Ultra Low Bit Rate Speech Coding Using an Ergodic Hidden Markov Model.
Matthew LeeAdriane Swaim DureyElliot MooreMark ClementsPublished in: ICASSP (1) (2005)
Keyphrases
- hidden markov models
- low bit rate
- coding method
- speech recognition
- shape coding
- adaptive quantization
- speech signal
- phoneme recognition
- automatic speech recognition
- bit rate
- coding scheme
- image compression
- noisy speech
- multi stream
- video compression
- linear prediction
- image quality
- visual quality
- high speed
- inter frame
- visual speech
- compression ratio
- rate distortion
- viterbi algorithm
- compression algorithm
- video conferencing
- speaker dependent
- multimedia
- video quality
- motion compensation
- noisy environments
- higher order
- markov chain
- speech enhancement
- multiresolution
- probabilistic model
- multi modal
- speaker identification
- subband
- video coding
- peak signal to noise ratio
- bitstream