`
`EflflLEKIPEH)AflREHMS]IJ
`CKDLELICHNICUKTWIJEHS
`
`JUNE 1992
`
`VOLUME 10
`
`NUMBER 5
`
`ISACEM
`
`(ISSN 0?33-8716)
`
`&
`‘E-
`In
`
`—
`
`W:
`F'-:3
`ea
`in
`
`2 EEu
`
`.
`4::
`
`
`
`.
`_,.A PUBLICATION OF THE IEEE COMMUNICATIONS SOCIETY
`
`3 Si"EECI-I. AND IMAGE CODING
`
`5.75
`
`EE2=
`
`Gue's't"'EiiittIr—N-. Htihing
`
`...........',"Guest Editorial . . N. H nbing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`.
`
`.
`
`.
`.
`
`;‘-PAPERS
`
`
`
`. .N. Joyrmt
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`ifiignal Compression: Technology Targets and Roses rch Directions [invited Paper”) .
`.5‘. Wang. A. Sekey, rind A. Gerrho
`.
`.
`.
`mn Objective Measure for Predicting Subjective Quality of Speech Coders (Invited Paper)
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`_
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`-A Low-Delay CELP Coder for the CCITT I"6__ Rb/s Speech Coding Standard (Invited .”:'Ip£’r)
`I
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. .J.-H. C,‘iien.1€. V. Cox. lr’.-C. Lin. N. Jayanr. and M. J. Meiclrner
`__.A H igh-Quality Multirate Real-Time CELP Coder Urtu.".'ed Paper) .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. P. Kmart and K. .S‘ivrunimm'iari
`;Teclmiques for improving the Performance of CELP-Type Speech Coders Un.ur'Ica' Paper) .
`.
`.
`.
`.
`. 1. A. GEJ".§'O.’l and M./1. Jariiik
`-‘Two-Channel Conjugate Vector Quantizer for Noisy Channei Speech Coding .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. T. Mariya
`Weighted Optimum Bit Allocations to Orthogonal Transforms for Picture Coding .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. B. Macq
`-'j’_Image Coding with the Discrete Cosine-Ill Transform .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. O. K. Ersoy and A. Nouira
`-Unified Variable-Length Transform Coding and Image-Adaptive Vector Quantization .
`.
`. L. Wang, M. Goldberg, rm.1l' .S'. Sfiiieri
`3-Adaptive Transform Tree Coding of Images .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. W. .'rt. Peariman, P. Jakordar. and M. M. Leimg
`
`'-Spectral Entropy-Activity Classification in Adaptive Transform Coding .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. . .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. R. Marrer and U. Franky
`
`5_Shape-Gain Vector Quantization for Noisy Channels with Applications to Image Coding .
`.
`.
`.
`. J. Rosebroc.-'< and P. W. Be.r.rh'ch
`;'Subband Image Coding Using Entropy-Coded Quantization over Noisy Channels .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. N. Tamrbe and N. Farm:-d:‘n
`5._:A Progressive Scheme for Digital Image I-ialfloning, Coding of Halftoncs. and Reconstruction .
`.
`. S. Ko."l'.I’a.t‘ and D. Ana.s'ro.r.'rr’otr
`_§Sing|e Bit-Map Block Truncation Coding of Color Images .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. Y. Wu and D. C. Coii
`_-Iplerframe Hierarchical Address-Vector Quantization .
`.
`.
`.
`. . .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. N. M. Nasrabadi’, C. Y. Chan, and J. U. Roy
`I.-A Fast Feature-Based Block Matching Algorithm Using integral Projections .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. J.-.5‘. Kiri: and R.-H. Park
`--A Transformation for the Calculation of Filter Pairs for Perfect
`licconstruction in Suhband Coding with Linequincnnx
`Subsampiing .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`. J. De Lamar’.-‘(rerun and G. Sdiriarief
`
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`. . .
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`.
`
`. . .
`
`.
`
`.
`
`.
`
`
`
`.
`
`
`
`RPX Exhibit 1134
`RPX v. DAE
`
`RPX Exhibit 1134
`
`RPX V. DAE
`
`
`
`
`
`-CALL FOR PAPERS
`.'
`
`Litllegrity of Public Teiecornmunication Networks .
`
`
`
`
`
`
`&
`IEEE COMMUNICATIONS SOCIETY E"
`
`G)
`
`
`
`S. B. WEINSTEIN, Director of Publications
`Belleore, Rm. 2L-287
`445 South Street
`Morristown, NJ 07960-1910
`
`IEEE COMMUNICATIONS SOCIETY
`JOURNAL Editorial Board I992
`
`W. H. TRAEITER. Editor-iii-Chief
`De . Elec.
`n .
`Univ. Missouri-Rolla
`Rolla, MO 65401-0249
`tranter@ee.ttmr.edu
`
`EU; L. Mli:DO;:)A:li;:t, Associate Editor
`e core, m.
`-
`l
`'
`445 South Street
`Morristown. NJ 07960-I910
`suc@thi.irripcr.be|lcore.com
`
`A. M. BUSH
`National Science Foundation
`[300 G. Street. NW
`Washington, DC 20550
`
`N. K. CHEUNG
`Bellcore, Rm. Nvc 32-219
`331 Newman Springs Rd.
`Red Bank, NJ 07701
`iijf.-_ HIAYE
`ep. E cc. En .
`Cfgloordia Univ.
`1
`5 De Maisonneuve.
`Montreal, Quebec,
`Canada H3G IM8
`
`M K
`. AWASHIMA
`Fujitsu Inst. Comp. Sci.
`l-l't'-25 Shinkamata
`Ota-I(i.i
`Tokyo 144-00. Japan
`%eM’§“§,"°£"C°m M E,
`Calf’,-etgn Univ
`1’
`-
`Colonel B Drgve
`Ottawa gm
`Canada‘ KlS'5B6
`NI.rt;LI%xBi1:icHUK
`A
`l Labs.
`600 Mountain Ave.
`Murray Hill, NJ 07974
`
`3-
`
`Senior Editorss
`. RAPPAPORT
`T.
`The Bradley Dep. Elcc. Eng.
`615 Whiltemore Hall
`Virginia Polytech. Inst. 3:
`‘Silo; Univfim
`0
`B ac
`urg,
`
`2406l-
`
`l I l
`
`D L T
`ENNENHOUSE
`.
`.
`M.I.T., Rm. NE43-S38
`545 Technology Square
`Cambfidgc. MA 92139
`
`A. A. ROBIIOCR
`_
`Italtel
`Via A. di‘To-cqucvilie 13
`20154 Milan, Italy
`W} D. SIlRCOS£t(3E28
`Be lcore, m.
`-
`445 South Street
`Morristown, NJ {N960-l9!0
`
`6
`
`IR. Yitriugosiii
`a 5.
`ujiisu
`‘l0lS Karnilmdanaka Nakahara-ltttrl
`Kawasaki 2| 1, Japan
`
`THE INSTITUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS, INC.
`Officers
`
`.
`ARVID G. LARSON, l/ice Presiclettt, Professional Activitles
`MERRILL W. BUCKLEY, J R.. President
`J . T. CAIN, Vice President, Pirblicatioii Acti'vlti'e.r
`MARTHA SLOAN, President-Elect
`huts T. GANDIA, Vice Piifesidenr, lffegionol Airtivgrtier
`KARSTEN E.“?Rfi.NGElD3Secretary
`ARCO W. MIGLIARO.
`ice Presi ent, Stan or s Activities
`THEODORE
`.
`ISSEY, R.. Treasurer
`FERNANDO ALDA NA, Vice Presideiit, Technical Acti‘vi‘tie.s
`EDWARD A. PARRISH, Vice President, Educational Activities
`FREDERICK T. ANDREWS, JR., Director, Division IH—Comniunr'coti’on.r Technology Division
`
`Headquarters Staff
`ERIC HERZ. Executive Director and General Manager
`THOMAS W. BARTLETT, Associate General Mai-ioger—Finance and Adininistrotion
`WILLIAM D. CRAWLEY, Associate General Manager-Programs‘
`JOHN H. POWERS. Associate Genera! Monoger— Volunteer Services
`
`DONALD CHRISTIANSEN, Editor, IEEE Spectrum
`IRVING ENGELSON, Staff Director, Technical Activities
`LEO FANNINO, Staff Director, Professional Activities
`W. R. HABINGREITHER, .S‘tofi'Dtrector, Customer Service Center
`PHYLLIS HALL, Staff Dlrector, Publishing Services
`
`MELVIN I. 0I.I<I~:N, Staff Director, Field Services
`EDWARD ROSENBERG, Controller
`ANDREW G. SALEM, Stofl Dlrector. Standards
`RUDOLF A. STAMPFL, Staff Director, Ediicotiorinl Actr‘vr‘ti'e.t
`
`Publications Department
`Publications Managers: ANN I-I. BURGMEYER, GAIL S. FERENC
`Managing Editor.‘ VALERIE CAMMARATA
`Associate Ea’i’ror.- SHARVN L. PERRY
`
`IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS is published nine times a year in January, February, April, May, June, Aagusi.SepIemb=1‘-it
`October. and December by The Institute of Electrical and Electronics Engineers, Inc. Responsibility For the contents rests upon the authors and not upon tl:l'_Gj
`IEEE, the SocietyfCot.mcil,‘or its members. IEEE Headquarters: 345 East 4‘? Street, New York. NY l00I'?. NY Telephone: 212-705 + extension: lnforrtlfl?
`lion -1900; General Manager -79 I0; Public Information -7867: Publishing Services -7560; S ectrum -7556. Telecopiers: N Y (Headquarters) 212-‘I52-4929,;
`NY (Publications) 2l2-705-7632; NY Telex: 236-4-ll (international messages only). IE E Service Center (for orders, subscriptions, address chan
`'-
`Bducational Activities, Region Sectionfstudent Services, Standards): 445 Hoes Lane, PO. Box 1331, Piscataway, NJ 03855-1331. NJ Telephone: In 011,;
`matiori: 908-981-0060; 908-56 + extension: Controller -5365; Technical Activities -3900. IEEE Washington Oflice (for U.S. professional activities): 1828-‘
`L Street, NW, Suite I202, Washington. DC 20036-5 l 04. Washington Telephone: 202-785-00] 1'. Priee,i'Publieatiori Information: Individual copies: IEEE1
`members $10.00 {first copy only]. nonmembers $20.00 per copy. (Note: Add $4.00 postage and handling charge to any order from 51.00 to $50.00. iijicludilllz
`prepaid orders.) Member and nonmember subscription riccs avaiiable on request. Available in microfiche and microfilm. Copyright and Reprint Pei'i:nl3--
`alone: Abstracting is permitted with credit to the source.
`ibrariesare permitted to photocopy beyond the limits of the U.S. Copyright Law for
`rivate use oi’-'
`patrons: I) those post~l9Ti articles that carry a code at the bottom of the lirst
`age, provided the per-copy feeindieated in the code is pai
`tlirough Ihl-‘v.
`Copyright Clearance Center, 29 Congress Street, Salem, MA 01970; 2) pie-1 78 articles without fee. Instructors are permitted to photocopy isolated:
`articles for noncommercial classroom use without fee. For all other copying. reprint, or republication permission, write to Copyrights and Permissions;
`Department, IEEE Publishing Services, 445 Hoes Lane, P. O. Box 1331, Piscatawa . NJ 03 355- [33]. Copyright @) I992 by The Institute of Electrical anti-
`Electronics Engineers. Inc. All rights reserved. Second-class postage paid at New orlt, NY. and at additional mailing offices. Postmaster: Send addrcs-‘ls
`changes to IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, IEEE, 445 Hoes Lane. PD. Box 1331, Piscataway, NJ 03855-t33t. GS _
`Registration No. 125634188.
`
`
`
`
`
`The l'ield of interest of the IEEE Communications Society consists of all telecommunications including telephone, telegraphy, facsimile. and pain
`point television, by electromagnetic propagation including radio; wire; aerial, undcrground,_coaiiial, and submarine cables; waveguides, oommunica
`satellites, and losers; in marine, aeronautical. space. and fixed station services: repeaters. radio relaying. signal storage. and regeneration: telecomrituni "
`tam error detection and correction: multiplexing and carrier techniques: communication switching systems: data communications: and communion
`l. eory.
`-
`_
`In addition to the above, this JOURNAL or the IEEE TRANSACTIONS ON COMMUNICATIONS contains papers pertaining to analog and digital sig '
`processing and modulation, audio and video encoding techniques, the theory and design of transmitters, receivers, and ‘repeaters for communications I
`optical and sonic media, the design and analysis of computer communication systems. and the development of communication_sol‘tware. Coniributio -'
`theory enhancing the understanding of communication systems and techniques are included, as are discussions of the social implications of the developm
`of communication technology. All members of the IEEE are eligible for membership in the Society upon payment of the annual Society membership fgg;
`Sl5.00. Members may receive this JOURNAL or the IEEE TRANSACTIONS ON COMMUNICATIONS upon palyment ofan additional 3 l 5.00 ($30.00 total),
`both publications upon payment ofan additional $30.00 ($45.00 total). For information on joining, write to I e IEEE at the address below. Member copr‘
`Transacti'on.t,lJottrnaIs are for personal use on!y.
`
`'
`
`..-
`
`
`
`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`
`
`t"l'!E!\" r-I
`
`('lf.Z A LOW-IJEI..r\Y CELP CODER
`
`Bil
`
`speech coding standards now already exist for specific ap-
`plications. To avoid proliferation of different 16 kb/s
`standards and the potential difliculty of interworking be-
`tween different standards, "in June 1988 the CCITT de-
`
`cided to investigate the possibility of establishing a single
`16 kb/s speech coding standard for universal applica-
`tions. The CCITT’s intended applications include video-
`phone, cordless telephone, digital satellite systems, Dig-
`ital Circuit Multiplication Equipment (DCME), Public
`Switched Telephone Network (PSTN), Integrated Service
`Digital Network (ISDN), digital leased lines, voice store
`and forward systems, voice messages for recorded an-
`nouncements,
`land-digital mobile
`radio,
`packetized
`speech, etc. [12].
`Because of the variety of applications this 16 kb/s
`standard has to serve, the CCITT determined a stringent
`set of performance requirements and objectives for the
`candidate coding algorithms. Not every application would
`need every requirement to be met. Yet, to be accepted by
`the CCITT as a universal 16 kb/s standard. a candidate
`algorithm must meet all of the requirements. The major
`requirement was that the speech quality should be roughly
`comparable to that of G.72l while the one-way encoder!
`decoder delay should not exceed 5 ms (the objective was
`52 ms) [12]. In other words, the CCITT was looking for
`a toll-quality low-delay speech coder at 16 kb/s.
`The CCITT specifies the speech quality requirements
`in terms of qdu, or the quantization distortion unit. By
`definition, one encoding with a 64 kb/s (3.71 1 PCM co-
`dec introduces 1 qdu of distortion (CCITT Recommen-
`dation G. l 13). For asynchronous tandem connections, the
`qdu of individual speech coders is supposed to be addi-
`tive. For example, N asynchronous tandeming stages of
`G.7ll codecs should result in N qdu. Another example is
`that a single encoding ofa 32 kb/s 0.721 ADPCM codec
`is rated at 3.5 qdu, and therefore two ADPCM encodings
`should be rated at 7 qdu and four encodings at 14 qdu.
`The CCITT performance requirements for the 16 kb/s
`standard specify that, for a clear channel {i.e., no bit er-
`rors), a candidate coder should produce 4 qdu or less for
`a single encoding and 14 qdu or less for three asynchro-
`nous tandeming stages. In effect, this says that a candidate
`coder can be slightly worse than G.72l ADPCM for a
`single encoding, and three asynchronous encodings of the
`candidate coder should match the speech quality of four
`asynchronous encodings of G.72l ADPCM. For noisy
`channels, the CCITT required that, for random bit errors
`at a bit error rate (BER) of l0‘3 or 104, a candidate coder
`should produce decoded speech quaiity not worse than that
`of G.721 ADPCM under the same conditions.
`In addi-
`
`tion, the coder should pass network signaling tones such
`as DTMF and CCITT Signaling Systems No. S, 6. and 7.
`It was not so difficult to meet each one of these quality
`requirements individually. However,
`in 1988,
`it was a
`major challenge to create a 16 kb/s coder that would meet
`all ofthese requirements simultaneously. Furthermore, the
`addition of the 5 ms low-delay requirement made such an
`attempt even more difficult.
`
`Because of the low-delay requirement. none of the 16
`kb/s coders mentioned above (CELP, MPLPC. APC,
`ATC, and SBOADPCM) could be used in their current
`form. With all these well—established coders ruled out. the
`
`only hope seemed to be backwai'd-adaptive predictive
`coders which derive their predictor coefficients from pre-
`viously quantized speech and thus do not need to bulfer a
`large frame of speech samples.
`(The G.72l ADPCM
`coder belongs to this category.)
`Prior to the CClTT’s recent standardization effort at 16
`
`kb/s, several researchers had previously reported their
`work on low-delay speech coding at 16 kb/s. Jayant and
`Ramamoorthy I13], [14] used an adaptive postfilter to en-
`hance 16 kb/s ADPCM speech and achieved a mean
`opinion score (MOS) [15] of 3.5 at nearly zero coding
`delay. Cox at at‘. [16] combined SBC and vector quanti-
`zation (VQ) and achieved an MOS of roughly 3.5—3.7 at
`a coding delay of 15 ms. Berouti er a.-’. [17] reduced the
`coding delay of an MPLPC coder to 2 ms by reducing the
`frame size to 1 ms. However.
`the speech quality was
`equivalent to 5.5—bit log PCM—a significant degradation.
`Taniguchi et at.
`[18] developed an ADPCM coder with
`mnltiquantizers where the best quantizer was selected once
`every 2.5 ms. The coding delay of their real-time proto-
`type codec was 8.3 ms. With the help of postfiltering, the
`coder produced speech quality “nearly equivalent
`to a
`7-bit it-law PCM" [18], but this was achieved with a non-
`standard 6.4 kHz sampling rate and a resulting nonstan-
`dard bandwidth of the speech signal. Gibson et at‘. [19],
`[20] studied backward—adaptive predictive tree coders and
`predictive trellis coders which should have low coding de-
`lays. Unfortunately, they did not report the exact coding
`delay or the subjective speech quality. Iyengar and Kabal
`121] also developed a backward-adaptive predictive tree
`coder with a 1 ms coding delay and a level of speech qual-
`ity equivalent to 7-bit log PCM. Watts and Cuperman {22]
`proposed a vector generalization of ADPCM with a delay
`between I and 1.5 ms. They did not report the subjective
`speech quality of the coder.
`Since 1988, when the CCITT announced its intention
`
`to standardize a 16 kb/s low-delay speech coder, there
`has been a great deal of research activity in the area of
`low-delay speech coding at 16 kb/s [23]—[38].
`In re-
`sponse to the CClTT’s standardization effort, we have
`created a 16 kb/s coder called low-delay CELP, or LD-
`CELP, which achieves high speech quality with a one-
`way coding delay less than 2 ms [24], [29], [32], [33],
`[35}, [38].
`
`The LD—CELP coder is a predictive coder that com-
`bines: 1) high-order backward—adaptive linear prediction;
`2) backward gain—adaptive vector quantization [39], [40]
`for excitation; 3)
`the analysis-by-synthesis excitation
`codebook search of CELP; and 4) adaptive postfiltering
`[I4], [41]. The low coding delay is achieved by using
`backward-adaptive prediction to avoid the long speech
`buffet‘ required by forward—adaptive prediction, and by
`using a small excitation vector size of only five samples,
`or 0.625 ms (assuming the standard 8 kHz sampling rate).
`
`
`
`332
`
`EEEIE JOL:'R|\l.I\|_ UN .‘$ELi-'.C']‘I€D ARIEAS IN CO.\-1MUNIC.-\'l"lONS.
`
`\"()i__ Iii. N0. 5. JUNE N93
`
`With the processing delay and transmission delay also in-
`cluded,
`the total one—way coding delay can be less than
`2 ins. This not only surpasses the CCITT delay require-
`ment of 5 ms but actually meets" the objective of 2 ms.
`This LD-CELP coder was submitted by AT&T to the
`CCITT and has been the only candidate coder since 1989.
`This coder has been implemented in real—time hardware
`using the AT&T WE“? DSP32C floating-point digital sig-
`nal processor, and the resulting ha1'dware prototype
`LD-CELP coder has been used_ in the official CCITT lab-
`oratory tests.
`
`In the standardization process, there were two phases
`of laboratory testing. The first phase of testing was con-
`ducted in late 1989 and early I990, while the second phase
`was in early I991. The LD-CELP cotter submitted for the
`first phase of testing (called the Phase 1 coder from here
`on) met all of the CCITT’s performance requirements ex-
`cept for the requirement of three asynchronous tandems.
`Based on the Phase 1
`test results,
`the Speech Quality
`Experts Group (SQEG) of the CCITT indicated that the
`LD-CELP coder could be standardized for point-to—point
`applications but not for networking applications where
`tandeming may occur, unless the coder could be improved
`to meet the tandeming performance requirement
`in the
`Phase 2 test.
`
`In late 1990 to early 1991, we improved the LD-CELP
`coder’s tandeming performance significantly and pro-
`duced what we called the Phase 2 cotter. The hardware
`
`prototype Phase 2 coder was then tested in the second
`phase of laboratory testing in 1991. From the Phase 2 test
`results, the SQEG concluded that the 16 kb/s LD-CELP
`coder “has a performance eqttittctfetrt to or better than
`G. 721,“ and “meets at! the speech qtrutity require.-ueurs
`set by Study Group XV and tested by Study Group XH ”
`[42]. Therefore,
`“the SQEG recommends
`that
`the
`I6 kb/s LD-CELP cortec can be stuttdarrtt'zed as o CCITT
`G Series Recottttttettdrttfott as t'egut'ds to its speech quot-
`tty” [42]. According to the current standardization sched
`ule, this 16 kb/s LD-CELP coder is expected to be stan-
`dardized by the first half of 1992.
`In this paper, we will describe the 16 kb/s LD-CELP
`coding algorithm,
`its
`implementation, and its perfor-
`mance. Section II
`introduces system concepts and pro-
`vides an overview ofthe LD-CELP coder. Section III de-
`
`scribes the LD-CELP coding algorithm. Section IV
`discusses the implementation issues. Section V describes
`the subjective and objective performance, and Section VI
`gives some concluding remarks.
`
`II. SYSTEM Cot~JCt'-;t='rs AND OVERVIEW
`
`In this section, we review the conventional CELP al-
`gorithm [1] and then give an overview of the LD-CELP
`algorithm and point out the differences between conven-
`tional CELP and LD-CELP. Along the way, we also dis-
`cuss the issue of coding delay.
`
`A. Review of Conventional CELP
`
`A typical example of the conventional CELP speech
`coder is shown in Fig. 1. The CELP coder is based on the
`
`“source-filter" speech production model [43], with the
`short—term synthesis filter modeling the vocal tract and the
`excitation VQ, together with the long-term synthesis fi].
`ter, modeling the glottai excitation. The CELP coder syn-
`thesizes speech by passing a gain-scaled excitation se- -
`quence through long-term and short-term synthesis filters,
`Both synthesis filters are all-pole filters containing either
`a long~term or a short-term predictor in a feedback loop.
`Basically.
`the CELP coder encodes speech frame—by-
`fralne, and within each frame it attempts to find the best
`predictors, gain, and excitation such that a perceptually
`weighted mean—squared error (MSE) between the input
`speech and the synthesized speech is minimized.
`The long-term predictor is often referred to as the pitch
`predictor, because its main function is to exploit the pitch
`periodicity in voiced speech. Typically, a one-tap pitch
`predictor is used,
`in which case the predictor transfer
`function is:
`
`P.(z) = 62 "’
`
`(1)
`
`where p is the bulk delay or pitch period, and B is the
`predictor tap. The short-term predictor is sometimes re-
`ferred to as the LPC predictor, because it is aiso used in
`the well—known LPC (linear predictive coding) vocoders
`which operate at 2.4 kb/s or below. The LPC predictor
`is typically a 10th-order predictor with a transfer function
`of:
`
`10
`
`P2(Z) =
`
`a.-zit
`
`(2)
`
`through am are the predictor coefficients. The
`where a.
`excitation VQ codebook contains a table of codebook vec-
`tors (or codevectors) of equal length. The codevectors are
`typically populated by Gaussian random numbers with
`possible center clipping.
`_
`In the actual encoding process, the encoder first buffers
`an input speech frame of about 20 ms or so, and then
`performs linear predictive analysis [43] (or LPC rutut'ysr's)
`on the buffered speech. The resulting LPC parameters are
`then quantized. The pitch predictor parameters, including
`the pitch period and the predictor tap, are then determined
`either in an open-loop fashion [1] or in a closed-loop fash-
`ion [44]. The quantized LPC parameters and pitch pre-
`dictor parameters are both sent as side information to the
`decoder. This scheme is called fot‘wut'd-adoptive preo't'c-
`tiou.
`
`The input speech frame is further subdivided into sev-
`eral equal-length subfrumes, or vectors, typically of size
`4 to 8 ms. Then, for each vector, the encoder passes each
`candidate codevector in the excitation VQ codebook
`through the gain scaling unit and the two synthesis filters,
`and then compares the corresponding filtered output vec-
`tor with the input speech vector and computes the asso-
`ciated perceptually weighted MSE. distortion. The en-
`coder repeats this process for all candidate excitation
`codevectors and then identifies the codevector that mini-
`
`mizes the perceptually weighted MSE distortion. This
`
`
`
`
`
`CHEN at air. A LOW-DELAY CEl.P LTODER
`
`333
`
`multiplex
`
`Encode
`and
`
`
`quantization
`
`LPC
`analysis &
`
`
`
`
`
`Long-term
`synthesis
`
`
`
`
`
`
`
`Output
`postfiltered
`
`\m_,:_..._/
`
`Long-term
`synthesis
`filter
`
`{bl
`
` j/
`Short-term
`synthesis
`filter
`
`Fig. 1.
`
`{:1} Typical conventional CELP encoder. {bl Corresponding CELP
`decoder.
`
`is
`process is called a closed-loop Search (sometimes it
`called an (maIysis—by-.rymItesis procedure). Vector quan-
`tization of the excitation using a closed-loop search is the
`main feature of CELP;
`it
`is also the main reason why
`CELP coders outperform other linear predictive coders
`such as APC and MPLPC.
`
`it is possible to jointly optimize the ex-
`Theoretically,
`citation codevector and the gain.
`In practice, however,
`almost all conventional CELP coders separately quantize
`the gain subsequent
`to the closed—loop excitation VQ
`codebook search. There are at least two reasons for this.
`First, performing such sequential optimization produces
`a much lower computational complexity than a joint op-
`timization approach. Second, with the gain typically
`quantized to 5 b or so, the resolution in gain quantization
`is high enough that the performance difference between
`the two approaches is essentially negligible.
`
`five ditiercnt kinds of infor-
`In conventional CELP.
`mation are encoded and sent to the decoder. These are:
`
`1) the LPC parameters; 2} the pitch period; 3) the pitch
`predictor tap; 4) the excitation gain; and 5) the excitation
`“shape“ codevectors. The decoder decodes such infor-
`mation and reproduces speech frame-by-frame by exciting
`the two cascaded synthesis filters with the seated excita-
`tion vector sequence. Usually, a postfilter of the type pro-
`posed in [41] is used in a CELP decoder to enhance the
`perceptual quality of decoded speech.
`
`B. Cridfrtg De.-'rt_v
`
`In the Introduction, we mentioned that a conventional
`
`CELP coder typically has a one-way coding delay of 50
`to 60 ms. The one-way coding delay is defined as the
`elapsed time from the instant a speech sample arrives at
`
`
`
`
`
`334
`
`IEEE JOURNAL ON SELECTED ARE.-XS IN CUMMUNlf.'.-\Tl0NS. VOL.
`
`ID. NO. 5. JUNE I993
`
`the encoder input to the instant when that same speech
`sample appears at the decoder output, less any delay added
`by other communication equipment (such as modems) in
`between the encoder—decode'r pair, and the signal propa-
`gation delay which depends on distance. In other words,
`it
`is as if the encoder and the decoder were connected
`
`back-to-back by wires at the same physical location with-
`out any equipment in between. This definition makes the
`coding delay dependent only on the coding algorithm. not
`on other equipment or communication distance.
`By such definition,
`the coding delay of CELP coders
`can be roughly determined in terms of the speech frame
`size used. The coding delay consists of three kinds of de-
`lay: 1) algorithmic buffering delay; 2) processing delay;
`and 3) bit transmission delay. These delays are explained
`below. First, due to the forward-adaptive LPC analysis,
`the CELP encoder first has to buffer one frame of speech
`samples before the encoding of the first sample in that
`frame can be started. Such buffering introduces at least
`one frame worth of buffering delay. Second, assume the
`real—time hardware is just fast enough to run the coder in
`real-time (which is usually the case). Then,
`it will take
`almost one frame worth of processing delay to perform
`the encoding and decoding of the buffered speech frame.
`Third, suppose the encoder does not start sending bits cor-
`responding to a given speech frame until the encoding of
`the entire frame is completed and the decoder does not
`start decoding a speech frame until all of the bits of that
`frame have been received, then one additional frame of
`
`bit transmission delay will be introduced. This is because
`it is one frame worth of time from the instant the first bit
`of the frame is sent to the instant the last bit of the frame
`
`is received, provided that the bit rate of the communica-
`tion channel is the same as the bit rate of the speech coder.
`Hence, the total one—way coding delay of a CELP coder
`is roughly three frames.
`Of course, the above delay analysis is oversimplified.
`In practice,
`it
`is possible to reduce the processing delay
`without using a faster processor. This can be achieved by
`sending bits out “on-the~fly“ at the encoder as soon as
`certain bits become available, and by decoding bits “on-
`the—fiy“ at the decoder as soon as the received bits are
`suflicient to start decoding the first speech sample in the
`frame. (Some constraints on bit timing have to be care-
`fully observed, of course.) Having a fast processor can
`further cut down the processing delay. The use of a faster
`communication channel (e.g., in multiplexed cornmuni-
`cation systems) can also reduce the bit transmission de-
`lay. On the other hand, for ease of implementation, some-
`times a designer Inay choose to have a coding delay longer
`than three frames. Therefore, the actual coding delay is
`likely to vary between two and four frames, depending on
`the actual implementation and application. If we take 2.5
`to 3 frames as the average, then a CELP coder with a 20
`ms frame size is likely to have a coding delay of 50 to 60
`ms. This long coding delay is mainly due to the large 20
`ms frame buffer that is required by the forward adaptation
`of-LPC predictor coeflicients.
`
`C. Ovet‘vt'eu-' of LD-CELP
`
`In LD-CELP, we reduce the coding delay by making
`the LPC predictor “backward-adaptive." (See [15, ch. 4,
`6] for a comprehensive discussion of forward and back-
`ward adaptation.) This means that the predictor coeffi-
`cients will not be derived from the input speech samples
`yet to be coded, but rather from previously quantized
`speech samples. Since previously quantized speech sam-
`ples are also available at the decoder, the decoder can also
`derive the predictor coefficients using the same procedure
`used in the encoder. Thus, there is no need to transmit
`
`any side information bits to specify the predictor cocffi—
`cients. More importantly, with backward adaptation, there
`is no need to buffer about 20 ms of input speech for for-
`ward—adaptive LPC analysis. Therefore,
`the excitation
`vector (subframe) becomes the basic bulfer unit, and the
`
`one—way coding delay becomes roughly three times the
`vector size (or vector dimension). Since the vector di-
`mension is much smaller than 20 ms, the coding delay is
`greatly reduced.
`From the very beginning, our goal was to bypass the 5
`ms delay requirement and to achieve the delay objective
`of2 ms or less. The following simple analysis shows that
`we also needed to make the excitation gain backward-
`adaptive in order to achieve this goal. With the 8 kHz
`standard sampling rate, a delay of 2 ms corresponds to 16
`samples. Since the coding delay is roughly three times the
`vector dimension, to achieve a one~way delay of 16 sam-
`ples or less, the largest vector dimension we can use is
`five samples (0.625 ms). With a vector dimension of five
`samples and a bit rate of two bits/sample. we only have
`10 bits to encode the excitation gain as well as the exci-
`tation shape codevector. The excitation gain typically re-
`quires four to five bits of resolution in conventional CELP.
`In our case, however, it would be highly inefficient to use
`40-50% of the total bit rate to encode the slowly varying
`and somewhat redundant gain term. To achieve better
`coding efficiency and speech quality under the 2 ms delay
`constraint, we use backw(ti'a’ gat'ri—aa'aptive vector quan-
`rizarioii [40] for the excitation. By making the excitation
`gain backward—adaptive, we derive the excitation gain
`from the gain information embedded in previously quan-
`tized excitation, and there is no need to send any bits to
`specify the excitation gain, since the decoder can derive
`the same gain in the same manner.
`A simplified block diagram of the 16 kb/s LD-CELP
`coder is shown in Fig. 2. In this coder, the excitation vec-
`tor has a dimension of five samples, so the one~way cod-
`ing delay is less than 2 ms. The pitch predictor in con—
`ventional CELP is eliminated, and a 50th—order LPC
`
`predictor is used. The LPC predictor coefficients are up-
`dated oncc every four speech vectors (2.5 ms) by pet‘-
`forming LPC analysis on previously quantized speech.
`The excitation gain is updated once every vector by using
`a l0th—0rder adaptive linear predictor in the logarithmic
`gain domain. The coeflicients of this log-gain predictor
`are updated once every four vectors by performing linear
`
`
`
`
`
`CHBN er (IL: A LOW-DELAY CELP CODER
`
`335
`
`Excitation
`VQ
`oodebook
`
`
` Perceptual
`
`
`weighting
`
`filter
`
`Output
`postfiltered
`
`Adaptive
`postfilter
`
`VQ index
`from
`
`Excitation
`VQ
`codebook
`
`Backward
`
`
`
`
`
`
`gain
`adaptation
`
`
`
`Fig. 2.
`
`(a) 16 kb/s |ow—delay Ci-ELP encoder. (b) 16 lib/s low-delay CELP
`decoder.
`
`(b)
`
`predictive analysis on the logarithmic gains of previously
`quantized and scaled excitation vectors. The 10th-order
`perceptual weighting filter is also updated once every four
`vectors, but it is updated by performing an LPC analysis
`on the input speech. To reduce the codebook search com-
`plexity, the 10-bit excitation VQ codebook is a product
`code