... from the lower panel that the companding architecture is able to follow harmonic and formant transitions with EURASIP Journal on Audio, Speech, and Music Processing Speech Hz Companding off 8000 ... models and human performance in tasks related to speech coding and speech recognition,” IEEE Transactions on Speech and Audio Processing, vol 2, no 1, part 2, pp 115–132, 1994 [5] A Van Schaik and ... and 16.3% in total error Even for the multicondition recognizer, companding is observed to result in significant improvements in recognition performance for most noise types For example, for speech...