit whose required word length may be shorter than 16 bits, the main band extracting circuit 30 extracts only the signal of the main band of the unit from the buffer memory 28, and After being sent to the A converter 33 and converted into an analog audio signal, it is output from the output terminal 60. Here, when the signal is passed to the D / A converter 33, the read main band signal may be output as it is. However, when the sub band signal of another unit is recorded in the main band signal. In this case, it is possible to reduce the deterioration of the sound quality by outputting the portion with zeros.

In the above example, only the direction in which the number of quantization bits of the CD is increased has been described. However, a similar method can be used in the direction in which the sampling frequency is increased.

The present invention can be used not only for CDs but also for digital audio package media such as DAT and transmission systems such as broadcasting. The present method can be used not only for audio signals but also for image signals.

As described above, according to the embodiment of the present invention,
Broadband signals such as digital audio signals
In the case of recording and reproducing on a package medium such as the above, the signal of the entire band of the audio signal is divided into a signal of a main band and a signal of a sub band, and the signal of the main band is subjected to, for example, straight PCM, High-efficiency coding with entropy coding and nonlinear quantization, etc.
Obtaining the number of bits that are auditory redundant from the straight PCM word of the signal in the main band and allocating the filtered sub-band signal to the redundant bits reduces the deterioration of the signal in the main band, In addition, signals in the sub-band can be efficiently recorded or transmitted and reproduced.

That is, according to the embodiment of the present invention, since the compatibility with the conventional CD standard is maintained, in the conventional reproducing system, the signal quality of the main band is close to full word or higher. In the system adopting the present invention, in addition to the signal of the main band, the signal of the sub-band which has been encoded with high efficiency can also be reproduced, so that a sound with higher sound quality can be reproduced.

[0079]

As is apparent from the above description, in the present invention, the signal of the entire band of the digital signal is divided into the signal of the main band and the signal of the sub-band, and the required word length is less than a predetermined value. The signal of the main band is obtained by filtering the signal of the sub band, and adding the filtered signal of the sub band to the signal of the main band having the required word length less than a predetermined value. Signals in important sub-bands can be recorded on a recording medium or transmitted to a transmission medium while preventing signal deterioration.

[Brief description of the drawings]

FIG. 1 shows one of recording systems of a recording / reproducing apparatus according to an embodiment of the present invention.
FIG. 3 is a block circuit diagram illustrating a configuration of a pass.

FIG. 2 shows two of the recording systems of the recording and reproducing apparatus according to the embodiment of the present invention.
FIG. 4 is a block circuit diagram illustrating a configuration example of a path.

FIG. 3 is a block circuit diagram showing a specific configuration of a required word length calculation circuit in the configuration of the first pass.

FIG. 4 is a diagram showing a noise-shaped main band signal component and a signal component (quantization noise spectrum) of a difference signal before and after noise shaping.

FIG. 5 is a diagram illustrating a signal component of a main signal after rounding and a signal component of a difference signal before and after rounding (quantization error spectrum).
FIG.

FIG. 6 shows two of the recording systems of the recording / reproducing apparatus according to the embodiment of the present invention.
FIG. 18 is a block circuit diagram illustrating another configuration example of the pass.

FIG. 7 is a diagram showing a signal waveform and a spectrum component of a sub-band signal recorded on a medium in a recording system of the recording / reproducing apparatus of the present invention.

FIG. 8 is a diagram used for describing a subcode information format.

FIG. 9 is a block circuit diagram showing a configuration of a reproducing system of the recording / reproducing apparatus according to the embodiment of the present invention.

FIG. 10 is a diagram showing a quantization noise spectrum when a signal of a sub-band is added to a signal of a main band in a reproducing system of the recording / reproducing apparatus of the embodiment of the present invention.

[Explanation of symbols]

1 unit cutout circuit, 2 required word length calculation circuit, 3 rounding circuit, 4 word length check circuit, 5
Adder, 6,7 primary recording device, 9,10 power calculation circuit, 11 embedding position search circuit, 13 coefficient ROM, 15 LPF, 17 data packing circuit, 19 n-bit rounding circuit, 20 adder, 21
Allowable noise amount calculation circuit, 22 FFT circuit, 23
Maximum value detecting circuit, 24 judging circuit, 25 n + 1 circuit, 26 data reading circuit, 27 state checking circuit, 28 buffer memory, 29, 30 main band extracting circuit, 31 sub band extracting circuit,
32 adder, 33 D / A converter

Claims (67)

[Claims] 1. A signal of a main band whose required word length is less than a predetermined value is obtained by dividing a signal of a whole band of a digital signal into a signal of a main band and a signal of a sub band. And adding the filtered sub-band signal to the main band signal having the required word length less than a predetermined value. 2. The signal of the main band is a noise-shaped signal, and the signal of the sub-band is a difference signal between the signal of the entire band and the signal of the main band.
The digital signal processing method according to the above. 3. The signal of the main band is a signal rounded off and rounded, and the signal of the sub-band is a difference signal between the signal of the entire band and the signal of the main band. Item 1
The digital signal processing method according to the above. 4. The digital signal processing method according to claim 1, wherein the signal in the sub-band is a signal including a higher frequency component than the signal in the main band. 5. The digital signal processing method according to claim 1, wherein information indicating the properties of the signal components of the sub-band is added to the additional information constituting a part of the signal of the main band. 6. The digital signal processing method according to claim 2, wherein information indicating the property of the signal component of the sub-band is added to the additional information constituting a part of the signal of the main band. 7. The digital signal processing method according to claim 3, wherein information indicating the property of the signal component of the sub-band is added to the additional information constituting a part of the signal of the main band. 8. The digital signal processing method according to claim 4, wherein information indicating the properties of the signal components of the sub-band is added to additional information constituting a part of the signal of the main band. 9. The digital signal processing method according to claim 1, wherein said filtering is performed by a low-pass filter. 10. The signal of the sub-band is a straight P signal.
2. A CM-coded code.
The digital signal processing method according to the above. 11. The digital signal processing method according to claim 1, wherein the signal of the sub-band is an entropy-coded code. 12. The signal of the sub-band is auditory and / or
2. The digital signal processing method according to claim 1, wherein the code is a code which has been efficiently coded using visual information. 13. The digital signal processing method according to claim 1, wherein said sub-band signal is a code subjected to irreversible compression coding. 14. The signal of the sub-band is a straight P signal.
2. The digital signal processing method according to claim 1, wherein the code is a code obtained by further entropy coding a code that has been CM-coded. 15. The signal of the sub-band is auditory and / or
2. The digital signal processing method according to claim 1, wherein the code is a code obtained by further entropy coding a code that has been highly efficient coded using visual information. 16. The digital signal processing method according to claim 1, wherein the sub-band signal is a code obtained by further performing entropy coding on a code subjected to irreversible compression coding. 17. The apparatus according to claim 10, wherein information indicating that said straight PCM coding is used is added to additional information constituting a part of said main band signal.
The digital signal processing method according to the above. 18. The digital signal processing method according to claim 11, wherein information indicating that the entropy coding is used is added to additional information constituting a part of the signal of the main band. 19. A method of adding information indicating that high-efficiency encoding using auditory and / or visual information has been used to additional information constituting a part of the signal of the main band. 13. The digital signal processing method according to claim 12, wherein 20. The digital signal processing method according to claim 13, wherein information indicating that the irreversible compression encoding has been used is added to additional information constituting a part of the signal of the main band. . 21. The apparatus according to claim 14, wherein information indicating that the straight PCM coding and the entropy coding are used is added to additional information constituting a part of the signal of the main band. Digital signal processing method. 22. Information indicating that high-efficiency coding and entropy coding using auditory and / or visual information have been used is added to additional information constituting a part of the signal of the main band. The digital signal processing method according to claim 15, wherein 23. An apparatus according to claim 16, wherein information indicating that said irreversible compression coding and entropy coding are used is added to additional information constituting a part of said main band signal. Digital signal processing method. 24. The digital signal processing method according to claim 1, wherein whether to perform the filtering is determined adaptively from the state of the signal in the sub-band. 25. The digital signal processing method according to claim 1, wherein whether to perform the filtering is determined adaptively from the state of the signal in the main band. 26. The digital signal processing method according to claim 1, wherein whether or not to perform the filtering is adaptively determined from the state of the signals in all the bands before the division. 27. The digital signal processing method according to claim 24, wherein the determination is made according to the magnitude of the amplitude of the signal in the sub-band. 28. The method according to claim 25, wherein the determination is made according to the magnitude of the amplitude of the signal in the main band.
The digital signal processing method according to the above. 29. The digital signal processing method according to claim 26, wherein the determination is made according to the magnitude of the amplitude of the signal of the entire band before the division. 30. The digital signal processing method according to claim 24, wherein the determination is made according to the magnitude of the spectral component of the signal of the sub-band. 31. The digital signal processing method according to claim 25, wherein the determination is made according to the magnitude of the spectrum component of the signal of the main band. 32. The digital signal processing method according to claim 26, wherein the determination is made according to the magnitude of the spectral component of the signal of the entire band before the division. 33. The digital signal processing method according to claim 24, wherein said determination is made according to the magnitude of entropy of said sub-band signal. 34. The digital signal processing method according to claim 25, wherein said determination is made according to the magnitude of entropy of said signal of said main band. 35. The digital signal processing method according to claim 26, wherein said determination is made according to the magnitude of entropy of the signal of the entire band before said division. 36. An apparatus according to claim 1, wherein information indicating that the required word length is a main band exceeding a predetermined value is added to additional information constituting a part of the signal of the main band. The digital signal processing method according to the above. 37. The signal of the main band is a signal having a sampling frequency of 44.1 kHz and a quantization bit number of 16, and the signal of the sub-band is a signal of a band not less than the main band. 2. The digital signal processing method according to claim 1, wherein: 38. Dividing means for dividing a signal of the entire band of a digital signal into a signal of a main band and a signal of a sub-band, and a searching means for searching for a signal of a main band whose required word length is less than a predetermined value. Filtering means for filtering the signal of the sub-band, and adding means for adding the signal of the filtered sub-band to a signal of the main band whose required word length is less than a predetermined value. Digital signal processor. 39. A noise shaping means for noise-shaping the signal in the main band, and a difference between the signal in the entire band and the signal in the main band to generate the signal in the sub-band. 39. The digital signal processing device according to claim 38, further comprising: a sub-band signal generating unit. 40. The dividing means, comprising: a rounding means for rounding the signal of the main band to round the signal of the main band; 39. The digital signal processing apparatus according to claim 38, further comprising: a sub-band signal generation unit that performs the sub-band signal generation. 41. The digital signal processing apparatus according to claim 38, wherein the signal in the sub-band is a signal including a frequency component higher than the signal in the main band. 42. A property calculating means for calculating information representing the property of the signal component of the sub-band, wherein the adding means adds the additional information constituting a part of the signal of the main band to the additional information from the property calculating means. 39. The digital signal processing device according to claim 38, wherein information is added. 43. A property calculating means for calculating information representing the property of the signal component of the sub-band, wherein the adding means adds the additional information constituting a part of the signal of the main band to the additional information from the property calculating means. 40. The digital signal processing device according to claim 39, wherein information is added. 44. A property calculating means for calculating information indicating the property of the signal component of the sub-band, wherein the adding means adds the additional information constituting a part of the signal of the main band to the additional information from the property calculating means. 41. The digital signal processing device according to claim 40, wherein information is added. 45. A property calculating means for calculating information indicating the property of the signal component of the sub-band, wherein the adding means adds the additional information constituting a part of the signal of the main band from the property calculating means. 42. The digital signal processing device according to claim 41, wherein information is added. 46. The digital signal processing device according to claim 38, wherein said filtering means is a low-pass filter. 47. The adding means adds information indicating that the signal of the sub-band is a straight PCM-encoded code to the additional information constituting a part of the signal of the main band. The digital signal processing device according to claim 38, wherein 48. The method according to claim 48, wherein the adding means adds information indicating that the signal of the sub-band is an entropy-coded code to additional information constituting a part of the signal of the main band. 39. The digital signal processing device according to claim 38. 49. The adding means, as a part of the signal of the main band, displays information indicating that the signal of the sub-band is a code which has been encoded with high efficiency using auditory and / or visual information. 39. The digital signal processing device according to claim 38, wherein the digital signal processing device is added to the additional information constituting 50. The adding means adds information indicating that the signal of the sub-band is a code subjected to irreversible compression coding to the additional information constituting a part of the signal of the main band. 39. The digital signal processing device according to claim 38, wherein 51. An appending means for transmitting information indicating that a signal of the sub-band is a code obtained by further entropy-encoding a code of straight PCM encoding, and The digital signal processing device according to claim 38, wherein the digital signal processing device is added to information. 52. The above-mentioned addition means, as for the above-mentioned sub-band, the information which shows that it is the code which further entropy-encodes the code which has been highly efficient coded using auditory and / or visual information, 39. The digital signal processing apparatus according to claim 38, wherein the digital signal processing apparatus is added to additional information constituting a part of a main band signal. 53. The addition means comprises information indicating that the signal of the sub-band is a code obtained by further performing entropy coding on a code subjected to irreversible compression coding, and constitutes a part of the signal of the main band. 39. The digital signal processing device according to claim 38, wherein the digital signal processing device is added to the attached information. 54. The digital signal processing apparatus according to claim 38, further comprising: a determination unit that adaptively determines whether to perform the filtering based on a state of a signal in a sub-band. 55. The digital signal processing apparatus according to claim 38, further comprising: a determination unit that adaptively determines whether to perform the filtering based on the state of the signal in the main band. 56. The digital signal processing apparatus according to claim 38, further comprising: a decision unit for deciding whether to perform said filtering adaptively based on the state of signals in all bands before said division. 57. The digital signal processing apparatus according to claim 54, wherein said determination means makes said determination according to the magnitude of the amplitude of the signal in the sub-band. 58. The digital signal processing apparatus according to claim 55, wherein said determination means makes said determination according to the magnitude of the amplitude of said signal in said main band. 59. The digital signal processing apparatus according to claim 56, wherein said determining means makes said determination according to the magnitude of the amplitude of the signal of the entire band before said division. 60. The digital signal processing apparatus according to claim 54, wherein said determining means makes said determination according to the magnitude of the spectral component of the signal of said sub-band. 61. The digital signal processing apparatus according to claim 55, wherein said determining means makes said determination according to the magnitude of the spectral component of the signal of said main band. 62. The digital signal processing apparatus according to claim 56, wherein said determining means makes said determination based on the magnitude of the spectral component of the signal of the entire band before said division. 63. The digital signal processing apparatus according to claim 54, wherein said determining means makes said determination according to the magnitude of entropy of the signal of said sub-band. 64. The digital signal processing apparatus according to claim 55, wherein said determining means makes said determination according to the magnitude of entropy of the signal of said main band. 65. The digital signal processing apparatus according to claim 56, wherein said determining means makes said determination according to the entropy magnitude of the signal of the entire band before said division. 66. The adding means adds information indicating that the required word length is a main band exceeding a predetermined value to additional information constituting a part of the signal of the main band. The digital signal processing device according to claim 38, wherein 67. The signal of the main band is a signal having a sampling frequency of 44.1 kHz and a quantization bit number of 16, and the signal of the sub-band is a signal of a band not less than the main band. Claim 38.
The digital signal processing device according to claim 1.