{"title":"一种改进的4kbit /s CELP语音编码算法","authors":"Yanning Bai, C. Bao","doi":"10.1109/CHINSL.2004.1409609","DOIUrl":null,"url":null,"abstract":"The paper presents a 4 kbit/s CELP speech coder that utilizes the nonuniform and part-searching-area algebraic codebook technologies to overcome the insufficient number of signed pulses in a fixed codebook (FCB). The nonuniform algebraic codebook is based on the nonuniform statistical properties of the FCB. The part-searching-area utilizes the periodicity of the FCB excitation signal at low bit rates. The latter is only employed when the pitch delay is small enough. We also find that preserving the continuity of pitch is very important for voiced segments if these two technologies are used. So different pitch-detection methods are employed for voiced/unvoiced frames. Subjective and objective test results indicate that the qualities of reconstructed speech are improved, especially for female speakers.","PeriodicalId":212562,"journal":{"name":"2004 International Symposium on Chinese Spoken Language Processing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2004-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"An improved 4 kbit/s CELP speech coding algorithm\",\"authors\":\"Yanning Bai, C. Bao\",\"doi\":\"10.1109/CHINSL.2004.1409609\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The paper presents a 4 kbit/s CELP speech coder that utilizes the nonuniform and part-searching-area algebraic codebook technologies to overcome the insufficient number of signed pulses in a fixed codebook (FCB). The nonuniform algebraic codebook is based on the nonuniform statistical properties of the FCB. The part-searching-area utilizes the periodicity of the FCB excitation signal at low bit rates. The latter is only employed when the pitch delay is small enough. We also find that preserving the continuity of pitch is very important for voiced segments if these two technologies are used. So different pitch-detection methods are employed for voiced/unvoiced frames. Subjective and objective test results indicate that the qualities of reconstructed speech are improved, especially for female speakers.\",\"PeriodicalId\":212562,\"journal\":{\"name\":\"2004 International Symposium on Chinese Spoken Language Processing\",\"volume\":\"1 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2004-12-15\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2004 International Symposium on Chinese Spoken Language Processing\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/CHINSL.2004.1409609\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2004 International Symposium on Chinese Spoken Language Processing","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/CHINSL.2004.1409609","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
The paper presents a 4 kbit/s CELP speech coder that utilizes the nonuniform and part-searching-area algebraic codebook technologies to overcome the insufficient number of signed pulses in a fixed codebook (FCB). The nonuniform algebraic codebook is based on the nonuniform statistical properties of the FCB. The part-searching-area utilizes the periodicity of the FCB excitation signal at low bit rates. The latter is only employed when the pitch delay is small enough. We also find that preserving the continuity of pitch is very important for voiced segments if these two technologies are used. So different pitch-detection methods are employed for voiced/unvoiced frames. Subjective and objective test results indicate that the qualities of reconstructed speech are improved, especially for female speakers.