DCT_M model for excitation parameter in low bit rate vocoder

Abstract

The description precision of an excitation signal greatly influences the quality of reconstructed speech in low bit rate vocoders. To improve the reconstruction quality, the DCT_M model is proposed to express the excitation spectral parameter, which transforms the variable length vector to fixed dimension vector through DCT transformation. It then quantizes the fixed length vector using multi-stage vector quantization. Tests show that the proposed method can keep the shape of the entire spectral envelope and reduce model error thus greatly improve the description precision. Test results in the sine excitation linear prediction (SELP) vocoder show that the DCT_M model can improve the naturalness of reconstructed speech, with subjective test score of 65%.

References

1. Li J L Researchon Low Bit Rate Speech Coding AlgorithmDissertation for the Master DegreeBeijingTsinghua University 2004 (in Chinese)
2. He T H Zhang J W Cui H J et al.A model for describing the excitation parametersof vocoderAudio Engineering 2003 (4)5255(in Chinese)
3. Li C Lupini P Shlomot E et al.Coding of variable dimension speech spectral vectorsusing weighted nonsquare transform vector quantizationIEEE Transactions on Speech and Audio Processing 2001 9(6)622631. doi:10.1109/89.943340
4. Nishiguchi M Matsumoto J Wakasuki R et al.Vector quantized MBE with simplified V/UV divisionat 3.0 kbit/sIn: Proceedings of IEEE InternationalConference on Acoustics, Speech, and Signal Processing 1993 2151154. doi:10.1109/ICASSP.1993.319255
5. Das A Rao A V Gersho A Variable dimension vector quantization of speech spectrafor low rate vocodersIn: Proceedings ofthe Data Compression Conference 1994 420429