Cepstral normalisation and the signal to noise ratio spectrum in automatic speech recognition.
Philip N. GarnerPublished in: Speech Commun. (2011)
Keyphrases
- automatic speech recognition
- speech recognition
- signal to noise ratio
- spectral subtraction
- speech signal
- hidden markov models
- linear prediction
- noise reduction
- pattern recognition
- word error rate
- noisy environments
- speech retrieval
- conversational speech
- language model
- recognition errors
- minimum mean square error
- speech synthesis
- broadcast news
- speech corpus
- spoken words
- handwriting recognition
- matched filter
- spontaneous speech
- edge detection
- acoustic features
- word recognition
- speech enhancement
- bit error rate
- received signal
- gaussian noise
- cross correlation
- image processing
- neural network