Frequency and time filtering of filter-bank energies for HMM speech recognition

Climent Nadeu, José B. Mariño, Javier Hernando, Albino Nogueiras

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.

doi: 10.21437/ICSLP.1996-91

Cite as: Nadeu, C., Mariño, J.B., Hernando, J., Nogueiras, A. (1996) Frequency and time filtering of filter-bank energies for HMM speech recognition. Proc. 4th International Conference on Spoken Language Processing (ICSLP 1996), 430-433, doi: 10.21437/ICSLP.1996-91

@inproceedings{nadeu96_icslp,
  author={Climent Nadeu and José B. Mariño and Javier Hernando and Albino Nogueiras},
  title={{Frequency and time filtering of filter-bank energies for HMM speech recognition}},
  year=1996,
  booktitle={Proc. 4th International Conference on Spoken Language Processing (ICSLP 1996)},
  pages={430--433},
  doi={10.21437/ICSLP.1996-91}
}