Correction of the voice timbre distortions in telephone networks: method and evaluation

摘要

In a telephone link, the voice timbre is impaired by spectral distortions generated by the analog parts of the link. Our purpose is to restore a timbre as close as possible to that of the original voice of the speaker, using a blind equalizer centralized in the network, which compensates for the spectral distortions. We propose a spectral equalization algorithm, which consists in matching the long-term spectrum of the processed signal to a reference spectrum within a limited frequency bandwidth (200-3150 Hz). Subjective evaluations show a satisfactory restoration of the timbre of the speakers, within the limits of the chosen equalization band. The A-law quantization of the output samples of the equalizer induces however a disturbing noise at the reception end. A subjective evaluation shows that speakers' voices with corrected timbre, even with quantization noise, are preferred to the same voices at the output of a link without timbre correction (and without noise). In order to make the reference spectrum more appropriate to the various speakers' voices, we classify them according to their long-term spectra and use a specific reference spectrum for each class. This leads to a decrease of the spectral distortion induced by the equalizer, significantly perceived as an improvement of the timbre correction, as a subjective test shows.