Experiments in the Automatic Segmentation of Continuous Speech
Automatic segmentation of continuous speech is an important operation of the acoustic classifier of a speech understanding system.
The paper presents the results of a segmentation algorithm based on the linguistic description of global spectral features and suprasegmentals of speech.
A syntax-controlled analysis of these descriptions permits the detection of segments of continuous speech where co-articulation affects the pronunciation of several phonemes.
The paper presents the results of a segmentation algorithm based on the linguistic description of global spectral features and suprasegmentals of speech.
A syntax-controlled analysis of these descriptions permits the detection of segments of continuous speech where co-articulation affects the pronunciation of several phonemes.
Document Type: Research Article
Publication date: 01 January 1976
- Access Key
- Free content
- Partial Free content
- New content
- Open access content
- Partial Open access content
- Subscribed content
- Partial Subscribed content
- Free trial content