Voice Interactive Systems in Severe Noise Conditions,
Abstract
In a comparison of word recognition performance between human listeners and automatic speech recognition systems (ASR), the human listeners performs much better, especially in severe noise conditions. An application engineer can try to optimize the performance of an ASR system by selecting the optimal noise cancelling microphone and vocabulary for voice input. Some results from a study on the effect of signal handling and vocabulary configuration on the performance of voice input and voice output systems will be discussed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1987
- Accession Number
- ADP005603
Entities
People
- G. Langhout
- H. J. Steeneken