R. Salami, C. Laftamme, J. Adoul, A. Kataoba, S. Hayashi, C. Lamblin, D. Massaloux, S. Proust, P. Kroon, Y. Shoham
{"title":"Description Of The Proposed ITU-T 8 Kb/S Speech Coding Standard","authors":"R. Salami, C. Laftamme, J. Adoul, A. Kataoba, S. Hayashi, C. Lamblin, D. Massaloux, S. Proust, P. Kroon, Y. Shoham","doi":"10.1109/SCFT.1995.658098","DOIUrl":"https://doi.org/10.1109/SCFT.1995.658098","url":null,"abstract":"","PeriodicalId":395024,"journal":{"name":"Proceedings. IEEE Workshop on Speech Coding for Telecommunications","volume":"553 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115732907","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Low Complexity Decomposition and Coding of Prototype Waveforms","authors":"L. S. Burnett, G. J. Bradley","doi":"10.1109/SCFT.1995.658108","DOIUrl":"https://doi.org/10.1109/SCFT.1995.658108","url":null,"abstract":"It has recently been shwn [ I ] that high-quality, lowrate Prototype Wavefom coding can be achieved using waveform interpolation. l3e high rate of quantised protorypes necessary is facilitated by prototype &composi#wn to a Slowly Evolving Waveform (SEW) and Rapidly Evolving Waveform (REW). However, the decomposition is inherently open-loop and the Fourier coefficient filtering required sign@cantly increases the compkxiry of Protorype Wavefonn coding. Titis paper addresses these problems by use of a closed-loop prototype quantisation technique, which does not require explicit REW and SEW computation. Results show that similar speech quality can be generated by the closedloop scheme, but at a significantly reduced computational complexity.","PeriodicalId":395024,"journal":{"name":"Proceedings. IEEE Workshop on Speech Coding for Telecommunications","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127045053","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Speech Predictor Design for Diversity Communication Systems","authors":"V. Vaishampayan, A. A. Siddiqui","doi":"10.1109/SCFT.1995.658131","DOIUrl":"https://doi.org/10.1109/SCFT.1995.658131","url":null,"abstract":"The problem of speech predictor design is addressed for systems that use diversity to overcome channel impairments. A specific design approach is presented along with some performance results.","PeriodicalId":395024,"journal":{"name":"Proceedings. IEEE Workshop on Speech Coding for Telecommunications","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122288023","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Variable Bitrate Wideband Speech Coding Using Perceptually Motivated Thresholds","authors":"J. Paulus","doi":"10.1109/SCFT.1995.658114","DOIUrl":"https://doi.org/10.1109/SCFT.1995.658114","url":null,"abstract":"This paper proposes a variable bitrate wideband CELP coding scheme (50-7000 Hz) with a one-way coder-decoder delay of 25 ms only. The bitrate varies between 6.4 kbit/s and 14.8 kbit/s with a mean of 11.2 kbit/s at a voice activity of 95%. The variation of the bitrate is based on a voiced-unvoiced-silence classification of the speech frame to be encoded. For voiced frames a perceptually based focussed LTP analysis is used, and f o r unvoiced or silent f m m e s the LTP filter is omitted. Additionally, in adjacent frames with small spectral changes LPC parameters are repeated using an adaptive LPC codebook. For the classification of the dijferent modes a cepstral distance measure, and a new adaptive detector is used. By informal listening tests the speech quality was rated higher than the CCITT G.722 wideband ccdec operating at 48 kbit/s.","PeriodicalId":395024,"journal":{"name":"Proceedings. IEEE Workshop on Speech Coding for Telecommunications","volume":"111 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122785670","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Low-complexity Wideband Speech Coding","authors":"Juin Chen","doi":"10.1109/SCFT.1995.658110","DOIUrl":"https://doi.org/10.1109/SCFT.1995.658110","url":null,"abstract":"Wideband speech coding has been gaining popularity in recent years. Due to its higher sampling rate, wideband speech inherently requires higher processing power than telephone-bandwidth speech when the same coding algorithm is used. However, today many wideband coding applications demand a complexity that is even lower than that of most state-of-the-art telephone-bandwidth coders. To meet this complexity challenge, we created two wideband coders that are fundamentally different from and simpler than popular narrowband coders. The first one is a transform coder based on the modified discrete cosine transform (MDCT), LPC spectral fit, and pitch harmonic fit. The second algorithm is called UTransform Predictive Coding”, or TPC. It uses short-term and long-term prediction to remove the redundancy in speech. The prediction residual is quantized in the frequency domain based on a calculated noise masking threshold. In its current form, the TPC coder uses only open-loop quantization and therefore has a low complexity. We estimate that its complexity is only about half of that of the ITU-T 16 kb/s G.728 LD-CELP narrowband coder. The speech quality of TPC is almost transparent at 32 kb/s, very good at 24 kb/s, and acceptable at 16 kb/s. Work is in progress to further improve the speech quality at 16 kb/s.","PeriodicalId":395024,"journal":{"name":"Proceedings. IEEE Workshop on Speech Coding for Telecommunications","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131058526","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"2.4KBPS M-LCELP Speech Coding With Psychoacoustic Measure","authors":"J. Takizawa, K. Ozawa","doi":"10.1109/SCFT.1995.658151","DOIUrl":"https://doi.org/10.1109/SCFT.1995.658151","url":null,"abstract":"This paper proposes 2.4 Ebps speech coder based on the M-LCELP (Multi-mode Learned Gode Excited LpC) technique and .psychoacoustic measure, which has been developed for private wireless network systems. These systems require both lover bit rate and higher robustness against severe noisy channel. The subjective quality fiom the informal listening test shows that speech quality for 2.4Ebps M-LCELP coder with the psychoacoustic memrre slightly exceeds that of 2.4Ebps MLCELP without the psychoacoustic measure. Speech quality of both 2-4Ebps M-LCELP coders i s higher than that of the 4.8kbps US. Federal Standard 1016 CELP coder under the error free condition.","PeriodicalId":395024,"journal":{"name":"Proceedings. IEEE Workshop on Speech Coding for Telecommunications","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114783667","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A Harmonic Coder with a Direct Coding and Decoding of Bark Spectra","authors":"H. Tasaki, T. Yamaura, S. Takahashi","doi":"10.1109/SCFT.1995.658133","DOIUrl":"https://doi.org/10.1109/SCFT.1995.658133","url":null,"abstract":"In this paper, we propose a new harmonic coder that directly encodes and decodes the bark spectra (BS). In the decoder, the simultaneous equations of t rans fomat ion from BS to harmonic amplitudes have n o determinable solution because order of BS differs from that of harmonic amplitudes. To solve them, we introduce some assumptions about relations between harmonic amplitudes and FFT spectra. Using these assumptions, approximate solutions of harmonic amplitudes are obtained by simple matr i z operations. Experimental results show that the proposed coder has better subjective quulity than the harmonic coder using LSP.","PeriodicalId":395024,"journal":{"name":"Proceedings. IEEE Workshop on Speech Coding for Telecommunications","volume":"175 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115050694","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A variable-rate multipulse-deconvolution codec for wideband speech","authors":"V. Abreu-Sernández","doi":"10.1109/SCFT.1995.658112","DOIUrl":"https://doi.org/10.1109/SCFT.1995.658112","url":null,"abstract":"The wideband speech coder proposed in this paper is based on a MultiPulse scheme developed in our Department [l] in 1992. The MP coder is a full-band scheme which makes forward linear prediction analysis every frame of 16 ms. The bit rate for this coder is able to swatch, every 2 msec, between five possible rates from 16 kbps to 33 kbps. The quality measures show a lower objective quality for the MP coder when wmparing with the standard but a similar subjective quality. 1 Wideband Speech Coding Wideband speech coding (50-7000 Hz) is required for the emerging high-quality communications. The channel capacity of these new teleservices allows to spend more than 16 kbps for the speech codec without the bandwith restriction of the conventional telephony. The reference coder for the wideband speech signals is the ITU-T standard G.722, a two-bands ADPCM coder with three possible bit rates, 64 kbps, 56 kbps and 48 kbps [2]. The current activities are concentrated on coding at 32 kbps without decreasing the quality of the standard at 64 kbps and with the goal of wideband speech coding at 16 kbps [3]. Few schemes for wideband speech coding have been proposed until now, most of them based on the CELP algorithm [4, 51. The main drawbacks of a traditional CELP, working with wideband speech, are the huge computational complexity and the memory requirements to store the codebooks. Our proposed MultiPulse coder makes use of a spiky deconvolution algorithm to obtain the sparse exOThis work has been partially funded by Northern Telecom and by the Spanish Research National Plan under grant no. TIC92-0800-C05-02. citation. The main characteristics of this coder are the low computational complexity and the flexibility to operate at a wide range of bit rates, by simply selecting the suitable pulse rate. For telephonic speech, we get a Variable Rate version with six different operating rates from 9.1 kbps to 4.8 kbps, offering a high to toll speech quality [6, 71.","PeriodicalId":395024,"journal":{"name":"Proceedings. IEEE Workshop on Speech Coding for Telecommunications","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128960377","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Synthesis Methods In Sinusoidal And Waveform-interpolation Coders","authors":"D. Sen, W. Kleijn","doi":"10.1109/SCFT.1995.658136","DOIUrl":"https://doi.org/10.1109/SCFT.1995.658136","url":null,"abstract":"Waveform-interpolation (WI) coders can use various r e p resentations for the characteristic waveforms. For the commonly used Fourier-series representation, the parameterization of a WI coder is similar to that of cextain sinusoidal coders (the harmonic coders). We briefly mention the basic differences in the analysis between these procedures and then focus on the differences in the frequency-domain synthesis methods commonly associated with these two types of coders. Each synthesis method can be used with either analysis method. The method commonly associated with WI coders results in a more regnlar Structiire, allowing a fast recursive procedure to be used. This method also allows for a decoupling of the interpolation d the formant structure and the fundamental frequency, which is desirable when interpolating over longer intervals.","PeriodicalId":395024,"journal":{"name":"Proceedings. IEEE Workshop on Speech Coding for Telecommunications","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129049993","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}