Pronunciation recognition of English phonemes /\textipa{@}/, /æ/, /\textipa{A}: / and /\textipa{2}/ using Formants and Mel Frequency Cepstral Coefficients.
Keith Y. PatarroyoVladimir Vargas-CalderónPublished in: CoRR (2017)
Keyphrases
- speech signal
- mel frequency cepstral coefficients
- speech recognition
- cepstral coefficients
- speaker identification
- automatic speech recognition
- speaker independent
- language identification
- linear predictive
- vocal tract
- speech recognition systems
- hidden markov models
- speech synthesis
- acoustic features
- noisy environments
- broadcast news
- speaker recognition
- linear prediction
- language model
- pattern recognition
- spectral analysis
- speaker diarization
- natural language
- machine translation
- text to speech
- sound source
- gaussian mixture model
- multi modal
- low level
- machine learning