An Investigation of Fundamental Frequency Pattern Prediction for Japanese Electrolaryngeal Speech Enhancement Based on Frame-Wise Phoneme Representations.
Mohammad EshghiTomoki TodaPublished in: IEEE Access (2024)
Keyphrases
- speech signal
- speech enhancement
- fundamental frequency
- speech recognition
- linear prediction
- automatic speech recognition
- noisy environments
- hidden markov models
- vocal tract
- speech synthesis
- noise reduction
- background noise
- noisy speech
- prediction error
- speaker identification
- single channel
- pairwise
- signal to noise ratio
- language model
- sound source
- non stationary
- natural images
- computational complexity
- pattern recognition