Method and apparatus for speech detection of PCM multiplexed voice channels
AssigneeUniversite de Sherbrooke
InventorsJean-Pierre Adoul, Fouad Daaboul
Life of this patent
Estimating the term…
Abstract
The disclosure herein describes a method and a system for speech detection on PCM multiplexed voice channels; for each channel, a decision is reached every M samples regarding the channel activity; in addition, the nature of speech is detected as: voiced (compact or non-compact) or unvoiced (fricative or non-fricative) when the channel is active; pure silence, white noise or echo when the channel is inactive. The decision is based on the joint value of the amplitude, zero crossing of the signal and zero crossing of the signal derivative.
Claims12
A speech detector for use in a PCM multiplexed voice channel system comprising:
- means for processing a predetermined batch of consecutive PCM samples;
- means for sequentially computing a series of parameters during processing of said predetermined batch of consecutive PCM samples, said parameters consisting of a function of the amplitude of the vocal signal, the zero crossing of the voice signal, the zero crossing of the derivative of the vocal signal;
- and means for determining the status of each channel from information received as a result of the computing of said parameters over said predetermined batch.
A method of speech detection in a PCM multiplexed voice channel system, comprising:
- processing a predetermined batch of consecutive PCM samples;
- sequentially computing a series of parameters during processing of said predetermined batch of consecutive PCM samples, said parameters consisting of a function of the amplitude of the vocal signal, the zero crossing of the vocal signal, the zero crossing of the derivative of the vocal signal;
- and determining the status of said channel from information received as a result of the computing of said parameters over said predetermined batch.
Dependencies are inferred from the claim text.