Interspeech'2005 - Eurospeech
A method for efficient quantization of speech spectral models is proposed. The impulse response of the model is coded in two spectral subbands of unequal width. The distortion measure used is a squared error between impulse responses. The quality obtained at 22 bits per frame is as good as that of Split Vector Quantization of Line Spectral Frequencies at 24 bits per frame.
Bibliographic reference. Erkelens, Jan (2005): "High-quality memoryless subband coding of impulse responses at 22 bits per frame", In INTERSPEECH-2005, 2737-2740.