5th International Conference on Spoken Language Processing
We examine the distinctive feature [voice] that separates the voiced from the unvoiced sounds for the case of stop consonants. We conduct acoustic phonetic analyses on a large database and demonstrate the superior separability using a temporal measure (voice onset time; VOT) rather than spectral measures. We describe several algorithms to automatically estimate the VOT from continuous speech and compare them on a speech recognition problem to reduce error rates by as much as 53 percent over a baseline HMM based system.
Bibliographic reference. Ramesh, Padma / Niyogi, Partha (1998): "The voicing feature for stop consonants: acoustic phonetic analyses and automatic speech recognition experiments", In ICSLP-1998, paper 0881.