Low bit rate multichannel audio coding methods and apparatus using non-linear adaptive bit allocation

Description Submit all comments and votes

BACKGROUND OF THE INVENTION

1. Field of the Invention

This invention relates to a low bit rate encoder and a low bit rate encoding method for compression-encoding audio signals of multi-channel system, a low bit rate decoder and a low bit rate decoding method for decoding compression-coded signals, and recording media on which signals encoded by such encoder/encoding method are recorded, which are used for cinema film projection systems or stereo or multi-sound acoustic systems such as video tape recorder or video disc player, etc.

2. Description of the Related Art

Various efficient encoding techniques and devices for audio or speech signals, etc. are known.

As an example of the efficient encoding technique, there is a blocking frequency band division system, which is the so called transform coding, for blocking, e.g., an audio signal, etc. in a time domain, to thereby transform signals in the time domain each of blocks into signals on the frequency domain for every block of time (orthogonal transform) thereafter to divide them into signal components in a plurality of frequency bands to encode those signal components every respective frequency band.

Moreover, there can be enumerated sub-band coding (SBC) which is non-blocking frequency band division system in which an audio signals, etc. in the time domain is divided into signal components in a plurality of frequency bands without blocking such signals every unit time thereafter to encode the signals.

Further, there have been proposed efficient coding techniques and devices in which the sub-band coding and the transform coding described above are combined. In this case, e.g., an input signal is divided into signal components in a plurality of frequency bands by the sub-band coding thereafter to orthogonally transform signals for every respective frequency bands into signals in the frequency domain to implement coding to these orthogonally transformed signal components in the frequency domain.

Here, as a filter for frequency band division of the above-described sub-band coding, there is, e.g., a filter of QMF, etc. Such filter is described in, e.g., the literature "Digital coding of speech in subbands" R. E. Crochiere, Bell Syst. Tech. J., Vol. 55, No. 8, 1976. This filter of QMF serves to halve the frequency band into bands of equal bandwidth. This filter is characterized in that so called aliasing does not take place in synthesizing the above-mentioned divided frequency bands at later processing stage.

Moreover, in the literature "Polyphase Quadrature filters-A new subband coding technique", Joseph H. Rothweiler ICASSP 83, BOSTON, filter division technique of equal bandwidth is described. This polyphase quadrature filter is characterized in that division can be made at a time in dividing a signal into signal components in a plurality of frequency bands of equal bandwidth.

Further, as the above-described orthogonal transform processing, there is, e.g., such an orthogonal transform system to divide an input audio signal into blocks by a predetermined unit time (frame) to carry out Fast Fourier Transform (FFT), Discrete Cosine Transform (DCT), or Modified DCT Transform (MDCT), etc. for every respective blocks to thereby transform signals in the time domain into those in the frequency domain.

This MDCT is described in the literature "Subband/Transform Coding Using Filter Bank Designs Based on Time Domain Aliasing Cancellation", J. P. Princen A. B. Bradley, Univ. of Surrey Royal Melbourne Inst. of Tech. ICASSP 1987.

Further, as frequency division width in the case of encoding (quantizing) respective frequency components divided into frequency bands, there is band division in which, e.g., hearing sense characteristic of the human being is taken into consideration. Namely, there are instances where an audio signal is divided into signal components in plural (e.g., 25) bands by a bandwidth such that the bandwidth becomes broader according as frequency shifts to higher frequency band side, which is generally called critical band.

In addition, in encoding data every respective bands at this time, coding by a predetermined bit allocation every respective bands or adaptive bit allocation every respective bands is carried out.

For example, in encoding coefficient data obtained after MDCT processing, coding is carried out by adaptive allocation bit number with respect to MDCT coefficient data obtained by MDCT processing for every respective band for every respective blocks.

As the bit allocation technique and device therefore, the following two techniques and device are known.

For example, in the literature "Adaptive Transform Coding of Speech Signals", IEEE Transactions of Acounstics, Speech, and Signal Processing, vol. ASSP-25, No. 4, August 1977, bit allocation is carried out on the basis of magnitudes of signals every respective bands.

Moreover, for example, in the literature "The critical band coder--digital encoding of the perceptual requirements of the auditory system", M. A. Kransner MIT, ICASSP 1980, there are described the technique and the device in which necessary signal-to-noise ratios are obtained every frequency bands by making used of the hearing sense masking to carry out fixed bit allocation.

Meanwhile, in the efficient compression encoding system for audio signals using subband coding, etc. as described above, such a system to compress audio data by making use of the characteristic of the hearing sense of the human being so that its data quantity becomes equal to about 1/5 has been already put into practice.

It should be noted that there is a system called ATRAC (Adaptive TRansform Acoustic Coding, a Trade Mark of SONY Corporation) used in, e.g., MD (Mini Disc, a Trade Mark of SONY Corporation) as the efficient encoding system of compressing audio data so that its data quantity becomes equal to about 1/5.

However, in the efficient coding system utilizing the characteristic of the hearing sense of the human being, there are instances where a sound of a musical instrument or a voice of a human being, etc. obtained by compression-coding a speech signal thereafter decoded, might be changed from the original sound although such a phenomenon takes place to a little degree. Particularly, in the case where this efficient coding system utilizing the characteristic of the hearing sense is used as a recording format for recording media for which fidelity reproduction of original sound is required, realization of higher sound quality is required.

On the contrary, a format of such an efficient coding system (ATRAC system, etc.) to compress audio signal so that its signal (data) quantity becomes equal to about 1/5 has been already put into practice, and hardware employing such a format is being popularized.

Accordingly, implementation of change or expansion having no compatibility of the format is disadvantageous not only to manufacturers (makers) which have used the format but also to general users.

For this reason, it is expected that high sound quality be attained by encoding or decoding device without changing the format itself.

As the method of realization of higher sound quality except for the above, it is conceivable to mix linear PCM sound into ordinary compressed data. However, since compressed data of the efficient coding system and linear data are different in length of frame and time length per each frame, it is difficult synchronize at the time of reproduction. Accordingly, it is very difficult to use these data of two formats at the same time.

Further, not only in the case of ordinary audio equipment, but also in, cinema film projection system, high definition television, or stereo or multi-sound acoustic system such as video tape recorder or video disc player, etc., audio signals of 4 to 8 channels are being handled. It is also expected that efficient coding to reduce the bit rate would apply to such plural channel systems.

Particularly, in the cinema film, there are instances where, digital audio signals of 8 channels, namely of left channel, left center channel, center channel, right center channel, right channel, surround left channel, surround right channel and sub-woofer channel are recorded. In this case, the above-mentioned efficient coding to reduce bit rate is required.

It is difficult to provide on cinema film an area capable of 8 channels of linearly quantized audio data of sampling frequency of 44.1 kHz and 16 bits as used in so called CD (Compact Disc), etc. Accordingly, compression of the audio data is required.

It should be noted that channels of 8 channel data recorded on the cinema film respectively correspond to left speaker, left center speaker, center speaker, right center speaker, right speaker, surround left speaker, surround right speaker, and sub-woofer speaker, which are disposed on the screen side where, pictures reproduced from the picture recording areas of cinema film are projected by projector.

The center speaker is disposed in the center on the screen side, and serves to output reproduced sound by audio data of center channel. This center speaker outputs the most important reproduced sound, e.g., speech of actor, etc.

The sub-woofer speaker serves to output reproduced sound by audio data of sub-woofer channel. This sub-woofer speaker effectively outputs sound which feels as vibration rather than sound in low frequency band, e.g., sound of explosion, and is frequently used effectively in scene of explosion.

The left speaker and the right speaker are disposed on left and right sides of the screen, and serve to output reproduced sound by audio data of left channel and reproduced sound by audio data of right channel, respectively. These left and right speakers exhibit stereo sound effect.

The left center speaker is disposed between the left speaker and the center speaker, and the right center speaker is disposed between the center speaker and the right speaker. The left center speaker outputs reproduced sound by audio data of left channel, and the right center speaker outputs reproduced sound by audio data of right center channel. These left and right center speakers perform auxiliary roles of the left and right speakers, respectively.

Particularly, in movie-theater having large screen and large number of persons to be admitted, etc., there is the drawback that localization of sound image becomes unstable in dependency upon seat positions. However, the above-mentioned left and right center speakers are added to thereby exhibit effect in creating more realistic localization of sound image.

Further, the surround left and right speakers are disposed so as to surround spectators' seats. These surround left and right speakers serve to respectively output reproduced sound by audio data of surround left channel and reproduced sound by audio data of surround right channel, and have the effect to provide reverberation or impression surrounded by hand clapping or shout of joy. Thus, it is possible to create sound image in more three-dimensional manner.

In addition, since a defect, is apt to take place on the surface of a medium of cinema film, if digital data is recorded as is, missing data occurs to a great degree. Such a recording system cannot be employed from a practical point of view. For this reason, error correcting code ability is the very important.

Accordingly, with respect to the data compression, it is necessary to carry out compression processing to such a degree that recording can be made in the recording area on the film by taking bits for correcting code into consideration.

From facts as described above, as the compression method of compressing digital audio data of 8 channels as described above, there is applied the efficient coding system (e.g., the ATRAC system) to attain sound quality comparable to CD by carrying out optimum bit allocation by taking into consideration the characteristic of the hearing sense of the human being as described above.

However, with this efficient coding system, sound of general musical instrument or voice of the human being, etc. is varied from original sound similarly to the above although such a phenomenon takes place to a little degree. For this reason, in the case where such a system is employed in recording format for which reproduction having fidelity to original sound is required, any means for realizing higher sound quality is required.

This problem always exists as long as in the case where systems except for the above-mentioned efficient coding system is used as multi-channel recording format in the cinema film, irreversible compression system is employed from a viewpoint of ensuring of the recording area.

Moreover, in a system for implementing efficient coding to audio signals of the multi-channel system as described above, data of respective channels are independently caused to undergo compression processing.

For this reason, even if, e.g., a certain one channel is in unvoiced sound state, fixed bit (byte) allocation amount is allocated to that channel.

Giving fixed bit allocation amount to the channel in unvoiced sound state as stated above is redundant.

Moreover, since bit allocation amounts are the same also with respect to a channel of signal of low level and a channel of signal of high level, if bit allocation amounts are evaluated over respective channels, redundant bits exist.

It is considered that particularly in the case where bit allocation amounts are fixed every respective channels, redundancy as described above becomes more conspicuous.

OBJECT AND SUMMARY OF THE INVENTION

With the above in view, an object of this invention is to provide an encoder and an encoding method capable of eliminating redundancy of bit allocation amount in compression-coding in the multi-channel system and of realizing a higher quality of compression-coding, a decoder and a decoding method corresponding thereto, and recording media on which compression-coded signals are recorded.

To achieve the above-mentioned object, in accordance with this invention, there is provided a low bit rate encoder for compression-encoding digital audio signals of a plurality of channels by making use of both the property of the audio signal and the hearing sense the human being, the encoder comprising: energy detecting means for detecting energies of the digital audio signals every digital audio signals of the respective channels; bit allocation amount determining means for determining bit allocation amounts to the respective channels on the basis of the detected result; compression-coding means for compression-coding the digital audio signals on the basis of bit allocation amounts allocated every respective channels in accordance with the determined bit allocation amounts; and multiplexing means for multiplexing the compression-coded signals every respective channels. The bit allocation amount determining means is operative to determine respective bit allocation amounts so that the relationship between energy and bit allocation amount of digital audio signal represents a non-linear characteristic such that according as energy of the digital audio signal increases, bit allocation amount increases as a whole. Further, variable bit allocation is carried out between channels with respect to samples in the time region and samples in frequency region of the audio signals of a plurality of channels.

In the low bit rate encoder of the first embodiment according to this invention, the energy detecting means is an amplitude information detecting means for detecting amplitude information of digital audio signals of respective channels before undergone compression-coding. Further, the bit allocation amount determining means determines bit allocation amounts to respective channels on the basis of a change at point of time of the amplitude information.

In this case, the bit allocation amount determining means calculates (determines), by a predetermined conversion formula, bit allocation amounts with respect to peak values of amplitude information of respective channels on the basis of the hearing sense characteristic, thus to determine amounts of bits to be allocated to respective channels on the basis of the conversion result.

Moreover, the bit allocation amount determining means respectively determines estimated amounts of bit amounts to be allocated to respective channels from the predetermined conversion formula to allocate bit allocation amounts of respective channels in proportion to the respective estimated amounts to thereby allow the total bit allocation quantity of all channels to be fixed.

On the other hand, the low bit rate decoder of the first embodiment according to this invention includes decoding means for decoding signals of respective channels encoded by the low bit rate encoder of the first embodiment.

Further, in the low bit rate encoder of the second embodiment according to this invention, the energy detecting a means is means for detecting change at a point of time of a predetermined scale factor (normalized value of the two-dimensional areas of time and frequency (block floating units)) with respect to signals of the respective channels, and the bit allocation amount determining means serves to carry out variable bit allocation between channels dependent upon change of scale factors.

Also in low bit rate encoder of the second embodiment, the bit allocation amount determining means calculates (determines), by a predetermined conversion formula, bit allocation amount with respect to change in point of time of sum total of scale factors of respective channels on the basis of the characteristic of the hearing sense of the human being to determine bit amounts to be allocated to the respective channels on the basis of the conversion result.

Further, the bit allocation amount determining means respectively calculates (determines) estimated amounts of bit amounts to be allocated to respective channels from the predetermined conversion formula to allocate bit allocation amounts of respective channels in proportion to respective estimated amounts to thereby allow total bit allocation amount of all channels to be fixed.

In addition, the low bit rate decoder of the second embodiment according to this invention includes decoding means for decoding signals of respective channels encoded by the low bit rate encoder of the second embodiment.

In accordance with this invention, in compression-coding audio data of a plurality of channels, since there is employed an approach to determine bit allocation amounts for respective channels on the basis of changes at a point of time of energies of respective channels thus to carry out compression-coding, bit allocation in correspondence with information amounts can be carried out with respect to respective channels.

In addition, in accordance with this invention, in compression-coding audio data of a plurality of channels, the relationships between energies and bit allocation amounts at respective channels are made non-linear to carry out compression-coding on the basis of the bit allocation amounts. For this reason, it is possible to carry out bit allocation in correspondence with information amounts with respect to respective channels.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a circuit diagram showing, in a block form, outline of the configuration of a low bit rate encoder of a first embodiment according to this invention.

FIG. 2 is a circuit diagram showing, in a block form, outline of the configuration of a low bit rate decoder of first and second embodiments according to this invention.

FIG. 3 is a circuit diagram illustrated in a block form for explaining bit allocation in a low bit rate encoder of the ATRAC system and low bit rate encoder of the embodiment according to this invention.

FIG. 4 is a view for explaining the state of recording of data within sound frame.

FIG. 5 is a graph for explaining bit allocation amount in the first embodiment.

FIG. 6 is a flowchart for explaining the operation of determination of bit allocation amount in the first embodiment.

FIG. 7 is a circuit diagram showing, in a block form, outline of the configuration of a low bit rate encoder of a second embodiment according to this invention.

FIG. 8 is a graph for explaining bit allocation amount in the second embodiment.

FIG. 9 is a flowchart for explaining the operation of determination of bit allocation amount in the second embodiment.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

Preferred embodiments of this invention will now be described with reference to the attached drawings.

The fundamental configuration of a first embodiment according to this invention is shown in FIGS. 1 and 2. The configuration of a low bit rate encoder of the first embodiment is shown in FIG. 1, and the configuration of a low bit rate decoder of the first embodiment is shown in FIG. 2.

The configuration of the encoder shown in FIG. 1 will be first described.

Audio signals of a plurality of channels (ch1, ch2, . . . , chn) are sent to sampling and quantizing elements 100.sub.1 .about.100.sub.n corresponding to respective channels via input terminals 20.sub.1 .about.20.sub.n and transmission paths 1.sub.1 .about.1.sub.n similarly corresponding to respective channels. At these sampling and quantizing elements 100.sub.1 .about.100.sub.n, audio signals of respective channels are converted into quantized signals. Quantized signals from these sampling and quantizing elements 100.sub.1 .about.100.sub.n are sent to amplitude information detecting circuit 200 and delay lines 300.sub.1 .about.300.sub.n via respective transmission lines 2.sub.1 .about.2.sub.n.

The amplitude information detecting circuit 200 detects amplitude information from quantized signals of respective channels. Namely, this amplitude information detecting circuit 200 detects peak values of amplitude information for every periods corresponding to the number of samples (hereinafter referred to as time blocks) of audio data processed at a time by encoding elements 400.sub.1 .about.400.sub.n which will be described later to send (transfer) these peak values to bit allocation determining circuit 500 via transmission lines 4.sub.1 .about.4.sub.n corresponding to respective channels. It should be noted that this amplitude information detecting circuit 200 may be of a structure to detect amplitude information by signals from transmission lines 1.sub.1 .about.1.sub.n.

The bit allocation determining circuit 500 determines, of conversion, bit allocation amounts for every respective channels from peak values of every respective channel in a manner described later to send (transfer) these bit allocation amounts to respective encoding elements 400.sub.1 .about.400.sub.n via transmission lines 5.sub.1 .about.5.sub.n.

Moreover, the delay lines 300.sub.1 .about.300.sub.n delay signals which have been received through transmission lines 2.sub.1 .about.2.sub.n by the time blocks to send (transfer) these delayed signals to respective encoding elements 400.sub.1 .about.400.sub.n through respective transmission lines 3.sub.1 .about.3.sub.n.

Respective encoding elements 400.sub.1 .about.400.sub.n carry out a compressing operation for every time block. Bit allocation amounts received through transmission lines 5.sub.1 .about.5.sub.n at this time reflect peak information of signals received through the transmission lines 3.sub.1 to 3.sub.n. Respective encoding elements 400.sub.1 to 400.sub.n compress signals which has been received through the transmission lines 3.sub.1 .about.3.sub.n so that their bit allocation amounts are equal to bit allocation amounts which have been received through the transmission lines 5.sub.1 .about.5.sub.n to send (transfer) these compressed signals to formatter 600 via respective transmission lines 6.sub.1 .about.6.sub.n.

The formatter 600 implements error correcting processing to the compressed signals for every channel which has been received via the transmission lines 6.sub.1 .about.6.sub.n in accordance with a predetermined format to compose them into a bit stream for transmission or for recording onto recording medium. This bit stream is outputted from output terminal 21 via transmission line 7.

Further, this bit stream is written into predetermined areas 28 on cinema film 27 by laser recording unit 26, for example. In the figure, reference numeral 29 denotes perforations adapted so that sprockets of projector (not shown) for film feeding are engaged therewith. The recording areas 28 are provided, e.g., between the perforations 29.

The configuration of low bit rate decoder of this embodiment will now be described with reference to FIG. 2.

Bit stream composed by the encoder (low bit rate encoder) of FIG. 1 is transmitted or is recorded onto a recording medium. This recorded bit stream is delivered to input terminal 22 via a predetermined reproducing unit (not shown), and is then sent from this input terminal 22 via transmission line 8 to deformatter 700.

This deformatter 700 decomposes the bit stream which has been sent through the transmission line 8 into compressed signals for every respective channel in accordance with a predetermined format. The decomposed compressed signals of every respective channel are sent to decoding elements 800.sub.1 .about.800.sub.n via corresponding to transmission lines 9.sub.1 .about.9.sub.n.

Respective decoding elements 800.sub.1 .about.800.sub.n expand the compressed signals which have been sent via the respective transmission lines 9.sub.1 .about.9.sub.n to send them to D/A (digital/analog) converters 900.sub.1 .about.900.sub.n via corresponding respective transmission lines 10.sub.1 .about.10.sub.n.

Respective D/A converters 900.sub.1 .about.900.sub.n convert the expanded signals (digital signals) which have been sent via the respe