In this paper, we present an improved harmonic coder for the wideband. In an effort to improve spectrum modelling with harmonic representation, we include in the structure, a post error spectrum modelling that is essentially useful for noisy signals or mixed spectra of high pitch voices. The bit constrained error spectrum modelling procedure coupled with an efficient quantization of the harmonic model parameters allow us to develop a low fixed bit rate coder. The detailed structure of all the analysis, synthesis and quantization steps are outlined for the developed 16kbs coder on the [50-7000 Hz] bandwidth. Informal listening tests show that the quality obtained is comparable to the ITU_T Recommendation G722 at 48 kbs
Cite as: Amodio, A., Feng, G. (1999) A wideband speech coder based on harmonic coding at 16KBS. Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999), 1539-1542, doi: 10.21437/Eurospeech.1999-346
@inproceedings{amodio99_eurospeech,
author={A. Amodio and Gang Feng},
title={{A wideband speech coder based on harmonic coding at 16KBS}},
year=1999,
booktitle={Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999)},
pages={1539--1542},
doi={10.21437/Eurospeech.1999-346}
}