AU683614B2

AU683614B2 - Process for reducing data in the transmission and/or storageof digital signals from several interdependent channels

Info

Publication number: AU683614B2
Application number: AU40593/93A
Authority: AU
Inventors: Karl-Heinz Brandenburg; Ernst Eberlein; Jurgen Herre; Dieter Seitzer
Original assignee: Fraunhofer Gesellschaft zur Forderung der Angewandten Forschung eV
Current assignee: Fraunhofer Gesellschaft zur Forderung der Angewandten Forschung eV
Priority date: 1992-05-25
Filing date: 1993-05-18
Publication date: 1997-11-20
Anticipated expiration: 2013-05-18
Also published as: CA2118402A1; WO1993025015A1; JPH08501666A; KR950701786A; EP0642719B1; NO944285D0; KR100240439B1; UA34457C2; CA2118402C; DE59303771D1; RU2119259C1; JP3552232B2; DE4217276C1; NO307280B1; EP0642719A1; ATE142829T1; NO944285L; AU4059393A

Abstract

The description relates to a process for reducing data in the transmission and/or storage of digital signals from several interdependent channels. As the signals from the channels are not mutually independent, it is possible further to reduce the quantity of data during transmission. In prior art processes for coding signals from two stereo channels, it is possible with a poor signal composition to cause interference through the coding which is not covered by the signal actually present in the channel and detract from listening enjoyment. In the process of the invention, by the use of the spectral values of corresponding blocks, the necessary data rate for separate coding of the two channels and for joint coding is balanced. The channels are jointly coded as long as the data rate for joint coding does not exceed the data rate for the separated coding of the channels concerned by a predeterminable value. In this case the spectral values of the corresponding blocks are similar so that joint coding is preferred. With very dissimilar signals, balancing results in separate signal transmission. The joint coding of several channels is temporarily suspended in order to be certain of preventing interference.

Description

OPI DATE 30/12/93 AOJP DATE 10/03/94 APPLN. ID 40593/931111IIIIIIIiIIJlIIII111 PCT NUMBER PCT/DE93/00448 111 i hlllElllllllllU AU9340593 (51) Internationale Patentlassifl kati on 5 H04H 5/00, H04S 1/00 F104B 1/66 (11) Internationale Veriiffentlichungsnummer: Al (43) Internationales Veriiffentlichungsdatum: 9. Dez WO 93/25015 ~ember 1993 (09. 12,93) (21) Internationales Aktenzeichen: (22) Internationales; Anmeldedatum: PCT/DE93/00448 I8, Mai 1993 (18.05.93) (81) llestimmutigsstaaten: AU, CA, JP, KR, NO, RU, UA, US, europflisches Patent (AT, BE, CH, DE, DK, ES, FR, GB, GR, IE, IT, LU, MC, NL, PT, SE).

Ver~ffentlicht k1it inzternationalen Recherchenberich,.

Prioritfitsdaten: P 42 17276.4 25. Mai 1992 (25.05.92) (71) Anmelder (fir alle Bestimrnungsstaaten ausser US): FRAUNHOFER-GESELLSCHAFT ZUR FORDE- RUNG DER ANGEWANDTEN FORSCHUNG EN.

IDE/DEJ; Leonrodstrage 54, D-8000 Milnchen 19 (DE).

(72) Erfinder; und Erfinder/Anmrelder (nur flir US) HERRE, J~lrgen [DE/ DE]; Am R6thelheim 9, D-8520 Erlangen SElT- ZER, Dieter [DE/DE]; Humboldtstrage 14, D-8520 Erlangen BRANDENBURG, Karl-Heinz [DE/DE]; Haagstra~e 32, D-8520 Erlangen EBERLEIN, Ernst [DE/DE]; Waldstraae 28 b, D-8521 Grog.enseebach (DE).

6-',8 3014 (54)Title: PROCESS FOR REDUCING DATA IN THE TRANSMISSION AND/OR STORAGE OF DIGITAL SIGNALS FROM SEVERAL INTERDEPENDENT CHANNELS (54) Bezeicltnung: VERFAHREN ZUR REDUZIERUNG VON DATEN BE] DER GBERTRAGUNG UND/ODER SPEI- CR ERUNG DIGITALER SIGNALE MEHRERER VONEINANDER ABHANGIGER KANXLE lath Listening level calculation 2a Pnlhresholid LLThrshldR 2b PE(Threshold R.R(f)) 6a PEtMin Threshold. M(f))2o ELtPC).(R 6b PE Min Threshold. S(f)) Lit) LOT BerechmV, der 2 RU) B.,hnea ltMS Stereo on] _P (57) Abstract PE(Minsthw, Sit) 66 The description relates to a process for reducing data in the transmission and/or storage of digital signals from several interdependent channels. As the signals from the channels are not mutually independent, it is possible further to reduce thle quantity of data during transmission. In prior art processes for coding signals from two stereo channels, it is possible with a poor signal composition to cause interference through the coding which is not covered by the signal actually present in the channel and detract from listening enjoyment. In the process of the invention, by the use of the spectral values of corresponding blocks, the necessary data rate for separate coding or the two channels and for joint coding is balanced. The channels are jointly coded as long as the data rate for joint coding does not exceed the data rate for the separated coding of the channels concerned by a predeterminable value. In this case the spectral values of the corresponding blocks are similar so that joint coding is preferred. With very dissimilar signals, balancing results in separate signal transmission. The joint coding of severa.l channels is temporarily suspended in order to be certain of preventing interference,

DESCRIPTION

PROCESS FOR REDUCING DATA IN THE TRANSMISSION AND/OR STORAGE OF DIGITAL SIGNALS OF SEVERAL INTERDEPENDENT CHANNELS Technical Field The invention relates to a process for reducing data in the transmission and/or storage of digital signals of several interdependent channels.

Processes in which audio signals are transmitted in frequencycoded manner are e.g. known from PCT Publications WO 88/001811 and WO 89/08357 and reference should be made to these documents for the explanation of terms not explained here.

For the bit stream of the signal to be transmitted and the design of the coder the standardization committee Moving Pictures Experts Group (MPEG) of the International Standardization Organisation (ISO) has established ISO standard 11172-3.

Numerous known processes for the data-reduced coding of digital audio signals code the signals in the frequency range and use for the transmission of the signals from the time range into the frequency range (in spectral values) a suitable imaging method, e.g. FFT, DCT, MDCT, polyphase filter bank or hybrid filter bank.

These processes permit a high degree of utilization of signal redundancy and irrelevance with respect to the properties of the human ear. If in the transmission of signals of several channels the signals are not independent of one another, an additional reduction of the data quantity to be transmitted is possible. This prerequisite is fulfilled e.g. in the case of signals in the channels of a quadrophonic or sterophonic audio signal.

Recently further work has been carried out on the development of the international standard for the transmission of several 2 channels. The publication "MUSICAM-Surround: A Universal Multichannel Coding System Compatible with ISO 11172-3", 93 rd AES Convention, 1992, San Francisco, proposes a process for the transmission of up to five channels. For example, two stereo channels and in addition a middle channel and two side channels (3/2 stereophony) or two stereo channels and three commentary channels can be transmitted. Also in this process the signals are interdependent in the individual channels.

Prior Art A process for the utilization of redundancy/irrelevance between the two channels of a stereo audio signal is described in the publication by J.D. Johnston, "Perceptual Transform Coding of Wideband Stereo Signals, ICASSP 89. In this so-called M/S coding (middle/side coding) in place of the left and right channel, the sum middle) and the difference side) of the stereo signal are coded. This leads to a saving in the quantity of data to be transmitted.

The dependence of the signals of two stereo channels is also utilized in the intensity stereo process known from the publication "Subband Coding of Stereophonic Digital Audio Signals", Proc. of the ICASSP, 1991, pp 3601 to 3604. In this process the monosignal and an additional information on the left/right distribution of the signal are transmitted.

In both processes in the case of an unfavourable signal composition considerable interference can occur. For example, a widely differing signal composition in the left and right channel in M/S coding leads to defects, which cannot be concealed by the signal present in the channel. Therefore e.g. a loud saxophone signal, which is contained almost exclusively in the right channel, causes interference on the left channel, which is not concealed and can therefore clearly be heard. When using 3 the intensity stereo process the three-dimensional sound impression is disturbed if the left and right channels have widely differing spectral compositions.

Thus, the known processes are only usable if unfavourable signal compositions are not to be expected, or if it is possible to accept interference or noise in favour of reducing the data quantity. In addition, the known processes are restricted to stereo coding.

Description of the Invention The problem of the invention is to provide a process for the reduction of data in the transmission and/or storage of digital signals of several interdependent channels, in which the dependence of the signals in the different channels is utilized and which does not lead to subjectively perceptible interference to the transmitted signals. This problem is solved by a process having the features of claim I.

According to the invention, the signals of the different channels are initially passed into spectral ranges. Then, with the aid of the spectral values of corresponding blocks of several channels there is a balancing of the necessary data rate for a separate coding of the channels, i.e. for a monocoding of the channels and for a joint coding of several channels. A joint coding of the particular channel takes place for as long as the data rate for the joint coding does not exceed by a predeterminable value the data rate for the separate coding of the channels.

In this case the spectral values of the corresponding blocks are similar, so that a joint coding is preferred.

If the data rate for the joint coding exceeds by the predeterminable value the data rate for the monocoding, there is a monocoding of the channels. In this case the spectral values of the 4 corresponding blocks of the channels are so dissimilar that a very high data rate would be required in order not to produce defects in the corresponding channels due to the coding process and which cannot be concealed by the signal actually present in these channels. With a view to reliably avoiding such disturbing defects, in the process according to the invention temporarily no use is made of joint coding.

Further developments of the invention are characterized in the subclaims.

According to claim 2 for the balancing of the necessary data rates for the monocoding and/or the joint coding an estimator SF is formed, which indicates with which bit number each sample must be coded so as to keep below the monitoring threshold the interference caused by the coding process when the latter is predetermined.

In coding processes providing a fixed bit allocation for the individual frequency bands, the estimator can be calculated beforehand (cf. e.g. Theile Stoll G. Link M. Low bit-rate coding of high quality audio signals, in EBU Review No. 230, 1988, pp 158-181).

With coding processes having a variable bit number, such as is described in PCT Publication WO 88/001811, the estimator (SF) is to be chosen appropriately for the corresponding process. A preferred estimator is given hereinafter.

Thus, in the process according to the invention, the SF of the possible coding process, namely monocoding and joint coding of several channels undergo blockwise comparison. On the basis of the comparison a decision is made as to whether moirocoding or joint coding of several channels be carried out on the signal.

5 When comparing the estimators of monocoding and/or joint coding, according to claim 3 the necessary data rate for the separate coding of the signals is obtained by summation fe em the estimators of the particular channels.

In the case of stereo coding of two channels e.g. the estimator SF (LR) for the separate coding is formed by the addition of estimators SF and SF of the signals from the left (L) and the right channels.

In an advantageous development of the process the joint coding of the channels according to claim 4 takes place through the formation of linear combinations (matrixing of the input signals. Following the decoding of the signal the matrixing is cancelled out again (dematrixing). When transmitting two stereo channels the coding takes place with the aid of middle/side (M/S) stereo coding. In this coding in place of the left and right channel the sum middle) and the difference side) of the signals from the left and right channels are coded, which leads tc a saving on the data quantity to be transmitted.

According to claim 5 the estimator SF (Mijk) of the necessary data rate for the joint coding of signals from channels Ki, Kj, Kk, etc. is formed by the addition of the estimators SF SF SF Mi being the ith matrix channel.

When coding two stereo channels the estimator of the necessary data rate for the joint coding of the signals is obtained by the summation of the SF of the middle signal and the side signal The middle signal is formed from the signal of the left and the signal of the right channels through the relationship M L R and the side channel by the rela- L R tionship S L %2

/T

\u$vr o n" 6 In an advantageous development of the invention according to claim 6, prior to the balancing of the necessary data rate, the necessary precision of the coding is established and for this purpose a maximum permitted interference is predetermined. In no channel must the interference which occurs exceed this permitted interference.

For the coding of two stereo channels, as a function of the particular signal for both channels, that monitoring threshold is calculated from the acoustic model which guarantees this condition. From the two calculated monitoring thresholds use is made of the minimum threshold for the coding of both channels.

This process ensures that the permitted interference or noise signal is never exceeded.

In an advantageous further development of the invention according to claim 7 the so-called Perceptual Entropy (PE) is used for the estimator. The PE of an audio signal is the minimum data rate, which is required with predetermined codings in order to be able to code and decode the audio signal without a subjectively perceptible change. The determination of the PE is e.g. described in J.D. Johnston, Estimation of Perceptual Entropy, Using Noise Masking Criteria, ICASSP 1988, pp 2524-2527.

According to claim 8 the balancing of the necessary data rate takes place with the aid of the relationship SF C0 SF (Ki, Kj, Kk C2, in which C 1 and C 2 are predeterminable constants. If the spectral values from the corresponding blocks of the particular channels fulfil the indicated relationship, joint coding takes place. In the case of stereo coding of two channels the relationship is: SF (MS) C 1 SF (LR) C2.

THe smaller the constants C 1 and C 2 the more similar must be the spectral values of the signals from the channels in order to fulfil the conditions for joint coding. There is a correspon- 7 ding increase in the probability that an interference signal produced by another channel during coding is concealed by the signal actually present in the channel. Preferred values for the constants C 1 and C 2 are given in claims 9 and The essential advantages of the invention are that without any noteworthy increase in the complexity of the transmission process an additional data reduction is obtained. There is a reliable avoidance of defects which can be caused by a joint coding of several channels in in each case different channels. The process according to the invention is independent of the specific construction of the coding process used and is therefore universally usable. The process merely requires an additional signal processing in the coder required in small numbers on the transmitter side, but not in the decoder used in large numbers by final consumers.

Brief Description of the Drawings An embodiment of the process according to the invention is described hereinafter with reference to the drawing where, for ease of representation, the coding of two stereo channels is used as an example. The drawing is a block circuit diagram of the determination of the switch-on criterion for stereo coding.

Way of Performing the Invention The samples L and R of stereo signals from the left and right channel of a stereo source are past with the aid of a fast Fourier transformation (FFT) blockwise from the time range into the frequency range in spectral values L R and from the latter the monitoring thresholds (la, Ib) for both channels are calculated.

Then, for balancing the necessary data rates, on the one hand 8 for separate coding and on the other for joint coding, estimators are determined. In the represented, preferred embodiment perceptual entropy (PE) is used as the estimator.

For determining the PE (LR) for separate coding the PE's are in each case separately determined (2a, 2b) for the left and right channels using the calculated monitoring threshold. The total PE (LR) is obtained therefrom by summation: PE (LR) PE PE For the determination of the PE (MS) of the joint coding of the two channels, the spectral values undergo, in th.e preferred embodiment, a middle/side matrixing For this purpose the middle signal is formed by the relationship: M L R 12 and the side signal by the relationship S a L R The following determination of the PE's of the signals is based on the minimum size from the calculation of the monitoring thresholds of the two channels (la, Ib).

With the aid of this minimum monitoring threshold the PE's of the middle signal and side signal are separately determined (6a, 6b).

The joint PE (MS) for the stereo coding is obtained therefrom by summation: PE (MS) PE PE Finally, the balancing of the necessary data rate takes place. The audio signals undergo stereo coding if the condition PE (MS) C1 PE (LR) C2 is fulfilled. If the condition is not fulfilled, stereo coding does not take place so as to avoid disturbing defects as a result of coding.

The described embodiment in no way constitutes a restriction to a specific coding process. In fact, the process according to 9the invention is universally usable independently of the particular coding process.

,IRA(

Claims

2. Process according to claim 1, characterized in that for the balancing of the necessary data rate an estimator (SF) is formed, which indicates which bit number is required per block, so that for a predetermined coding process the interference caused by the latter are kept below the monitoring threshold.
3. Process according to claim 2, characterized in that the estimator (SF (Ki, Kj, Kk of the necessary data rate for the separate coding of channel signals is formed by the addition of the estimators (SF SP (KJ) of the signals from the channels (Ki, Kj, etc.).
4. Process according to claim 3, characterized in that the joint coding takes place by the formation of linear combinations (matrixing) of input signals and these are cancelled out during the decoding process (dematrixing). Process according to one of the claims 2 to 4, characterized I RA4, -7'1 11 in that the estimator (SF (Mijk of the necessary data rates for the joint coding of signals from channels (Ki, Kj, Kk is formed by the addition of the estimators (SF SF (Mj), SF (Mk) in which Mi is the ith matrixed channel.
6. Process according to one of the claims 1 to 5, characterized in that for the decoded signal in the channels a permitted max- imum interference is predetermined, that for said maximum inter- ference the particular monitoring threshold in the channels is calculated and used for the determination of the estimator for the joint coding.
7. Process according to one of the claims 2 to 6, characterized in that the perceptual entropy (PE) of the audio signal is used as the estimator.
8. Process according to one of the claims 3 to 7, characterized in that there is a joint coding of the signals of several channels (Ki, Kj,, Kk if the following condition is fulfilled, SF (Mijk C1 SF (Ki, Kj, Kk C2, in which C, and C 2 are predeterminable constants.
9. Process according to claim 8, characterized in that the constant C1 is between the values 1 and 2. Process according to claims 8 or 9, characterized in that the constant C 2 is zero. 12 ABSTRACT A process for reducing data in the transmission and/or storage of digital signals of several interdependent channels is des- cribed. As the signals from the channels are not independent of one another, an additional data quantity reduction is possible dur- ing transmission. In known processes for the coding of signals from two stereo channels, in the case of an Lnfavourable signal composition interference can be caused by the coding process and cannot be concealed by the signal actually present in the channel and consequently reduce listening enjoyment. In the process according to the invention and with the aid of the spectral values of corresponding blocks of several channels, there is a balancing of the necessary data rate for a separate coding of the two channels and for a joint coding. There is a joint coding of the channels for as long as the data rate for the joint coding does not exceed by a predeterminable value the data rate for the separate coding o. che particular channel. In this case the spectral values of the corresponding blocks are similar, so that a joint coding is preferred. In the case of highly dissimilar signals the balancing leads to a separate transmission of the signals. In order to ensure that there is no interference, temporarily no use is made of the joint coding of several channels.