An investigation of cepstral parameterisations for large vocabulary speech recognition

We examined variants of MFCC and PLP cepstral parameterisations in the context of large vocabulary continuous speech recognition under di erent acoustical environmental conditions: Compared to MFCC, mel-frequency PLP uses a cubic root intensity-toloudness law, and an LPC analysis is applied to the mel-warped spectrum. In LPC-smoothed MFCC, the only di erence to MFCC is the additional LPC smoothing of the warped spectrum. While neither technique was able to signi cantly outperform the MFCC parameterisation in our setup which includes an LDA feature transformation, feature set combination via DMC at the acoustic likelihood level and via ROVER at the recognized word level delivered small but consistent improvements.