4th International Conference on Spoken Language Processing
Philadelphia, PA, USA
In speech recognition, a discriminative quefrency weighting can be achieved by somewhat decorrelating the frequency sequence of log mel-scaled filter-bank energies with a computationally inexpensive filter. In this paper, we show how the spectral parameters that result from this kind of frequency filtering, both alone and combined with filtering of their time trajectories, are competitive with respect to the conventional cepstral representations of speech signals.
Bibliographic reference. Nadeu, Climent / Mariño, José B. / Hernando, Javier / Nogueiras, Albino (1996): "Frequency and time filtering of filter-bank energies for HMM speech recognition", In ICSLP-1996, 430-433.