JP2002042416A - Voice signal recording method, its transmitting method, its recording device, its transmitting device, its recording medium and its transmitting medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve such a problem that a signal is instantaneous one which may be imperceivable for some person or in a case of some sound source since the signal in a level imperceivable as an original sound is singly added to a certain prescribed signal so that a sufficient effect to prevent copying is not obtained in a conventional manner. SOLUTION: A copy disturbing signal is generated by multiplying a making level by a level adjusting signal α which is set at each sub-band as a coefficient (step S13). The level adjusting signal α has a numerical value being <=1 and a larger value is selected when a copy disturbance effect is preferential. The copy disturbance signal is added to the whole frequency band of the signal of an original frequency area (step S14). Then a copy preventing signal is added even to a band where an information amount in the original frequency area signal is not required. The frequency area signal after addition is performed is converted into the PCM signal of a time area, outputted (step S15) and recorded in a recording medium (step S16).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio signal recording method, a transmission method, a recording device, a transmission device, a recording medium, and a transmission medium. The present invention relates to a recording method, a transmission method, a recording device, a transmission device, a recording medium, and a transmission medium of an audio signal for restricting the operation.

[0002]

2. Description of the Related Art In recent years, commercially available devices or personal computer software capable of easily digitally copying copyrighted works such as music and movies have appeared, and copyright protection has been strongly demanded. In order to protect this copyright, it is necessary to regulate by law, to implement technical protection, and to improve morals of consumers. Especially if it can be protected technically, it will be most effective.

[0003] As copyright protection measures in this technical aspect, there are:
Conventionally, when a PCM signal from a linear PCM sound source such as a compact disc (CD) is compressed and copied to a mini disc (MD), SCMS (Serial Copy) is used.
As a management system, a copy prohibition flag is set on the digital interface of the CD player to prohibit copying.

That is, in the SCMS, a copy control flag is added to a digital interface (IEC958) signal of a CD player so that copying is possible.
Three state controls are possible: first generation allowed and prohibited. The target recorder is a digital device such as a digital audio tape recorder (DAT), an MD recorder, a CD-R recorder, or the like, which is specified by law.

However, in the SCMS, since the first generation copy is permitted, unlimited copying is possible by exchanging the recording medium on the recorder side. Further, since the original CD has a large amount of information, it is compressed by the recent development of high-efficiency coding technology, and is stored in a smaller recording device such as an MD or an IC memory. High quality copies can be easily made.

On the other hand, in order to prevent such a state that a duplicate can be easily produced, a new audio medium such as a DVD (Digital Versatile Disk) audio standardized as next-generation audio is copy-limited by encryption and a secret key. The technology to do is adopted. However, at present, the market is mainly CD-based, and it is difficult to change this situation for the time being. Further, with the development of information communication represented by the Internet, music distribution services have created a new market, and means for protecting the rights of these contents are urgently desired.

[0007] In response to such a request, in particular, in a copy generating means (MP3, MD, etc.) utilizing information compression,
There has been proposed a method of restricting the compression copy of an audio signal using an audibility characteristic, which is an important technical element for reducing the amount of information. In addition, a method of restricting the compression copy of an audio signal using an instantaneous masking effect in a time domain has been conventionally proposed (Japanese Patent Laid-Open No. 11-232777). In the conventional compressed copy restriction method described in Japanese Patent Application Laid-Open No. H11-232777, a specific signal is added to the original audio or image data of the recording medium serving as the copy source and recorded, and the original data is once compressed and then reproduced. In this case, the specific signal is emphasized and the reproduced signal is distorted.

[0008]

However, the above-mentioned conventional method of restricting the compression and copy of an audio signal adds a signal of a level which is not perceived as an original sound to a specific signal alone. This does not cause deterioration in sound quality after the compression encoding process, and is not perceived depending on the person or sound source because it is instantaneous, and the effect of preventing duplicates may be insufficient. Also, SC
The MS method prevents the generation of duplicates corresponding to grandchildren, and countless copies can be made for the first generation.

[0009] The present invention has been made in view of the above points,
By inserting a copy interference signal based on human auditory characteristics into the original sound source over the entire band, the signal quality after compression is significantly deteriorated, and a recording method, a transmission method, and a sound signal of an audio signal capable of preventing generation of an easy duplicate, It is an object to provide a recording device, a transmission device, a recording medium, and a transmission medium.

[0010]

In order to achieve the above object, a sound signal recording method according to the present invention provides a first method of converting a pulse code modulated PCM sound signal into a frequency domain signal per unit time. And dividing the frequency domain signal obtained in the first step into subbands each having a unit of a critical bandwidth corresponding to human frequency resolution, and calculating a masking level for this frequency domain signal for each subband A second step of performing a convolution operation on the masking level over the entire band to determine a final masking level in each subband, and calculate a noise signal below this masking level as a copy interference signal in each subband unit And a copy interference signal are added to the frequency domain signal obtained in the first step. A fourth step and a fifth step of converting the fourth PCM audio signal of the signal obtained with time domain step, PC obtained by the fifth step
Recording the M audio signal on a recording medium.

According to the present invention, a PCM audio signal in the time domain is converted into a PCM audio signal in the time domain by inserting a copy disturbing signal at the very end of the masking level into a band which is not originally necessary for the signal in the frequency domain. Bits are allocated to a band that is not originally required, and the bit amount allocated to a band that is originally required for audibility can be reduced so that a sufficient amount of information is not provided.

In order to achieve the above object, in the transmission method of the present invention, instead of the sixth step of the recording method, the PCM audio signal obtained in the fifth step is transmitted to a transmission medium. It was made.

In order to achieve the above object, a recording apparatus or a transmission apparatus according to the present invention comprises a pulse code modulated PCM.
A time-frequency conversion circuit for converting an audio signal into a frequency domain signal for each unit time, and a frequency domain signal obtained by the time-frequency conversion circuit in units of a critical bandwidth corresponding to human frequency resolution. A masking amount calculating circuit that divides into subbands, calculates a masking level for this frequency domain signal for each subband, performs a convolution operation on the masking level over the entire band, and determines a final masking level in each subband A copy interference signal generation circuit that generates a noise signal equal to or less than the masking level determined by the masking amount calculation circuit as a copy interference signal for each subband, and a copy interference signal that is output from the time-frequency conversion circuit. An addition circuit that adds the signal to the frequency domain signal and outputs the added signal.
A frequency-time conversion circuit for converting a signal output from the addition circuit into a PCM audio signal in the time domain, and a recording unit or a transmission medium for recording the PCM audio signal output from the frequency-time conversion circuit on a recording medium Output means.

According to the present invention, the copy interference signal at the very end of the masking level generated by the copy interference signal generation circuit is added to an originally unnecessary band of the frequency domain signal by the addition circuit, and then the time domain signal is added by the frequency-time conversion circuit. By recording after converting to a PCM audio signal, bits are allocated to a band that is not originally required at the time of compression processing of a PCM audio signal reproduced from a recording medium or a PCM audio signal received via a transmission medium. It is possible to reduce the amount of bits allocated to the perceived important band so that a sufficient amount of information is not provided.

In order to achieve the above object, a recording medium or a transmission medium according to the present invention uses a pulse code modulated PCM.
The audio signal is subjected to time-frequency conversion and then divided into sub-bands in units of a critical bandwidth corresponding to the frequency resolution of human beings, a masking level is calculated for each of these sub-bands, and the masking levels are convolved over the entire band. An operation is performed to determine a final masking level in each subband, and a PCM audio signal in which a noise signal below the masking level calculated for each subband is added as a copy interference signal is recorded or transmitted. The configuration is as follows.

According to the recording medium or transmission medium of the present invention, the reproduced or received PCM audio signal is assigned bits to a band which is not originally required at the time of compression processing at the time of copy recording, so that the originally required audibility is obtained. The amount of bits allocated to the important band can be reduced so that a sufficient amount of information is not provided.

[0017]

Next, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a flowchart for explaining the operation of an embodiment of a method for recording an audio signal according to the present invention. This embodiment is intended to restrict copying by artificially guiding the sound quality improvement process by MNR (masking level to quantization noise level ratio) in the wrong direction.

Before describing the embodiment of FIG. 1, the background of the present invention will be described. At present, the audio signal compression scheme actually used is ATRA.
C (adaptive transform acoustic coding), MP3 (MPE
G-1 Audio Layer 3), Dollby-Digital,
AAC and the like exist. All of these audio signal compression methods fall into the category called adaptive transform coding. The basic encoding process is to expand the PCM data on the time axis into the frequency domain by orthogonal transform, and convert the energy of the sound source into a frequency band (sub-band) in units of a critical bandwidth corresponding to human frequency resolution. Information compression is realized by performing adaptive bit allocation using the auditory characteristics of the frequency domain signal at the break.

The factors that determine the sound quality of these compression methods are:
It depends on whether the necessary information amount is correctly given to the necessary band. If this information amount (bit allocation) is incorrect, the sound quality will naturally deteriorate. As one of the methods for determining the bit allocation, a method using MNR (masking level to quantization noise level ratio) is adopted. This MN
The basic idea of R is that the absolute audible level of the auditory characteristics and the relative audible level generated by the sound source are limited to a certain band, and the masking level of that band (human moment within a limited time instantaneously) is determined by convolution. Energy in the frequency domain that is not perceived by the ear) is obtained, and if the quantization noise is below the masking level, it is not detected. It can be said that the higher the MNR, the higher the sound quality. This is obtained for each of the divided bands of the individual coding schemes, and by improving the overall MNR, effective information compression is realized.

FIG. 5 is a flowchart for explaining the operation of general bit allocation obtained by the above-mentioned MNR. After frequency expansion of a PCM sample group of a certain time unit length, the lower band according to the human frequency resolution (critical band) is combined into a narrower band, a masking level is calculated for each band, and a convolution operation is performed in all bands. To obtain a final masking curve (step S
1).

Subsequently, the ratio (MNR) between the masking level and the quantization noise level generated by the provisionally allocated quantization bit length is calculated for each band (step S2), and the MNR becomes the worst in all the bands. A band is searched (step S3). Next, the bit allocation is increased by a predetermined amount for the band having the worst MNR (step S
4), it is determined whether or not the coding amount of the entire band after the increase is less than the allowable information amount (step S5). The increase in the quantization bit length improves the MNR of the band.

If the coding amount is less than the allowable information amount, the ratio (MNR) of the masking level and the quantization noise level generated by the temporarily allocated quantization bit length is calculated again for each band (step S2), MNR in all bands
Search for the worst band (step S3), increase the bit allocation amount for the worst band with the searched MNR (step S4), and determine whether the increased coding amount is less than the allowable information amount (step S4). Step S5). Hereinafter, in the same manner as described above, the quantization bit length is sequentially increased from the band having the worst MNR in the entire band, and the usable information amount (bit amount) determined by the coding rate is not exceeded. The above operation is repeated. When the coding amount becomes equal to or larger than the allowable information amount, the bit allocation amount is determined (Step S).
6).

FIG. 6 is a comparison diagram showing an example of a masking level in a normal sound source and a quantization noise level generated by a given bit allocation. In the figure, I denotes a temporal portion of the audio signal from the original sound source, expands it on the frequency domain, finds the masking amount for each band, and obtains the final masking level after convolution processing. It is. II indicates the quantization noise level based on the result of bit allocation according to the flowchart of FIG. The sound source in FIG. 6 indicates a normal sound source in which a copy interference signal described later is not inserted.

FIG. 7 shows an example in which bits are allocated according to FIG. 5 in accordance with the MNR obtained from these. As shown in FIG.
It can be seen that bit allocation is concentrated near the high-level spectrum and in the middle frequency region where sensitivity is high due to the nature of hearing. It can also be seen that there is a band to which no bit is allocated. According to the bit allocation flowchart shown in FIG. 5, the ratio (distance) between the masking level and the quantization noise level is almost constant as shown in FIG. 6, except for the band where no signal exists.

By the way, in the above-mentioned bit allocation, if the information amount is not sufficiently allocated to the band which is originally required, the quantization noise exceeds the level which cannot be perceived and appears as annoying sound. By artificially guiding the sound quality improvement process by MNR in the wrong direction to induce a mechanism to pour the amount of information into unnecessary bands, it is possible to cause overall sound quality deterioration. .

Further, even if the signal has a level equal to or lower than the masking level, bits are not allotted at all, but the bit allocation amount is determined by the ratio of the quantization noise level of the band caused by encoding to the masking level. . Therefore, if a signal just below the masking level is inserted in the low or high band (frequency band is not so important, but in normal music many bits are allocated in the middle band), this band (low band) And high frequencies), and the amount of bits allocated to the originally required band (middle frequency) is reduced, resulting in deterioration of sound quality.

When the value of MNR indicates a good overall value, the corresponding sample in time (audio source)
Can be said to be excellent or easy to encode. Conversely, when the MNR indicates a poor overall value, the sound quality is degraded and it can be said that the audio source is difficult to encode. Therefore, it can be seen that to lower the quality of the coded sound, the value of the MNR should be lowered, that is, the quantization noise level should be raised in all bands. The quantization noise level can be raised by inserting a noise signal having a certain distance and not perceivable from the signal level masked for each band.

This embodiment has been made by focusing on the above points. Returning to FIG. 1, first, PCM data on the time axis is expanded into a signal in the frequency domain by orthogonal transformation (step S1). S11). Subsequently, the signal in this frequency domain is divided into frequency bands (sub-bands) in units of a critical bandwidth corresponding to human frequency resolution, and a masking level is calculated for each frequency band of this frequency domain signal. The amount of masking is convolved over the entire band to determine the total masking level in each subband. Further, the final masking level is determined by considering the absolute audible level based on human auditory characteristics for each sub-band and selecting the higher masked signal level (step S12).

Subsequently, the above masking level is multiplied by a level adjustment signal α set for each subband as a coefficient to generate a copy interference signal (step S13).
The above-mentioned level adjustment signal α has a numerical value of 1 or less. A small value is selected when sound quality is emphasized, and a large value is selected when priority is given to the copy disturbance effect.

Subsequently, the generated copy interference signal is added over the entire frequency band of the original frequency domain signal (step S14). As a result, the copy protection signal is added to the band in which the amount of information in the signal in the original frequency domain was not necessary. Then, the frequency domain signal after the addition is converted into a time domain PCM signal and output (step S15).

The output PCM signal is recorded on a recording medium such as a CD (step S16). And
The recorded recording medium is shipped from the factory to the market as a pre-recorded medium. It should be noted that this embodiment is a CD
May be provided on the output side of a recording medium reproducing apparatus.

FIG. 2 is a diagram showing an example of a comparison between a masking level in a sound source to which a copy interference signal of a degree undetectable by the present invention is added and a quantization noise level caused by a given bit allocation.

In FIG. 3, III extracts a certain temporal portion of a signal obtained by adding a copy interference signal to an audio signal from an original sound source, develops it on a frequency domain, and then calculates a masking amount for each band. This is the final masking level after the calculation and convolution processing. IV indicates the quantization noise level. As shown in FIG. 2, when the copy interference signal is added to the audio signal, compared with the comparison diagram of FIG. 6 relating to a normal sound source to which the copy interference signal is not added, the information consumption of the band that was not necessary conventionally increases, It can be seen that the quantization noise level curve IV relatively approaches the masking level curve III by reducing the quantization bit length of the important band.

That is, the level adjustment signal α used for generating the copy interference signal is set for each sub-band such that the quantization noise level curve IV approaches the masking level curve III as much as possible. However, when the audio signal is normally reproduced from the original CD, in order to avoid a malfunction of the reproducing apparatus, the level adjustment signal α is set so that the quantization noise level does not become too large near the upper limit frequency of the audible frequency and below the lower limit frequency. Is set.

FIG. 3 shows an example in which bits are allocated according to the embodiment of FIG. 1 in accordance with the MNR obtained from them. As shown in FIG. 3, the information consumption of a band that has not been necessary conventionally increases, and the quantization bit length of an important band decreases. That is, in the present embodiment, a sufficient amount of information is given to an important band in terms of audibility by making the copy protection signal sneak into a band where the amount of information in the signal in the original frequency domain was not necessary. As a result, sound quality degradation is detected by the listener.

As described above, according to the present embodiment, the signal obtained by adding the copy interference signal to the PCM audio signal is, for example, a C signal.
By recording in D, if the purchaser of the CD reproduces the CD, the CD is not compressed, and the copy protection signal has a level that is not recognized in terms of audibility. The reproduction sound of the PCM audio signal is normally reproduced without being recognized above.

On the other hand, when the purchaser or the like
When the CM signal is subjected to compression processing and copied and recorded on an MD or IC memory, and a signal copied and recorded from the MD or IC memory is reproduced, as shown in FIG. For example, since the PCM signal is reproduced in a state where the bit allocation (quantization bit length) of the middle band is smaller than originally required and the bit allocation is also performed to the originally unnecessary band, the listener is In addition to the reproduced sound, noise is heard, and a sound that cannot be heard is obtained. For this reason, copying can be indirectly prevented.

When the above embodiment is applied to an output circuit of a CD player or the like, the copy obstruction operation according to the present embodiment can be realized even for an existing CD on which no copy obstruction signal is recorded. .

Next, the recording apparatus according to the present invention will be described. FIG. 4 shows a block diagram of an embodiment of the recording apparatus according to the present invention. In the figure, an input PCM audio signal is supplied to a time-frequency conversion circuit 11 composed of an input buffer, where it is grouped by unit time length and developed in frequency. The masking amount calculation circuit 12 is based on the frequency spectrum obtained by the time-frequency conversion circuit 11, and is divided into sub-bands by the above-described critical bandwidth.
First, the masking amount for each band determined from the relative audible level is calculated. The amount of masking is convolved over the entire band to determine the total masking level in each subband. Further, the final masking level is determined by considering the absolute audible level for each band and selecting the higher masked signal level.

The copy disturbing signal generating circuit 13 generates a copy disturbing signal by multiplying the masking level calculated by the masking amount calculating circuit 12 with a level adjustment signal α set for each band as a coefficient. The level adjustment signal α has a numerical value of 1 or less. A small value is selected when emphasizing sound quality, and a large value is selected when giving priority to the copy disturbance effect. Below the audible frequency such as 0 Hz or 20 kHz or near the audible frequency upper limit frequency, the level adjustment signal α may be reduced to avoid malfunction of the music signal reproducing device. In this way, a copy disturb signal equal to or lower than the masking level is generated.

The adder circuit 14 receives the copy interference signal from the copy interference signal generation circuit 13 and the time-frequency conversion circuit 1
The signal of the original frequency spectrum output from 1 is added on the frequency spectrum. Then, the frequency-time conversion circuit 15 expands the addition signal output from the addition circuit 14 from the frequency domain to the time domain, and generates a PCM signal having a copy protection effect in the time domain. The PCM signal thus obtained is recorded on a CD or other predetermined recording medium (not shown) through well-known recording means (not shown).

The PCM signal thus recorded is
Even if a copy disturbing signal is added, when reproduced as it is, the copy disturbing signal is not detected, and the quality equivalent to the original signal can be maintained. However, when the PCM signal thus recorded is copied and recorded on another recording medium, a compression encoding process is added at the time of recording, and the compression encoding method is one of the MD, MPEG1, and DCC. In an encoding method that compresses at a compression ratio of / 4 or less, the amount of information is consumed by the added copy interference signal, and the amount of quantization bits becomes insufficient as a whole, resulting in sound quality deterioration.

Thus, according to the present embodiment, it is possible to effectively prevent easy copying of a compressed copy of an audio signal due to quality deterioration occurring at the time of compression encoding, thereby protecting the rights of a content creator. It is possible.

Note that the present invention is not limited to the above embodiment, and the present invention is applied to, for example, a PCM audio signal distributed through the Internet or the like or a PCM audio signal transmitted by broadcasting or the like. You can also. Further, it goes without saying that a recording medium or a transmission medium recorded according to the present invention is also included in the present invention.

[0045]

As described above, according to the present invention,
By inserting a copy disturbing signal at the very end of the masking level into a band that is not required for the signal in the frequency domain and converting it into a PCM audio signal in the time domain, the compression processing of the converted PCM audio signal at the time of copy recording can be performed originally. Bits are allocated to bands that are not required, and the amount of bits allocated to bands that are originally required for auditory importance is reduced so as not to provide a sufficient amount of information. The consumption and the shortage of the quantization bit amount on the whole can cause sound quality degradation, thereby indirectly preventing copy recording.

Further, according to the present invention, since the copy disturbing signal is added over the entire band, the sound quality deterioration can be reliably perceived irrespective of the difference between the listener and the sound source, and the copy preventing effect can be obtained. It is possible to obtain a sufficient amount, and it is possible to indirectly prevent copy recording of the first generation.

[Brief description of the drawings]

FIG. 1 is a flowchart for explaining the operation of an embodiment of the method of the present invention.

FIG. 2 is a comparison diagram of an example of a masking level in a sound source to which a copy interference signal is added according to the present invention and a quantization noise level generated by a given bit allocation;

FIG. 3 is a diagram illustrating an example of bit allocation by a sound source to which a copy interference signal is added according to the present invention;

FIG. 4 is a block diagram of an embodiment of the device of the present invention.

FIG. 5 is a flowchart illustrating an example of general bit allocation.

FIG. 6 is a comparison diagram of an example of a masking level in a normal sound source and a quantization noise level generated by a given bit allocation.

FIG. 7 is a diagram showing an example of bit allocation in a normal sound source.

[Explanation of symbols]

 Reference Signs List 11 time-frequency conversion circuit 12 masking amount calculation circuit 13 copy interference signal generation circuit 14 addition circuit 15 frequency-time conversion circuit

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5D044 AB05 DE03 DE13 DE17 DE50 DE52 EF05 GK08 HL08 5D045 DA20

Claims (7)

[Claims] A first step of converting a pulse code modulated PCM audio signal into a frequency domain signal for each unit time; and converting the frequency domain signal obtained in the first step into a human frequency resolution. A second step of dividing the subband into subbands each having a critical bandwidth corresponding to and calculating a masking level for this frequency domain signal for each subband; performing a convolution operation on the masking level over the entire band; A third step of determining a final masking level in the sub-band, calculating a noise signal lower than the masking level as a copy interference signal for each sub-band, and generating the copy interference signal. A fourth step of adding to the frequency domain signal obtained in the step, and a fourth step obtained in the fourth step. PCM of the issue time-domain
A method of recording an audio signal, comprising: a fifth step of converting the signal into an audio signal; and a sixth step of recording the PCM audio signal obtained in the fifth step on a recording medium. 2. A first step of converting a pulse code modulated PCM audio signal into a frequency domain signal for each unit time, and converting the frequency domain signal obtained by the first step into a human frequency resolution. A second step of dividing the subband into subbands each having a critical bandwidth corresponding to and calculating a masking level for this frequency domain signal for each subband; performing a convolution operation on the masking level over the entire band; A third step of determining a final masking level in the sub-band, calculating a noise signal lower than the masking level as a copy interference signal for each sub-band, and generating the copy interference signal. A fourth step of adding to the frequency domain signal obtained in the step, and a fourth step obtained in the fourth step. PCM of the issue time-domain
A method for transmitting an audio signal, comprising: a fifth step of converting the signal into an audio signal; and a sixth step of transmitting the PCM audio signal obtained in the fifth step. 3. A time-frequency conversion circuit for converting a pulse code modulated PCM audio signal into a frequency domain signal for each unit time, and converting a frequency domain signal obtained by the time-frequency conversion circuit into a human signal. Divided into subbands with a critical bandwidth corresponding to the frequency resolution as a unit, calculate a masking level for this frequency domain signal for each subband, perform a convolution operation on the masking level over the entire band,
A masking amount calculating circuit for determining a final masking level in each of the subbands; and a noise signal having a masking level equal to or less than the masking level determined by the masking amount calculating circuit is calculated and generated as a copy interference signal for each subband. A copy interference signal generation circuit, an addition circuit for adding the copy interference signal to a frequency domain signal output from the time-frequency conversion circuit and outputting the added signal, and a time domain PCM audio signal output from the addition circuit. A recording apparatus comprising: a frequency-time conversion circuit for converting a signal into a signal; and recording means for recording the PCM audio signal output from the frequency-time conversion circuit on a recording medium. 4. A time-frequency conversion circuit that converts a pulse code modulated PCM audio signal into a frequency domain signal for each unit time, and converts a frequency domain signal obtained by the time-frequency conversion circuit into a human signal. Divided into subbands with a critical bandwidth corresponding to the frequency resolution as a unit, calculate a masking level for this frequency domain signal for each subband, perform a convolution operation on the masking level over the entire band,
A masking amount calculating circuit for determining a final masking level in each of the subbands; and a noise signal having a masking level equal to or less than the masking level determined by the masking amount calculating circuit is calculated and generated as a copy interference signal for each subband. A copy interference signal generation circuit, an addition circuit for adding the copy interference signal to a frequency domain signal output from the time-frequency conversion circuit and outputting the added signal, and a time domain PCM audio signal output from the addition circuit. A transmission apparatus comprising: a frequency-time conversion circuit that converts a signal into a signal; and an output unit that transmits the PCM audio signal output from the frequency-time conversion circuit. 5. A pulse code modulated PCM audio signal is subjected to time-frequency conversion and then divided into sub-bands each having a critical bandwidth corresponding to human frequency resolution, and a masking level is set for each of these sub-bands. Calculate, perform a convolution operation on those masking levels over the entire band, determine the final masking level in each of the subbands, and add a noise signal below the masking level calculated for each subband as a copy interference signal. A recording medium characterized by recording a PCM audio signal. 6. A pulse code modulated PCM audio signal is subjected to time-frequency conversion and then divided into sub-bands each having a critical bandwidth corresponding to human frequency resolution, and a masking level is set for each of these sub-bands. Calculate, perform a convolution operation on those masking levels over the entire band, determine the final masking level in each of the subbands, and add a noise signal below the masking level calculated for each subband as a copy interference signal. A transmission medium for transmitting a PCM audio signal. 7. The third step is to use a masking level obtained from an absolute audible level and a relative audible level based on a human auditory characteristic as an index, and to apply a level adjustment signal given to each of the sub-bands to the masking. The method according to claim 1, wherein the copy interference signal is generated by multiplying a level.