JPH08125544A - Method/device for compressing digital signal and recording medium - Google Patents

Abstract

PURPOSE: To prevent the efficiency deterioration of block floating and quantization by controlling the frequency size of a small block executing floating in accordance with a characteristic on the frequency axis of an input signal and executing optimum floating. CONSTITUTION: Spectrum data supplied to an input terminal 301 is transmitted to a change quantity calculation circuit 303. Change quantity for respective frequencies is calculated and it is transmitted to an integrating comparsion circuit 304. The circuit 304 integrates change quanitty data and compares it with a threshold. Then, the frequency when an integrating value exceeds the threshold is transmitted to a block floating unit decision circuit 305. A unit correction circuit 306 corrects the boundary of the block floating unit from the circuit 305 from output from an energy calculation circuit 307 and a standard unit output circuit 309 and outputs it to an adaptive bit encoding circuit. The circuit 306 evaluates the quantity of auxiliary data in the unit after correction by the output of the circuit 307, the efficiency of block floating and information quantity required at the time of quantization so as to decide the adoption of the unit.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to recording / reproducing of compressed data obtained by bit-compressing a digital audio signal or the like, a recording medium on which the compressed data is recorded, and a transmission system of the compressed data, and more particularly to a frequency of an input signal. A digital signal compression method in which the frequency size of a small block subdivided by time and frequency for floating and / or bit allocation for information compression is changed according to the change on the axis. And an apparatus and a recording medium.

[0002]

2. Description of the Related Art The applicant of the present application has previously proposed a technique for bit-compressing an input digital audio signal and recording it in bursts with a predetermined data amount as a recording unit, for example, Japanese Patent Application No. Hei 2-221364. , Japanese Patent Application No. 2-22136
No. 5, Japanese Patent Application No. 2-222821, Japanese Patent Application No. 2-2228
No. 23 specification, drawings, etc.

This technique uses a magneto-optical disk as a recording medium, and records and reproduces AD (adaptive difference) PCM audio data defined in audio data formats of so-called CD-I (CD-interactive) and CD-ROM XA. For example, 32 sectors of this ADPCM data and several sectors for linking for interleave processing are recorded as a recording unit on the magneto-optical disk in a burst manner.

Several modes can be selected for the ADPCM audio in the recording / reproducing apparatus using this magneto-optical disk. For example, the sampling frequency is doubled as compared with the reproduction time of a normal CD. Is 37.
A level A of 8 kHz, a level B of a sampling frequency of 37.8 kHz at a compression rate of 4 times, and a level C of a sampling frequency of 18.9 kHz at a compression rate of 8 times are specified. That is, for example, in the case of the level B, the digital audio data is compressed to approximately 1/4, and the reproduction time (play time) of the disc recorded in the level B mode is the standard CD format (CD -D
It is four times that of A format). This means that the recording / reproducing time of the standard 12 cm can be obtained with a smaller disc, and the device can be miniaturized.

However, the rotation speed of the disk is standard C
Since it is the same as D, for example, in the case of the above level B, compressed data for a reproduction time four times as long as the predetermined time can be obtained. Therefore, for example, the same compressed data is read four times in a time unit such as a sector or a cluster, and only the compressed data for one time is sent to the audio reproduction. Specifically, when scanning (tracking) a spiral recording track,
A reproduction operation is performed in such a form that a track jump is performed to return to the original track position for each rotation, and the same track is repeatedly tracked four times. This means that normal compressed data only needs to be obtained at least once out of, for example, four times of redundant reading, is resistant to errors due to disturbances, etc., and is particularly preferable when applied to portable small devices.

Furthermore, the applicant of the present invention has proposed a bit allocation method for realizing good compression efficiently and in Japanese Patent Application No. 4-369.
No. 52 specification and drawings. In this technique, when allocating bits, normalization is performed according to the representative value in each small block such as the so-called critical band, so-called block floating, and bit allocation depending on the signal size in each small block is performed.
The weighting is performed according to the band corresponding to the small block. According to this technique, good compression can be performed when there is no extreme variation in the size of the spectrum in each small block.

[0007]

However, when the above-mentioned technique is applied to compress digital data,
Signals containing extreme fluctuations in the size of the spectrum in each small block for performing block floating and special peak components, for example, when compressing a single signal or multiple sine waves or rectangular waves, Since the deviation of the data in the small blocks is large, the efficiency may decrease when the block floating is performed using the representative value in each small block. Furthermore, when the state of the input signal, that is, the variation of the spectrum size in each small block or the peak component that is specially changed, the efficiency of the block floating described above also changes, and the processing block in the time axis direction changes A large deviation in compression efficiency may occur. In addition, if bit allocation for compression is performed in a state where the block floating efficiency is reduced, redundant bit allocation may be necessary in order to keep the generated quantization noise below the allowable noise level. As a result, compression efficiency may be reduced or a large deviation may occur.

In this case, in most of the application examples using the above-mentioned technique, the upper limit of the usable bit rate is defined for the convenience of the recording medium and the transmission path. Therefore, the upper limit of the bit rate is set to the compression efficiency. If the processing block has a low compression ratio, the processing block having a high compression efficiency has an excessive number of bits, resulting in a reduction in the overall compression efficiency. In addition, if the above upper limit is matched with a processing block with high compression efficiency, the amount of information in the processing block with low compression efficiency may be insufficient, and the auditory problem may not be ignored. This problem becomes more serious as the usable bit rate becomes lower, and it becomes more difficult to realize a lower bit rate, in other words, a higher compression rate, as the deviation of the compression efficiency of the processing block due to the input signal increases. Become.

The present invention has been made in view of the above circumstances, and it is possible to compress the information without decreasing the efficiency of the block floating and the quantization regardless of the characteristics of the input signal and the bit allocation for the information compression. It is an object of the present invention to provide a digital signal compression method and apparatus to which the above method is applied, and a recording medium.

[0010]

DISCLOSURE OF THE INVENTION A digital signal compression method and apparatus according to the present invention have been proposed in order to achieve the above-mentioned object, and calculate the characteristic of the input signal on the frequency axis to obtain the input signal. It is divided into small blocks that are subdivided with respect to time and frequency, and is subjected to floating for compression, and the frequency size of the small blocks to which the floating is applied is controlled in accordance with the characteristics of the input signal on the frequency axis to optimize the operation. It is characterized by performing floating.

Further, the digital signal compression method and apparatus of the present invention obtains spectrum data on the frequency axis of an input signal, obtains an acceptable noise spectrum from the spectrum data, and obtains the obtained acceptable noise spectrum with time. And divided into small blocks subdivided in frequency,
Bit allocation for compression is performed, and bit allocation is optimized by changing the frequency size of a small block for bit allocation for compression according to the characteristics of the input signal on the frequency axis. .

Further, the digital signal compression method and apparatus of the present invention comprises a step or means for calculating the characteristics of the input signal on the frequency axis, and the compression of the input signal by dividing it into small blocks subdivided in time and frequency. For obtaining the floating noise for obtaining the spectrum data on the frequency axis, obtaining the acceptable noise spectrum from the spectrum data, and obtaining the acceptable noise spectrum with respect to time and frequency. Steps or means for allocating bits for compression and allocating bits for compression, and small blocks for time and frequency to perform floating for compression in accordance with the characteristics of the input signal on the frequency axis And / or frequency size of a small block subdivided with respect to time and frequency for bit allocation for compression A step or means for changing the block floating and bit allocation are optimized by changing the small block for the block floating and the small block for the bit allocation together or independently. To do.

In the digital signal compression method and apparatus of the present invention, the time and frequency for performing block floating for compression are divided into small blocks and / or the time and frequency for allocating bits for compression. Predetermining the frequency size of the subdivided small blocks, and the compression efficiency when adopting this predetermined size,
Compare with the compression efficiency when the frequency size of the small block that performs floating for compression and / or the small block that performs bit allocation for compression is changed according to the characteristics of the input signal on the frequency axis. Then, the size of the small block that enables information compression with higher efficiency is selected according to the comparison result. Further, the digital signal compression method and apparatus of the present invention include a small block subdivided in time and frequency for performing block floating for compression, and a small block subdivided in time and frequency for bit allocation for compression. Are commonly or independently configured according to the characteristics of the input signal on the frequency axis. Further, the method and apparatus of the present invention make the time length of the processing block for information compression variable in accordance with the input signal, and the change of the input signal of the processing block and the change of the input signal of the other processing block. And / or determining the temporal length of the processing block based on power, or energy or peak information, and / or the change in the input signal of the processing block and the time longer than the maximum of the processing block in time. The time length of the processing block is determined based on the change information of the input signal obtained from the width input signal. When determining the processing block length, the ratio that is involved in the determination of the element that determines the temporal length of the processing block is fixed or adapted to the input signal, and / or a predetermined ratio (for example, double, Four
2 times, 8 times, etc.) and use together or alone. In addition,
The input signal is an audio signal, and the frequency width of the block that controls the generation of at least most of the quantization noise is made wider in the higher frequencies.

Further, in the method and apparatus of the present invention, the orthogonal transform is used to divide the time axis signal into a plurality of bands on the frequency axis, and the orthogonal transform size is changed in the orthogonal transform, and is used at the time of the orthogonal transform. The shape of the window function is changed. Here, when dividing the time axis signal into a plurality of bands on the frequency axis, first, the time axis signal is divided into a plurality of bands, and a block composed of a plurality of samples is divided for each of the divided bands. Then, coefficient data is obtained by performing orthogonal transformation for each block in each band. In addition, the division frequency width in the division from the time axis signal before orthogonal transformation into a plurality of bands on the frequency axis is made wider in the higher region, and the division frequency width is made the same in two consecutive lowest bands. Furthermore, when allocating the bits, the compression code is not allocated to the main information and / or the sub-information for the signal components in the band substantially equal to or more than the signal pass band, and the quadrature is used for the division into the plurality of bands. -Use a mirror filter. Also,
A modified discrete cosine transform is used as the orthogonal transform, and when determining the temporal length of the processing block based on the change of the input signal of the processing block, a predetermined boundary value for determining the temporal length is used. Is variable according to the amplitude and frequency of the input signal. The boundary value takes a plurality of stepwise values according to the amplitude and frequency of the input signal. Further, when the processing block length is determined, the auditory characteristics that the signal of the other processing block exerts on the signal of the processing block is determined by determining the spectrum of the frequency axis and / or the energy of the orthogonal transform coefficient and / or Calculation is performed using power or peak information, and the temporal length of the processing block is determined. Allocation of bits for compression and / or block floating of spectrum and / or orthogonal transform coefficients on the frequency axis used when calculating the auditory characteristics of the signal of the other processing block on the signal of the processing block. It is also used as the spectrum and / or the orthogonal transform coefficient on the time axis after the orthogonal transform used for. Furthermore, when determining the temporal length of the processing block based on the change of the input signal of the processing block, the judgment based on the periodic change of the input signal and / or the repeating pulse or the periodic characteristic is performed. To do.

In the method and apparatus of the present invention, when determining the temporal length of the processing block based on the change of the input signal of the processing block, the boundary value is used as the amplitude of the input signal,
The function of making it variable according to the frequency and the auditory characteristic of the signal of the other processing block exerted on the signal of the processing block, the energy and / or power or peak of the spectrum on the frequency axis and / or the orthogonal transform coefficient. It is also possible to have a function of performing calculation using information and determining the temporal length of the processing block.

Next, the recording medium of the present invention has the above-mentioned digital signal compression method of the present invention or the compressed data compressed by the digital signal compression method of the present invention recorded therein.

The digital signal compression method and apparatus of the present invention also transmits compressed compressed data.

That is, the digital signal compression method and apparatus (high efficiency coding method and apparatus) according to the present invention calculates the characteristic of the input signal on the frequency axis and divides the input signal into time and frequency small blocks. The above problem is solved by performing the floating for compression, controlling the frequency size of the small blocks to be floated according to the characteristics of the input signal on the frequency axis, and performing the optimum floating. Solve.

Further, when bit allocation for compression is performed, a small block is subdivided in time and frequency for bit allocation for compression according to the characteristics of the input signal on the frequency axis. The above problem is solved by changing the frequency magnitude.

Further, it is more effective to change the small block for subdividing the time and frequency for performing the above-mentioned block floating and the small block for bit allocation either together or independently.

On the other hand, a small block subdivided into time and frequency for block floating and / or bit allocation is determined in advance with the frequency magnitude obtained according to the characteristics of the input signal on the frequency axis. It is more effective to compare the frequency sizes of the small blocks and select a size having a high compression efficiency comprehensively.

[0022]

According to the digital signal compression method and apparatus of the present invention, the block floating and / or the block floating and / or the bit floating efficiency for the input signal having the large deviation of the efficiency of the block allocation and / or the bit allocation are suppressed. Alternatively, compression with a small deviation in efficiency can be realized by selecting the frequency size of a small block subdivided with respect to time and frequency for bit allocation. As a result, it is possible to prevent a decrease in compression efficiency, obtain better sound quality at the same bit rate, and record or transmit at a lower bit rate at the same sound quality. Can be realized.

[0023]

1 is a block circuit diagram showing a schematic configuration of an embodiment of a digital signal compression apparatus (compressed data recording / reproducing apparatus) of the present invention to which the digital signal compression method of the present invention is applied.

In the compressed data recording / reproducing apparatus shown in FIG. 1, the magneto-optical disk 1 rotated by a spindle motor 51 is used as a recording medium. When recording data on the magneto-optical disk 1, so-called magnetic field modulation recording is performed by applying a modulation magnetic field according to the recording data with the magnetic head 54 while irradiating the optical head 53 with laser light, for example. Data is recorded along the recording track of. During reproduction, the recording track of the magneto-optical disc 1 is traced with laser light by the optical head 53 to perform magneto-optical reproduction.

The optical head 53 includes, for example, a laser light source such as a laser diode, a collimator lens, an objective lens,
It is composed of a polarization beam splitter, an optical component such as a cylindrical lens, and a photodetector having a light receiving portion of a predetermined pattern. The optical head 53 is provided at a position facing the magnetic head 54 through the magneto-optical disk 1. When recording data on the magneto-optical disk 1, a magnetic head 54 is driven by a head driving circuit 66 of a recording system to be described later to apply a modulation magnetic field according to the recording data, and the optical head 53 is used for the purpose of the magneto-optical disk 1. By irradiating the track with laser light, thermomagnetic recording is performed by the magnetic field modulation method. The optical head 53 also detects the reflected light of the laser light applied to the target track, detects a focus error by, for example, a so-called astigmatism method, and detects a tracking error by, for example, a so-called push-pull method. When reproducing data from the magneto-optical disk 1, the optical head 53 detects the focus error and the tracking error, and at the same time, detects the difference in the polarization angle (Kerr rotation angle) of the reflected light of the laser light from the target track and reproduces it. Generate a signal.

The output of the optical head 53 is supplied to the RF circuit 55. The RF circuit 55 extracts the focus error signal and the tracking error signal from the output of the optical head 53 and supplies them to the servo control circuit 56, and binarizes the reproduction signal to reproduce the reproduction system decoder 7 described later.
Feed to 1.

The servo control circuit 56 is composed of, for example, a focus servo control circuit, a tracking servo control circuit, a spindle motor servo control circuit, a sled servo control circuit and the like. The focus servo control circuit controls the focus of the optical system of the optical head 53 so that the focus error signal becomes zero. Further, the tracking servo control circuit controls the tracking of the optical system of the optical head 53 so that the tracking error signal becomes zero. Further, the spindle motor servo control circuit controls the spindle motor 51 so as to rotate the magneto-optical disk 1 at a predetermined rotation speed (for example, a constant linear speed). Further, the sled servo control circuit moves the optical head 53 and the magnetic head 54 to the target track position of the magneto-optical disk 1 designated by the system controller 57. The servo control circuit 56 that performs such various control operations sends information indicating the operating state of each unit controlled by the servo control circuit 56 to the system controller 57.

A key input operation unit 58 and a display unit 59 are connected to the system controller 57. The system controller 57 controls the recording system and the reproducing system in the operation mode designated by the operation input information from the key input operation unit 58. The system controller 7 also records the above-mentioned recording traced by the optical head 53 and the magnetic head 54 on the basis of address information in sector units reproduced from the recording track of the magneto-optical disc 1 by the header time, sub-code Q data and the like. Manages the recording and playback positions on the track. Further, the system controller 57
Based on the data compression rate and the reproduction position information on the recording track, control is performed to display the reproduction time on the display unit 59.

The reproduction time display is the reciprocal of the data compression ratio (absolute time information) with respect to address information (absolute time information) in sector units reproduced from the recording track of the magneto-optical disk 1 by so-called header time or so-called sub-code Q data. For example, in the case of 1/4 compression, the actual time information is obtained by multiplying by 4), and this is displayed on the display unit 59. Even at the time of recording, for example, when absolute time information is previously recorded (pre-formatted) on a recording track of a magneto-optical disk or the like, the pre-formatted absolute time information is read to determine the data compression rate. It is also possible to display the current position at the actual recording time by multiplying by the reciprocal.

Next, in the recording system of this disc recording / reproducing apparatus, the analog audio input signal AIN from the input terminal 60 is passed through the low pass filter 61 to the A / D converter 6
2 and the A / D converter 62 quantizes the analog audio input signal AIN. A / D converter 62
The digital audio signal obtained from the ATC (Ad
aptive Transform Coding) The PCM encoder 63 is supplied. Further, the digital audio input signal DIN from the input terminal 67 is supplied to the ATC encoder 63 via the digital input interface circuit 68. A
The TC encoder 63 sends the input signal AIN to the A / D
Bit compression (data compression) processing is performed on the digital audio PCM data quantized by the converter 62 and having a predetermined transfer rate. Although the compression rate will be described as 4 times here, the present embodiment has a configuration that does not depend on this magnification and can be arbitrarily selected depending on the application example.

Next, in the memory 64, writing and reading of data are controlled by the system controller 57,
It is used as a buffer memory for temporarily storing ATC data supplied from the ATC encoder 63 and recording the ATC data on a disk as needed. That is, for example, the compressed audio data supplied from the ATC encoder 63 has a standard C data transfer rate.
Data transfer rate of D-DA format (75 sectors /
1/4 of the second), that is, 18.75 sectors / second, and this compressed data is continuously written in the memory 14. As for this compressed data (ATC data), it is sufficient to record one sector for every four sectors as described above. However, since recording every four sectors is practically impossible, the continuous sectors described below will be used. I try to keep a record. This recording bursts at a data transfer rate (75 sectors / second) same as that of the standard CD-DA format by using a cluster composed of a plurality of predetermined sectors (for example, 32 sectors + several sectors) as a recording unit through a pause period. Is done in a regular manner. That is, in the memory 14, ATC audio data continuously written at a low transfer rate of 18.75 (= 75/4) sectors / second corresponding to the bit compression rate is recorded as the recording data of 75 sectors / second. It is read in bursts at the transfer rate. The overall data transfer rate of the read and recorded data, including the recording pause period, is as low as 18.75 sectors / sec, but within the time of the recording operation performed in bursts. The instantaneous data transfer rate in the above is the standard 75 sectors / sec. Therefore, when the disc rotation speed is the same as the standard CD-DA format (constant linear velocity), the same recording density and storage pattern as those of the CD-DA format are recorded.

The ATC audio data, that is, the recorded data, which is burst-read from the memory 64 at the (instantaneous) transfer rate of 75 sectors / second is recorded by the encoder 65.
Is supplied to. Here, from the memory 64 to the encoder 65
In the data string supplied to the above, the unit to be continuously recorded in one recording is a cluster composed of a plurality of sectors (for example, 32 sectors) and several sectors for cluster connection arranged at the front and rear positions of the cluster. This cluster connection sector is set to be longer than the interleave length in the encoder 65 so that interleaved data will not affect the data of other clusters.

The encoder 65, regarding the recording data supplied from the memory 64 in bursts as described above,
Encoding processing for error correction (parity addition and interleave processing), EFM encoding processing, and the like are performed. The recording data encoded by the encoder 65 is supplied to the magnetic head drive circuit 66. The magnetic head drive circuit 66 is connected to the magnetic head 54,
The magnetic head 54 is driven so as to apply the modulation magnetic field according to the recording data to the magneto-optical disk 1.

The system controller 57 controls the memory 64 as described above, and
By this memory control, the recording position is controlled so that the recording data read out in burst from the memory 64 is continuously recorded on the recording track of the magneto-optical disk 2.
The recording position is controlled by controlling the recording position of the recording data which is burst-read from the memory 64 by the system controller 57 and outputting a control signal for designating the recording position on the recording track of the magneto-optical disk 1 to the servo control circuit. By feeding 56.

Next, the reproducing system of this magneto-optical disk recording / reproducing unit will be described. This reproducing system is for reproducing the record data continuously recorded on the recording track of the magneto-optical disk 1 by the above-mentioned recording system,
The decoder 71 is provided with a reproduction output obtained by tracing a recording track of the magneto-optical disk 1 with a laser beam by the optical head 53 and binarized and supplied by the RF circuit 55. At this time, not only the magneto-optical disk but also the read-only optical disk same as the compact disk (CD: COMPACT DISC) can be read.

The decoder 71 corresponds to the encoder 65 in the above-mentioned recording system, and performs the above-mentioned decoding processing for error correction and EFM decoding on the reproduction output binarized by the RF circuit 55. By performing processing such as processing, the audio data is reproduced at a transfer rate of 75 sectors / second, which is faster than the normal transfer rate. This decoder 7
The reproduction data obtained by 1 is supplied to the memory 72.

In the memory 72, writing and reading of data are controlled by the system controller 57, and reproduced data supplied from the decoder 71 at a transfer rate of 75 sectors / second is written in bursts at the transfer rate of 75 sectors / second. Be done. Further, in the memory 72, the reproduction data written in a burst at the transfer rate of 75 sectors / sec is a regular transfer rate of 75 sectors / sec of 18.75.
It is read continuously at sectors / second.

The system controller 57 writes the reproduced data in the memory 72 at a transfer rate of 75 sectors / second, and also writes the reproduced data from the memory 72 in the above 18.7.
Memory control is performed so that data is continuously read at a transfer rate of 5 sectors / second. Further, the system controller 57 performs the above-mentioned memory control on the memory 72 and continuously reproduces the above-mentioned reproduction data written in burst from the memory 72 from the recording track of the magneto-optical disc 1 by this memory control. Controls the playback position. This playback position is controlled by the system controller 5
By controlling the reproduction position of the reproduction data read out from the memory 72 in burst by 7 and supplying the servo control circuit 56 with a control signal designating the reproduction position on the recording track of the magneto-optical disk 1 or the optical disk 1. Done.

ATC audio data obtained as reproduction data continuously read from the memory 72 at a transfer rate of 18.75 sectors / second is supplied to the ATC decoder 73. The ATC decoder 73 reproduces 16-bit digital audio data by expanding the ATC data four times (bit expanding). This ATC
The digital audio data from the decoder 73 is D
It is supplied to the / A converter 74.

The D / A converter 74 is the ATC decoder 73.
The digital audio data supplied from the converter is converted into an analog signal to form an analog audio output signal AOUT. The analog audio signal AOUT obtained by the D / A converter 74 is output from the output terminal 76 via the low pass filter 75.

Next, the high efficiency compression coding applied to the information compression of the digital signal compression apparatus of the present invention will be described in detail. That is, refer to FIG. 2 and subsequent figures for a technique for highly efficient encoding of an input digital signal such as an audio PCM signal using band division encoding (SBC), adaptive transform encoding (ATC) and adaptive bit allocation techniques. While explaining.

In the concrete high-efficiency encoder shown in FIG. 2, in the steady state, the input digital signal is divided into a plurality of frequency bands, and the two adjacent lowest bands have the same bandwidth. In the high frequency band, the higher the frequency band, the wider the bandwidth is selected, and the orthogonal transformation is performed for each frequency band. Bits are adaptively allocated and coded for each critical bandwidth (critical band) and for each band in which the critical bandwidth is subdivided in consideration of block floating efficiency in the middle and high frequencies. Normally, this block is the quantization noise generation block. This critical band is a frequency band divided in consideration of human auditory characteristics, and is the band of a pure tone when the pure tone is masked by narrow band noise of the same strength near the frequency of that pure tone. That is. This critical band has a wider bandwidth as it goes higher, and the entire frequency band of 0 to 22 kHz is divided into, for example, 25 critical bands.

Furthermore, in the embodiment of the present invention, the block size (processing block length) is adaptively changed according to the input signal before the orthogonal transformation, and the floating block is subjected to the floating processing according to the input signal. The unit size is also changed, and the quantization is performed so that the usage efficiency of the compressed information is optimized.

That is, in FIG. 2, for example, when the sampling frequency is 44.1 kHz, 0 to 0 are input to the input terminal 200.
A 22 kHz audio PCM signal is supplied.
This input signal is divided into a band of 0 to 11 kHz and a band of 11 to 22 kHz by a band dividing filter 201 including a filter such as so-called QMF, and 0 to 1 is divided.
A signal in the 1 kHz band is 0 to 5.5 kHz by the band division filter 202 which is also a so-called QMF filter or the like.
It is divided into a Hz band and a 5.5 kHz to 11 kHz band. 11 kHz to 22 kHz from the band division filter 201
The Hz band signal is sent to the MDCT circuit 203, which is an example of an orthogonal transform circuit, and the 5.
Signals in the 5 kHz to 11 kHz band are sent to the MDCT circuit 204.
0 to 5.5k from the band division filter 202
The signals in the Hz range are sent to the MDCT circuit 205, where they are subjected to MDCT processing.

Here, a QMF filter as an example of a method of dividing the above-mentioned input digital signal into a plurality of frequency bands is described in, for example, the document "Digital Coding."
Of Speech in Subvans "(" Digital coding
of speech in subbands "RE Crochiere, Bell Syst.
Tech. J., Vol.55, No.8 1976). This QMF filter divides the band into two equal bandwidths, and is characterized in that so-called aliasing does not occur when the divided bands are combined later.

In addition, the document "Polyphase Quadrature Filters-New Band Division Coding Technology"("Po
lyphase Quadrature filters -A new subband coding t
echnique ", Joseph H. Rothweiler ICASSP 83, BOSTON)
Describes an equal bandwidth filter partitioning technique.
This polyphase quadrature filter is characterized in that when a signal is divided into a plurality of bands of equal bandwidth, it can be divided at one time.

Further, as the above-mentioned orthogonal transform, for example, the input audio signal is divided into blocks in a predetermined unit time (frame), and the fast Fourier transform (F
FT), Discrete Cosine Transform (DCT), Modified DCT Transform (MDCT), etc. are used to convert the time axis into the frequency axis. For this MDCT, refer to the document "Subbands Using Filter Bank Design Based on Time Domain Aliasing Cancellation".
Transform Coding Using Filte
r Bank Designs Based on Time Domain Aliasing Cance
llation, "JPPrincen ABBradley, Univ. of Surrey
Royal Melbourne Inst. Of Tech. ICASSP 1987).

Here, each MDCT circuit 203, 204,
FIG. 3 shows a specific example of the standard input signal for the blocks for each band supplied to 205. In the specific example of FIG. 3, the three filter output signals have a plurality of orthogonal transform block sizes independently for each band,
The time resolution can be switched depending on the time characteristic of the signal, the frequency distribution, and the like. If the signal is quasi-stationary in time, then the orthogonal transform block size is 11.6 ms,
That is, the long mode (Long M) shown in FIG.
ode) and the signal is non-stationary,
The orthogonal transform block size is further divided into two and four.
The short mode (Short Mod) of FIG.
e) as in the case where all are divided into 4 to make 2.9 ms, or in the middle mode A (Middle) of FIG. 3C.
Mode A), middle mode B (M) of FIG.
As in the idle mode B), a part of the time division has a time resolution of 5.8 ms and a part has a time resolution of 2.9 ms by four divisions to adapt to an actual complicated input signal. It is apparent that the division of the orthogonal transform block size is more effective if the division is made more complicated if the scale of the processing device permits. The block size is determined by the block size determination circuits 206 to 208 shown in FIG. 2, transmitted to the MDCT circuits 203 to 205, and output from the output terminals 216 to 218 as block size information of the corresponding block.

Next, FIG. 4 shows a block circuit diagram showing a schematic configuration of a specific example of the block size determination circuit. Here, the block determination circuit 206 of FIG. 2 will be described as an example. Of the outputs of the band-splitting filter 201 including the QMF of FIG. 2, the output of 11 kHz to 22 kHz is the input terminal 4 of FIG.
01 to the power calculation circuit 404. Furthermore, among the outputs of the band division filter 202 of FIG.
The output of 5 kHz to 11 kHz is output to the power calculation circuit 405 via the input terminal 402 of FIG. 4, and the output of 0 to 5.5 kHz is output to the power calculation circuit 40 of the input terminal 403 of FIG.
6 respectively. Further, the block size determination circuits 207 and 208 in FIG. 2 have the input terminals 401 to 40 in FIG.
The operation is the same, except that the signal input to 3 is different from that of the block size determination circuit 206. The input terminals 401 to 403 of the block size determination circuits 206 to 208 have a matrix configuration, that is, the input terminal 401 of the block size determination circuit 207 has 5.5 kHz to 11 of the band division filter 202 of FIG.
The output of kHz is connected, and the output of 0 to 5.5 kHz is connected to the input terminal 402. The same applies to the block size determination circuit 208.

In FIG. 4, each power calculation circuit 404-
406 obtains the power of each frequency band by integrating the input time waveform for a fixed time. At this time, the integration time width needs to be equal to or smaller than the minimum time block of the above orthogonal transform block sizes. Also,
Other than the above calculation method, the same effect can be obtained by using the absolute value of the maximum amplitude or the average value of the amplitudes within the minimum time width of the orthogonal transform block size as the representative power. The output of the power calculation circuit 404 is sent to the change extraction circuit 408 and the power comparison circuit 409, and the outputs of the power calculation circuits 405 and 406 are sent to the power comparison circuit 409, respectively. The change extraction circuit 408 obtains the differential coefficient of the power sent from the power calculation circuit 404 and sends it as power change information to the block size primary determination circuit 410 and the memory 407. In the memory 407, the power change information sent from the change extraction circuit 408 is accumulated for the maximum time of the above orthogonal transform block size or more. This is because the orthogonally adjacent orthogonal transform blocks influence each other by the window processing at the time of orthogonal transform, so that the power change information of the immediately preceding block temporally adjacent to each other is obtained in the block size primary determination circuit 410. This is because it is necessary. In the block size primary determination circuit 410, the power variation information of the corresponding block sent from the variation extraction circuit 408 and the memory 40.
Based on the power change information of the block immediately before the corresponding block that is temporally adjacent to the block, the orthogonal transform block size of the corresponding frequency band is determined from the temporal displacement of the power in the corresponding frequency band. To do. At this time, when a displacement equal to or more than a certain amount is recognized, the orthogonal transform block size is selected in a shorter time. However, even if the displacement point (boundary value) is fixed, the effect can be obtained. Furthermore, a value proportional to the frequency, that is, when the frequency is high, a large displacement causes a short block size in time, and when the frequency is low, a small displacement causes a small displacement in a block size to be temporally determined. , More effective.
It is desirable that this value (boundary value) change smoothly, but it may be a stepwise change in a plurality of steps. The block size determined as described above is transmitted to the block size correction circuit 411.

On the other hand, in the power comparison circuit 409, the power information of each frequency band sent from each of the power calculation circuits 404 to 406 is compared at the same time and the time width in which the masking effect occurs on the time axis to calculate the power. Circuit 404
The effect of another frequency band on the output frequency band of the above is calculated and transmitted to the block size correction circuit 411. In the block size correction circuit 411, the block size primary determination circuit 41 is based on the masking information sent from the power comparison circuit 409 and the past block size information sent from each tap of the delay group consisting of delays 412 to 414.
Correct the block size sent from 0 to select a block size longer in time, and delay 412
And to the window shape determination circuit 415. The function of the block size correction circuit 411 is that even if the pre-echo becomes a problem in the corresponding frequency band, if a signal having a large amplitude exists in another frequency band, particularly in a band lower than the corresponding frequency band, the masking effect causes Pre-echo is not a problem in hearing,
Alternatively, it utilizes the property that the problem may be alleviated. Note that the masking is a phenomenon in which one signal is masked by another signal due to human auditory characteristics so that the other signal cannot be heard. There are an effect and a simultaneous masking effect by a signal on the frequency axis. Due to these masking effects, even if there is noise in the masked portion, this noise cannot be heard. Therefore, in the actual audio signal, the noise within the masked range is regarded as an acceptable noise.

Next, in the delays 412 to 414, the past orthogonal transform block sizes are recorded in order, and are output to the block size determination circuit 411 from each tap, that is, the output of the delays 412 to 414. At the same time, the output of the delay 412 is output to the output terminal 417 and the delay 41
The outputs of 2, 413 are connected to the window shape determining circuit 415. The outputs from the delays 412 to 414 serve in the block size correction circuit 411 to make use of the change in the block size in a longer time width in determining the block size of the corresponding block. If so, increase the selection of the block size that is shorter in time, and if the block size that is shorter in time has not been selected in the past, you can increase the selection of the block size that is longer in time. I am trying. Note that this delay group may be used by increasing or decreasing the number of taps depending on the actual configuration and scale of the device, except for the delays 412 and 413 required for the window determination circuit 415 and the output terminal 417. In the window shape determination circuit 415, the output of the block size correction circuit 411, that is, the block size immediately after the block of interest and the delay 4
2 from the output of 12, that is, the block size of the corresponding block and the output of the delay 413, that is, the size of the immediately preceding block that is adjacent to the corresponding block in time.
The shape of the window used in the DCT circuits 203 to 205 is determined and output to the output terminal 416. The output terminal 417 of FIG. 4, that is, the block size information and the output terminal 416, that is, the window shape information, are connected to the respective units as outputs of the block size determination circuits 206 to 208 of FIG.

The window shape determined by the window shape determination circuit 415 will be described. FIG. 5 shows the shapes of adjacent blocks and windows. As can be seen from (a) to (c) of FIG. 6, the window used for orthogonal transform has a portion overlapping with a temporally adjacent block as shown by a dotted line and a solid line in the figure, and thus the present embodiment In the example, since the shape overlapping the centers of adjacent blocks is adopted, the shape of the window changes depending on the orthogonal transform size of the adjacent blocks.

FIG. 6 shows details of the window shape. In FIG. 6, window functions f (n) and g (n +
N) is given as a function that satisfies the following equation (1).

F (n) × f (L-1-n) = g (n) × g (L-1-n) f (n) × f (n) + g (n) × g (n) = 1 (1) 0 ≦ n ≦ L−1.

If the adjacent conversion block lengths are the same, L in this equation (1) becomes the conversion block length as it is, but if the adjacent conversion block lengths are different, the shorter conversion block length is set to L. , And a longer transform block length is K, in a region where windows do not overlap, the following formula (2) is given.

F (n) = g (n) = 1 K ≦ n ≦ 3K / 2−L / 2 f (n) = g (n) = 0 3K / 2 + L ≦ n ≦ 2K (2) By making the overlapping portions of the windows as long as possible, the frequency resolution of the spectrum at the time of orthogonal transformation is made good. As is clear from the above explanation,
The shape of the window used for orthogonal transformation is determined after the orthogonal transformation sizes for three blocks that are temporally consecutive are determined. Therefore, the block of signals input from the input terminals 401 to 403 and the block of signals output from the output terminals 416 and 417 in FIG. 4 are different by one block in this embodiment.

Further, the power calculation circuits 405 and 40 shown in FIG.
6 and the power comparison circuit 409 can be omitted, the block size determination circuits 206 to 208 in FIG. 2 can be configured. Further, by fixing the window shape to the smallest block size in time that can be taken by the orthogonal transform block, the type is made one, and the delay 412 of FIG.
˜414, the block size correction circuit 411, and the window shape determination circuit 415 can be omitted. In particular, in the application example in which the processing time delay is not desired, the above-mentioned omission makes the configuration small in delay and works effectively.

Referring again to FIG. 2, each MDCT circuit 203
˜205, the spectrum data or MDCT coefficient data on the frequency axis obtained by the MDCT processing is transmitted to the adaptive bit allocation coding circuits 210 to 212, the block floating unit determination circuits 219 to 221, and the bit allocation calculation circuit 209. ing. The block floating unit determination circuits 219 to 221 determine a block floating unit, which is a unit for performing a floating process according to the degree of concentration of energy or power, from the previous spectrum data or MDCT coefficient data,
The data is transmitted to the adaptive bit allocation encoding circuits 210 to 212, and is output from the output terminals 216 to 218 as auxiliary information for decoding together with the block size information and the window shape information described above.

FIG. 7 is a block circuit diagram showing a schematic configuration of a specific example of the block floating unit decision circuits 219 to 221 shown in FIG. The operation of the block floating unit determination circuit will be described with reference to FIG. In FIG. 7, the input terminal 301 is supplied with spectrum data on the frequency axis or MDCT coefficient data from the MDCT circuits 203 to 205. This spectrum data or MDCT coefficient data is sent to the change amount calculation circuit 303. The change calculation circuit 303 calculates the change for each frequency by differentiating the intensity of the spectrum data or the MDCT coefficient data by frequency and obtaining the differential coefficient, and the integration comparison circuit 30.
4 is being transmitted. The integration / comparison circuit 304 integrates the variation data for each frequency in ascending order of frequency, and compares it with the threshold value output from the threshold value output circuit 308.
Furthermore, the frequency at the time when the integrated value exceeds the threshold is transmitted to the block floating unit determination circuit 305 as the frequency of the boundary of the unit, and the integrated data is set to zero, and the integration is started again, and all frequencies are The above-mentioned operation is performed for the change amount data. The threshold value output circuit 308 has a function of outputting the threshold value so that the amount of change in the block floating unit becomes equal to or less than a certain value. Good results have been obtained by allocating slopes so that Better results can be obtained by making this threshold variable depending on the application example, input signal, etc., and depending on the application example, a constant can be used to simplify the configuration.

Next, in the block floating unit determination circuit 305, the frequency of the unit boundary transmitted from the integration comparison circuit 304 and the output of the bit allocation calculation circuit 209 of FIG. The block floating unit is determined based on the bit allocation, which is transmitted to the unit correction circuit 306 and the energy calculation circuit 307. The boundary of the block floating unit at this time is basically determined by the frequency of the boundary of the unit transmitted from the integration / comparison circuit 304 described above, but the output of the bit allocation becomes zero in both adjacent units. If so, the advantage of separating the block floating units disappears, and therefore, the block floating units are modified to be the same block floating unit. Further, the unit correction circuit 306 corrects the boundary of the block floating unit transmitted from the block floating unit determination circuit 305 from the output of the energy calculation circuit 307 and the output of the standard unit output circuit 307, and outputs it from the output terminal 310, Adaptive bit allocation encoding circuits 210 to 212 in FIG.
And output terminals 216 to 218. The energy calculation circuit 307 calculates the energy for each unit by integrating the spectrum data or MDCT coefficient data in the unit transmitted from the block floating unit determination circuit 305 for each unit,
It is transmitted to the unit correction circuit 306. In the unit correction circuit 306, when the transmitted energy of each unit is equal to or less than a certain value, the corresponding unit is not independent but is made a unit common to the adjacent unit. This works so that the data having high intensity in a limited portion having a narrow width in frequency constitutes an independent unit and the data having small intensity in a narrow portion does not constitute an independent unit.

Further, in the standard unit output circuit 309, in the configuration and application of the standard block floating unit, the bandwidths of the two adjacent lowest bands are the same, and in the higher frequency band, the higher the frequency band, the wider the bandwidth. Widely selected, orthogonal transformation is performed for each frequency band,
The obtained spectrum data on the frequency axis is subdivided into the critical bandwidths in the low frequency range, taking into account the human auditory characteristics described above, in so-called critical bandwidths (critical bands), and in the high and mid frequency range, considering the block floating efficiency. It outputs the units divided for each band. In the application example, this standard block floating unit has a common band for quantization, and has the advantage of requiring less data (auxiliary data) other than the compressed spectrum data or MDCT coefficient data portion. have. On the other hand, if attention is paid only to the efficiency of the block floating, the case where all the parts that greatly change are configured as independent units is the best, but in that case, the above auxiliary data increases. Therefore, the effective use of comprehensive information is determined in consideration of the amount of auxiliary data.
Therefore, the unit correction circuit 306 of FIG. 7 evaluates the amount of auxiliary data in the unit after the correction by the output of the energy calculation circuit 307, the efficiency of block floating, and the amount of information required for quantization, and the standard Only when the efficiency of the unit is higher than that of the unit, the unit previously obtained is used.

Here, the operation of the block floating unit will be described with reference to FIG. FIG. 8A shows part of the spectrum data or MDCT coefficient data input from the input terminal 301 of FIG. 7. Here, in the configuration of the standard block floating unit, in the frequency part, three block floating units having equal frequency widths, N-1, N, N + 1, as shown by the broken line range, are configured. I shall. In this state, in the N-1 unit, only the N-1 (2) data is larger than the other data, and the block floating coefficient is determined by this data. If is normalized with the block floating coefficient,
That is, the result is that some of the most significant digits of the normalized data are zero.

In FIG. 8B, the solid line graph represents the differential coefficient dP / df obtained by differentiating the spectral data or MDCT coefficient data of FIG. 8A by the frequency. Further, the bar graph in the figure shows Σ (dP / df), which is a value obtained by integrating the above-mentioned differential coefficient, dP / df. here,
If the output value of the threshold value output circuit 308 in FIG. 7 is the value shown by the alternate long and short dash line in the figure, N-1 (2) and N-1 (3)
Since the integrated value, Σ (dP / df), exceeds the threshold value, the boundary with the block floating unit is set here. FIG. 8C is a diagram showing a block floating unit constituted by the boundaries obtained as described above. As is clear from the figure, in (c) of FIG. 8, the frequency width of the first unit, N′-1, is narrowed, and the deviation of each data in the unit is reduced. Compared with, the efficiency of block floating is improved. In addition, as described above, since the change is integrated from low-frequency data to high-frequency data, the unit boundary often occurs immediately before a large data change, and the peak data frequency It becomes easy to control the block floating of the data in the vicinity of the lower one and the quantization. Therefore, the masking to the lower frequency sound is less effective than the higher frequency sound, which is in accordance with the auditory characteristic, and good compression can be realized.

Referring again to FIG. 2, the bit allocation calculation circuit 2
09 is based on the spectrum data divided considering the block floating transmitted by the critical band and block floating unit determination circuits 219 to 221 described above, and the critical band and the block floating are considered in consideration of so-called masking effect. Obtaining the masking amount for each divided band, based on the masking amount and energy or peak value for each divided band considering the critical band and block floating, obtain the number of allocated bits for each band,
It is transmitted to the adaptive bit allocation encoding circuits 210 to 212. The adaptive bit allocation coding circuits 210 to 212 quantize each spectrum data (or MDCT coefficient data) according to the number of bits allocated for each band. The data encoded in this way is output to the output terminal 2
It is taken out via 13-215.

Next, FIG. 9 shows the bit allocation calculation circuit 20.
9 is a block circuit diagram showing a schematic configuration of one specific example of FIG.
The operation of the bit allocation calculation circuit will be described with reference to FIG. In FIG. 9, the input terminal 701 has
The spectrum data on the frequency axis or the MDCT coefficient data from the MDCT circuits 203 to 205 are supplied. The input data on the frequency axis is sent to the energy calculation circuit 702 for each band, and the energy of each divided band considering the masking amount, the critical band, and the block floating is, for example, the amplitude value of each amplitude value in the band. It can be obtained by calculating the sum. Instead of the energy for each band, a peak value, an average value, etc. of the amplitude value may be used. This energy calculation circuit 7
As the output from 02, for example, the spectrum of the sum value of each band is shown as SB in FIG. However, in FIG. 10, in order to simplify the illustration, the number of divided bands in consideration of the masking amount, the critical band, and the block floating is expressed by 12 bands (B1 to B12).

Here, in order to consider the influence of the spectrum SB on so-called masking, a convolution process is performed such that the spectrum SB is multiplied by a predetermined weighting function and added. Therefore, the output of the energy calculation circuit 702 for each band, that is, each value of the spectrum SB is sent to the convolution filter circuit 703. The convolution filter circuit 703 includes, for example, a plurality of delay elements that sequentially delay input data, and a plurality of multipliers that multiply outputs from these delay elements by a filter coefficient (weighting function) (for example, 25 corresponding to each band). Number of multipliers) and a summation adder that sums the outputs of the respective multipliers. By this convolution processing, FIG.
The sum of the parts indicated by the middle dotted line is taken.

Here, a specific example of the multiplication coefficient (filter coefficient) of each multiplier of the convolution filter circuit 703 will be described. When the coefficient of the multiplier M corresponding to an arbitrary band is 1, the multiplier is M-1 gives a coefficient of 0.15, multiplier M-2 gives a coefficient of 0.0019, and multiplier M-3 gives a coefficient of 0.0000.
086, multiplier M + 1 gives a coefficient of 0.4, multiplier M + 2
By multiplying the output of each delay element by a coefficient of 0.06 with a coefficient of 0.007 by a multiplier M + 3, the convolution processing of the spectrum SB is performed. However, M is an arbitrary integer of 1 to 25.

Next, the output of the convolution filter circuit 703 is sent to the subtractor 704. The subtractor 704 obtains a level α corresponding to an allowable noise level described later in the convoluted area. It should be noted that the level α corresponding to the permissible noise level (permissible noise level) is a level at which the permissible noise level for each band of the critical band is obtained by performing inverse convolution processing, as described later. Is. here,
The subtractor 704 is supplied with an allowance function (function expressing a masking level) for obtaining the level α. The level α is controlled by increasing or decreasing this allowance function. The permissible function is supplied from the (n-ai) function generating circuit 705 described below.

That is, the level α corresponding to the allowable noise level can be obtained by the following equation (3), where i is the number given in order from the lower band of the critical band. α = S- (n-ai) (3) In this equation (3), n and a are constants, a> 0, S is the intensity of the convolved Bark spectrum, and (3)
In the formula, (n-ai) is the allowable function. In this embodiment, n = 38 and a = 1 are set, and there is no sound quality deterioration at this time, and good coding can be performed.

In this way, the level α is obtained, and this data is transmitted to the divider 706. The divider 706 is for inverse convolution of the level α in the convolved area. Therefore, the masking spectrum can be obtained from the level α by performing this inverse convolution processing. That is, this masking spectrum becomes the allowable noise spectrum. Although the above-mentioned inverse convolution processing requires complicated calculation, in this embodiment, the inverse convolution is performed using the simplified divider 706.

Next, the masking spectrum is transmitted to the subtractor 708 via the synthesis circuit 707. Here, the subtracter 708 outputs the output from the energy detection circuit 702 for each band, that is, the spectrum S described above.
B is supplied via the delay circuit 709. Therefore, the subtraction operation of the masking spectrum and the spectrum SB is performed by the subtractor 708.
As shown, the spectrum SB is masked below the level indicated by the level of the masking spectrum MS.

The output from the subtractor 708 is taken out through the allowable noise correction circuit 710 and the output terminal 711. For example, RO in which the assigned bit number information is stored in advance.
M, etc. (not shown). This ROM or the like is assigned to each band in accordance with the output (the level of the difference between the energy of each band and the output of the noise level setting means) obtained from the subtraction circuit 708 through the allowable noise correction circuit 710. Outputs bit number information. This allocation bit number information is sent to the adaptive bit allocation coding circuits 210 to 212 in FIG. 2, whereby the spectrum data on the frequency axis from the MDCT circuits 203 to 205 in FIG. 2 are allocated to each band. It is quantized by the number of bits.

That is, in summary, in the adaptive bit allocation encoding circuits 210 to 212 of FIG. 2, the energy of each divided band considering the masking amount, critical band and block floating, and the output of the noise level setting means. The spectrum data for each band is quantized by the number of bits assigned according to the level of difference. The delay circuit 709 shown in FIG.
It is provided to delay the spectrum SB from the energy detection circuit 702 in consideration of the delay amount in each circuit before 07.

By the way, at the time of synthesizing by the above-mentioned synthesizing circuit 707, data indicating a so-called minimum audible curve RC which is the human auditory characteristic as shown in FIG. The masking spectrum MS can be combined. In this minimum audible curve, if the absolute noise level is below this minimum audible curve, the noise will not be heard. This minimum audible curve is different even if the coding is the same, for example, due to the difference in playback volume at the time of playback, but in a realistic digital system, for example, how to enter music into the 16-bit dynamic range is Since there is not much difference, if, for example, the quantization noise in the most audible frequency band around 4 kHz is not heard, it is considered that the quantization noise below the level of this minimum audible curve is not heard in other frequency bands. . Therefore, assuming that the system is used in such a manner that noise near the word length of the system of 4 kHz is inaudible, and the minimum noise curve RC and the masking spectrum MS are combined together to obtain an allowable noise level, In this case, the allowable noise level can be up to the shaded portion in FIG. In this embodiment, the level of 4 kHz of the minimum audible curve is set to the minimum level equivalent to 20 bits, for example. Also, this figure 1
2 also shows the signal spectrum SS.

Further, in the allowable noise correction circuit 710,
The allowable noise level in the output from the subtractor 708 is corrected based on the information on the equal loudness curve sent from the correction information output circuit 713, for example. Here, the equal loudness curve is a characteristic curve relating to human auditory characteristics, for example, a curve obtained by obtaining the sound pressure of sound at each frequency that sounds the same as a pure tone of 1 kHz, and connecting the curves. Also called sensitivity curve. Further, this equal loudness curve is the minimum audible curve R shown in FIG.
It draws a curve substantially the same as C. In this equal loudness curve, for example, at 4 kHz, even if the sound pressure drops by 8 to 10 dB from 1 kHz, it sounds as loud as 1 kHz, and conversely, at around 50 Hz, it must be about 15 dB higher than the sound pressure at 1 kHz. It doesn't sound the same. Therefore, it is understood that the noise exceeding the level of the minimum audible curve (allowable noise level) should have the frequency characteristic given by the curve corresponding to the equal loudness curve. From this, it can be seen that correcting the permissible noise level in consideration of the equal loudness curve is suitable for human hearing characteristics.

Further, the correction information output circuit 713 detects and outputs the output information amount (data amount) at the time of quantization in the adaptive bit allocation encoding circuits 210 to 212 and the bit rate target value of the final encoded data. The allowable noise level is corrected on the basis of the information on the error between the two. This is because the total number of bits obtained by performing temporary adaptive bit allocation in advance for all bit allocation unit blocks is a fixed number of bits (target value) determined by the bit rate of the final encoded output data. May have an error with respect to, and bit allocation is performed again so that the error may be zero. That is, when the total allocated bit number is smaller than the target value, the difference bit number is allocated to each unit block and added, and when the total allocated bit number is larger than the target value, the difference bit number is added to each unit block. It works by allocating to and shaving.

In order to perform the above operation, an error in the total allocated bit number from the target value is detected, and the correction information output circuit 713 corrects each allocated bit number according to the error data. Output the data. Here, when the error data indicates a bit number shortage, it can be considered that the data amount is larger than the target value because a large number of bits are used per unit block. Further, when the error data is data indicating a surplus of the number of bits, it can be considered that the number of bits per unit block is small and the amount of data is smaller than the target value. Therefore,
From the correction information output circuit 713, in accordance with the error data, the data of the correction value for correcting the allowable noise level in the output from the subtractor 708 based on, for example, the information data of the equal loudness curve is provided. It will be output. By transmitting the correction value as described above to the allowable noise correction circuit 710, the subtractor 70
The allowable noise level from 8 is corrected. In the system as described above, the data obtained by processing the orthogonal transform output spectrum by the sub information as the main information, the scale factor indicating the block floating state as the sub information, and the word length indicating the word length are obtained.
Sent from encoder to decoder.

FIG. 13 shows the ATC decoder 73 shown in FIG. 1, that is, the reproduction signal obtained by reproducing the signal which is highly efficient coded and recorded on the recording medium of the present invention as described above, or is transmitted through the transmission medium. 2 shows a decoding circuit for decoding the received signal received again. The quantized MDCT coefficient of each band, that is, the output terminals 213 to 21 of FIG.
Data equivalent to the output signal of No. 5 is given to the decoding circuit input 107 and used block size information and information about the frequency length of the small blocks for block floating and quantization, that is, the output of FIG. Terminal 216-
Data equivalent to the output signal of 218 is provided to the input terminal 108. The adaptive bit allocation decoding circuit 106 cancels the bit allocation using the adaptive bit allocation information. Next, in the inverse orthogonal transform (IMDCT) circuits 103 to 105, the signal on the frequency axis is converted to the signal on the time axis. The signals on the time axis of these partial bands are the band synthesis filter (IQM
F) Decoded into full band signals by the circuits 102 and 101.

The present invention is not limited to the above-described embodiment, and for example, the recording / reproducing medium and the signal compression device or decompression device, and further, the signal compression device and the decompression device are integrated. It is not necessary to provide such a connection, and it is also possible to connect between them by a data transfer line, an optical cable, communication by light or radio waves, etc. without using a recording medium. Furthermore, for example, the present invention can be applied not only to audio PCM signals, but also to signal processing devices for digital audio (speech) signals, digital video signals, and the like.

Further, the recording medium of the present invention records the data compressed by the digital signal compressing device, so that the recording capacity can be effectively utilized. Further, as the recording medium of the present invention, not only the above-mentioned optical disk but also various recording media such as a magnetic disk, an IC memory and a card having the memory built therein, and a magnetic tape can be used.

[0082]

As is apparent from the above description, according to the digital signal compression method and apparatus of the present invention, an input signal which causes a large deviation in efficiency of block floating and / or bit allocation, in other words, The efficiency of a signal including an extreme variation in the size of data on the frequency axis in each small block subdivided with respect to time and frequency for block floating and / or bit allocation and a particular peak component By selecting the frequency size of the small block so as to suppress the deviation to a small value, it is possible to realize the distribution of bits with a small deviation in the compression efficiency. As a result, it is possible to prevent a decrease in compression efficiency, obtain better sound quality at the same bit rate, and record or transmit at a lower bit rate at the same sound quality. Can be realized. Therefore, it is possible to perform highly efficient compression and decompression with excellent sound quality in terms of hearing.

Further, the recording medium of the present invention for recording the data compressed by the digital signal compression apparatus of the present invention can effectively utilize the storage capacity as compared with the conventional recording medium. Further, the path for transmitting the data compressed by the digital signal compression apparatus of the present invention can improve the utilization efficiency of the line as compared with the conventional one.

[Brief description of drawings]

FIG. 1 is a block circuit diagram showing a configuration example of a compressed data recording / reproducing apparatus (disk recording / reproducing apparatus) as an embodiment of a digital signal compressing apparatus of the present invention.

FIG. 2 is a block circuit diagram showing a specific example of a high-efficiency compression encoding encoder that can be used for bit rate compression encoding according to the present embodiment.

FIG. 3 is a diagram showing a structure of an orthogonal transform block at the time of bit compression.

FIG. 4 is a block circuit diagram showing a configuration example of a circuit that determines an orthogonal transform block size.

FIG. 5 is a diagram showing a relationship between a change in temporal length of orthogonal transform blocks temporally adjacent to each other and a window shape used in orthogonal transform.

FIG. 6 is a diagram showing a detailed example of the shape of a window used in orthogonal transformation.

FIG. 7 is a block circuit diagram showing a configuration example of a circuit that determines a block floating unit.

FIG. 8 is a diagram showing a spectrum, differentiation, and integration data clearly showing the progress of the operation of the block floating decision circuit in the present embodiment.

FIG. 9 is a block circuit diagram showing an example of a bit allocation calculation circuit using a convolution operation that realizes a bit allocation operation function.

FIG. 10 is a diagram showing spectra of bands divided in consideration of each critical band and block floating.

FIG. 11 is a diagram showing a masking spectrum.

FIG. 12 is a diagram in which a minimum audible curve and a masking spectrum are combined.

FIG. 13 is a block circuit diagram showing a specific example of a high-efficiency compression encoding decoder that can be used for the bit rate compression encoding of the above embodiment.

[Explanation of symbols]

 201, 202 Band division filter 203-205 Orthogonal transformation circuit (MDCT) 206-208 Block determination circuit 209 Bit allocation calculation circuit 210-212 Adaptive bit allocation coding circuit 303 Change calculation circuit 304 Integration comparison circuit 305 Block floating unit determination circuit 306 Unit correction circuit 307 Energy calculation circuit 308 Threshold output circuit 309 Standard unit output circuit 404 to 406 Power calculation circuit 407 Memory 408 Change extraction circuit 409 Power comparison circuit 410 Block size primary determination circuit 411 Block size correction circuit 412 to 412 414 Delay 415 Window shape determination circuit

Claims (49)

[Claims] 1. A digital signal compression method for compressing information of a digital signal, wherein characteristics of an input signal on a frequency axis are calculated, and the input signal is distributed into small blocks subdivided with respect to time and frequency, and is floated for compression. And controlling the frequency size of the small blocks to be floated in accordance with the characteristics of the input signal on the frequency axis to perform optimum floating. 2. A digital signal compression method for compressing information of a digital signal, obtaining spectrum data on a frequency axis of an input signal, obtaining an acceptable noise spectrum from the spectrum data, and obtaining the obtained acceptable noise spectrum. It is divided into small blocks that are subdivided in terms of time and frequency, bits are allocated for compression, and the frequency size of small blocks that are allocated for compression is changed according to the characteristics of the input signal on the frequency axis. A digital signal compression method characterized by optimizing bit allocation. 3. A digital signal compression method for compressing information of a digital signal, the step of calculating the characteristic of the input signal on the frequency axis, and the step of distributing the input signal into small blocks subdivided in time and frequency for compression. Floating step, obtaining spectrum data on the frequency axis, obtaining an acceptable noise spectrum from the spectrum data, and dividing the acceptable noise spectrum into small blocks in terms of time and frequency. Distributing and allocating bits for compression, and small blocks and / or bits for compression that are subdivided in time and frequency for performing floating for compression according to the characteristics of the input signal on the frequency axis The step of changing the frequency size of the small blocks subdivided with respect to the time and frequency of allocation, A digital signal compression method characterized by optimizing block floating and bit allocation by changing the small block for block floating and the small block for bit allocation both or independently. 4. A frequency size of a small block subdivided in time and frequency for performing block floating for compression and / or a subdivided small block in time and frequency for bit allocation for compression is determined in advance. , A small block for performing floating for compression and / or a small block for allocating bits for compression according to the compression efficiency when adopting this predetermined size and the characteristics of the input signal on the frequency axis 2. The compression efficiency when the frequency size of is changed is compared, and the size of a small block that can perform information compression with higher efficiency is selected according to the comparison result. The digital signal compression method according to claim 3. 5. A small block subdivided in time and frequency for performing block floating for compression and a small block subdivided in time and frequency for bit allocation for compression are arranged on the frequency axis of an input signal. 5. The digital signal compression method according to claim 3 or 4, wherein the digital signal compression method is configured in common or independently according to the characteristics of. 6. The time length of a processing block for information compression is made variable by adapting to an input signal, and a change in the input signal of the processing block and a change in the input signal of another processing block, and / or power. , Or the time length of the processing block is determined based on energy or peak information.
A method for compressing a digital signal according to item. 7. A time length of a processing block for information compression is variable in accordance with an input signal, and a change in the input signal of the processing block and an input signal having a time width longer than the maximum of the processing block in terms of time. 6. The digital signal compression method according to claim 1, wherein the temporal length of the processing block is determined based on the change information of the input signal obtained by. 8. A time length of a processing block for information compression is made variable in accordance with an input signal, and a change of an input signal of the processing block and a change of an input signal of another processing block, and / or a power. Alternatively, the time length of the processing block is determined based on the energy or peak information, or the time length of the processing block is made variable by adapting to the input signal, and the input signal of the processing block is changed. The length of the processing block is determined based on change information and change information of the input signal obtained from an input signal having a time width longer than the maximum of the processing block in terms of time. 5. The digital signal compression method according to any one of 5 above. 9. When determining the processing block length, the ratio that is involved in the determination of the element that determines the temporal length of the processing block is fixed or applied to the input signal, and / or
9. The digital signal compression method according to claim 8, wherein the digital signal compression method is used in combination or alone at a predetermined ratio. 10. The input signal is an audio signal, and the frequency width of a block that controls the generation of at least most of the quantization noise is made wider in a higher frequency range. 10. The digital signal compression method according to any one of 1. 11. An orthogonal transform is used to divide a time-axis signal into a plurality of bands on the frequency axis, and the shape of a window function used at the time of the orthogonal transform is changed along with the change of the orthogonal transform size in the orthogonal transform. 11. The digital signal compression method according to claim 10, wherein 12. When dividing the time axis signal into a plurality of bands on the frequency axis, the time axis signal is first divided into a plurality of bands, and each divided band is made up of a plurality of samples. A coefficient is obtained by forming a block and performing orthogonal transformation for each block of each band.
1. The digital signal compression method described in 1. 13. The digital signal compression method according to claim 12, wherein a division frequency width in dividing the time-axis signal before orthogonal transformation into a plurality of bands on the frequency axis is set to be wider in a substantially higher range. 14. The digital signal compression method according to claim 13, wherein the divided frequency widths are the same in two consecutive bands of the lowest band. 15. The bit allocation according to claim 14, wherein the allocation of the compression code to the main information and / or the sub information for the signal component in the band substantially equal to or larger than the signal pass band is blocked. Digital signal compression method. 16. The digital signal compression method according to claim 12, wherein a quadrature mirror filter is used for the division into the plurality of bands. 17. The method according to claim 11, wherein a modified discrete cosine transform is used as the orthogonal transform.
7. The digital signal compression method according to any one of 6 above. 18. When determining a temporal length of a processing block based on a change of an input signal of the processing block, a predetermined boundary value for determining the temporal length is set as an amplitude of the input signal,
10. The digital signal compression method according to claim 6, wherein the digital signal compression method is variable according to the frequency. 19. The digital signal compression method according to claim 18, wherein the boundary value takes a plurality of stepwise values according to the amplitude and frequency of the input signal. 20. When determining the processing block length,
The auditory characteristic that the signal of the other processing block exerts on the signal of the processing block is calculated by using the energy and / or power or peak information of the spectrum on the frequency axis and / or the orthogonal transformation coefficient, and the processing concerned. 9. The digital signal compression method according to claim 6, wherein the time length of the block is determined. 21. Allocation of bits for compression of spectrum and / or orthogonal transform coefficients on the frequency axis used when calculating the auditory characteristics of the signal of the other processing block on the signal of the processing block. 21. The digital signal compression method according to claim 20, wherein the digital signal compression method is shared with a spectrum on a time axis after orthogonal transformation and / or an orthogonal transformation coefficient used for block floating. 22. A digital signal compression method having the functions of the digital signal compression method according to claim 18 and the digital signal compression method according to claim 20. 23. When determining the temporal length of a processing block based on the change of the input signal of the processing block, the periodic change of the input signal and / or the repeating pulse or periodic characteristic is used. 23. The digital signal compression method according to claim 7, 8, 9, 18, 19, 20, 21 or 22, wherein the determination is performed. 24. In a digital signal compression apparatus for compressing information of a digital signal, a characteristic calculation means for calculating a characteristic of an input signal on a frequency axis, and the input signal is distributed to small blocks subdivided in time and frequency and compressed. And a block floating means for performing floating for controlling the frequency magnitude of the small blocks to be subjected to the floating according to the characteristics of the input signal on the frequency axis. And a digital signal compression device. 25. In a digital signal compression apparatus for compressing information of a digital signal, spectrum data forming means for obtaining spectrum data on a frequency axis of an input signal, and allowable noise spectrum calculating means for obtaining an allowable noise spectrum from the spectrum data. And an allowable noise spectrum calculated by the allowable noise spectrum calculation means, which is divided into small blocks subdivided in time and frequency, and bit allocation means for allocating bits for compression. A digital signal compression apparatus characterized in that bit allocation is optimized by changing the frequency size of a small block for bit allocation for compression according to the characteristics on the axis. 26. In a digital signal compression apparatus for compressing information of a digital signal, characteristic calculation means for calculating a characteristic of an input signal on a frequency axis, and the input signal is divided into small blocks subdivided in time and frequency, and compressed. By means of a block floating means for performing floating for this purpose, a spectrum data forming means for obtaining spectrum data on the frequency axis, an allowable noise spectrum calculating means for obtaining an allowable noise spectrum from the spectrum data, and an allowable noise spectrum calculating means. The obtained allowable noise spectrum is divided into small blocks that are subdivided with respect to time and frequency, and bit allocation means that performs bit allocation for compression and floating for compression according to the characteristics of the input signal on the frequency axis. A small block that is subdivided in terms of time and frequency of application And a size control means for changing the frequency size of a small block subdivided with respect to time and frequency for bit allocation for compression and / or compression, and the small block and bit allocation for the block floating. A digital signal compression device characterized by optimizing block floating and bit allocation by changing the small blocks together or independently. 27. A small block which is subdivided in time and frequency for applying block floating for compression, and / or
Alternatively, the frequency size of a small block subdivided with respect to the time and frequency for bit allocation for compression is set in advance, and the compression efficiency and the characteristics of the input signal on the frequency axis when this predetermined size is adopted. According to the above, a comparison is made with the compression efficiency when the frequency size of the small block to be subjected to the floating for compression and / or the small block to which the bit allocation for the compression is changed is changed. 27. The digital signal compressing apparatus according to claim 24, further comprising block determining means for selecting a size of a small block capable of performing information compression with higher efficiency. 28. A small block subdivided in time and frequency for performing block floating for compression and a small block subdivided in time and frequency for bit allocation for compression are arranged on the frequency axis of an input signal. 28. The digital signal compression apparatus according to claim 26 or 27, wherein the digital signal compression apparatus is configured in common or independently according to the characteristics of. 29. A time length of a processing block for information compression is made variable according to an input signal, and a change of an input signal of the processing block and a change of an input signal of another processing block, and / or a power. Or a processing block length determining means for determining the temporal length of the processing block based on energy or peak information.
The digital signal compression device according to any one of claims 4 to 28. 30. A time length of a processing block for information compression is made variable according to an input signal, and a change in the input signal of the processing block and an input signal having a time width longer than the maximum of the processing block in terms of time. The processing block length determining means for determining the temporal length of the processing block based on the change information of the input signal obtained by the above step is provided.
The digital signal compression device according to any one of claims 4 to 28. 31. A time length of a processing block for information compression is made variable by adapting to an input signal, and a change of an input signal of the processing block and a change of an input signal of another processing block and / or a power. Alternatively, the function of determining the temporal length of the corresponding processing block based on energy or peak information, and the temporal length of the processing block is made variable by adapting to the input signal, and the input signal of the processing block is changed. A processing block length determining means having a function of determining the length of the processing block on the basis of change information and change information of the input signal obtained by an input signal having a time width longer than the maximum of the processing block in terms of time. 29. The digital signal compression apparatus according to claim 24, wherein the digital signal compression apparatus is a digital signal compression apparatus. 32. The processing block length determining means uses a combination of a ratio, which is fixed or adapted to an input signal, and / or a predetermined ratio, which is involved in the determination of an element that determines the temporal length of a processing block. 32. The digital signal compression device according to claim 31, which is also used alone. 33. The input signal is an audio signal, and the frequency width of a block that controls the generation of at least most of the quantization noise is made wider in a higher frequency range. One of
The digital signal compression device according to the item. 34. Split orthogonal transform means for splitting a time axis signal into a plurality of bands on the frequency axis by using orthogonal transform, and a window function used in the orthogonal transform together with variable orthogonal transform size in the orthogonal transform. 34. The digital signal compressing apparatus according to claim 33, further comprising an orthogonal transform size and window function shape changing means for changing the shape of. 35. The division orthogonal transformation means divides the time axis signal into a plurality of bands, forms a block made up of a plurality of samples for each of the divided bands, and performs an orthogonal transformation for each block of each band. The digital signal compression apparatus according to claim 34, wherein coefficient data is obtained by performing the operation. 36. The digital signal compressing apparatus according to claim 35, wherein the division frequency width in dividing the time axis signal before orthogonal transformation into a plurality of bands on the frequency axis is made wider toward a substantially higher frequency band. 37. The digital signal compressing device according to claim 36, wherein the divided frequency widths are the same in two consecutive bands of the lowest band. 38. The allocation of a compression code to main information and / or sub-information for a signal component in a band substantially equal to or more than a signal pass band is prevented in the bit allocation. Digital signal compression device. 39. The digital signal compression apparatus according to claim 35, wherein a quadrature mirror filter is used for the division into the plurality of bands. 40. The method according to claim 34, wherein a modified discrete cosine transform is used as the orthogonal transform.
9. The digital signal compression device according to any one of 9. 41. When determining a temporal length of a processing block based on a change of an input signal of the processing block, a predetermined boundary value for determining the temporal length is set as an amplitude and a frequency of the input signal. 33. The digital signal compressing device according to claim 29, wherein the digital signal compressing device is variable according to. 42. The digital signal compression apparatus according to claim 41, wherein the boundary value takes a plurality of stepwise values according to the amplitude and frequency of the input signal. 43. The processing block length determining means determines an auditory characteristic of a signal of the other processing block on a signal of the processing block, energy of spectrum and / or orthogonal transform coefficient on a frequency axis, and / or 32. The digital signal compression apparatus according to claim 29, wherein the processing is performed using power or peak information to determine the temporal length of the processing block. 44. Bit allocation for compression of spectrum and / or orthogonal transform coefficients on the frequency axis used in calculating the auditory characteristics of the signal of the other processing block on the signal of the processing block. 44. The digital signal compression apparatus according to claim 43, wherein the digital signal compression apparatus is also used as a spectrum on a time axis after orthogonal transformation and / or an orthogonal transformation coefficient used for block floating. 45. A digital signal compressing apparatus having the functions of the digital signal compressing apparatus according to claim 41 and the digital signal compressing apparatus according to claim 43. 46. When determining the temporal length of a processing block based on the change of the input signal of the processing block, the periodic change of the input signal and / or the repeating pulse or periodic characteristic is used. 46. The digital signal compression device according to claim 30, 31, 32, 41, 42, 43, 44 or 45, wherein the determination is made. 47. The digital signal compression method according to any one of claims 1 to 23, or the digital signal compression method according to any one of claims 24 to 46. A recording medium characterized by recording compressed data compressed by. 48. One of claims 1 to 23, characterized in that compressed compressed data is transmitted.
A method for compressing a digital signal according to item. 49. The digital signal compressing device according to claim 24, which transmits compressed compressed data.