Theoretical analysis of musical noise in generalized spectral subtraction: Why should not use power/amplitude subtraction?
Takayuki InoueYu TakahashiHiroshi SaruwatariKiyohiro ShikanoKazunobu KondoPublished in: EUSIPCO (2010)
Keyphrases
- spectral subtraction
- theoretical analysis
- speech recognition
- hybrid method
- speech signal
- signal to noise ratio
- noise model
- automatic speech recognition
- hidden markov models
- noisy speech
- power spectrum
- background noise
- additive noise
- noise reduction
- noise level
- evolutionary algorithm
- pattern recognition
- image processing
- noise removal
- machine learning