Changes in the speech spectrum produced by compression algorithms may be a factor in lowering speech quality. We measured the sensitivity of human listeners to changes in the spectral characteristics of simple speech signals, and suggest that coded-speech quality could be improved by allocating distortion to regions of the speech signal where the change is not easily perceived.
|Original language||English (US)|
|Number of pages||4|
|Journal||Proceedings of the National Electronics Conference|
|Publication status||Published - Dec 1 1986|
ASJC Scopus subject areas