Exploring the mutual intelligibility breakdown caused by sculpting speech from a competing speech signal.
Martin CookeMaría Luisa García LecumberriPublished in: INTERSPEECH (2023)
Keyphrases
- speech signal
- speech recognition
- automatic speech recognition
- speaker identification
- vocal tract
- automatic speech recognition systems
- background noise
- noisy environments
- linear prediction
- hidden markov models
- speech synthesis
- non stationary
- spectral analysis
- speech enhancement
- language model
- speaker recognition
- pattern recognition
- speech sounds
- noisy speech
- speech quality
- spectral subtraction
- wavelet transform
- speaker diarization
- acoustic features
- probabilistic neural network
- machine learning
- sound source