Multi-stream and multi-band methods can improve the accuracy of speech recognition systems without overly increasing the complexity. However, they cannot be applied blindly. In this paper, we review our experience applying multi-stream and multi-band methods to the Broadcast News corpus. We found that multi-stream systems using different acoustic front-ends provide a significant improvement over single stream systems. However, despite the fact that they have been successful on smaller tasks, we have not yet been able to show any improvement using multi-band methods. We report various insights gained from the experience in applying these methods in a large-vocabulary task.
Cite as: Janin, A., Ellis, D., Morgan, N. (1999) Multi-stream speech recognition: ready for prime time? Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999), 591-594, doi: 10.21437/Eurospeech.1999-152
@inproceedings{janin99_eurospeech,
author={Adam Janin and Dan Ellis and Nelson Morgan},
title={{Multi-stream speech recognition: ready for prime time?}},
year=1999,
booktitle={Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999)},
pages={591--594},
doi={10.21437/Eurospeech.1999-152}
}