Compressed domain speech enhancement based on the joint modification of codebook gains.
Bingyin XiaChangchun BaoYan LiangXuan ZhouYu-wen HeRu-wei LiPublished in: ISSPIT (2011)
Keyphrases
- compressed domain
- speech enhancement
- noise reduction
- noisy environments
- motion vectors
- signal to noise ratio
- video analysis
- single channel
- bitstream
- inter frame
- vocal tract
- linear prediction
- speech signal
- smoothing algorithm
- image representation
- wiener filter
- multi channel
- macroblock
- motion estimation
- image features
- background noise
- motion compensation
- speech recognition
- bit rate
- edge detection