Intelligibility improvements using binaural diverse sub-band processing applied to speech corrupted with automobile noise

The paper reports on experiments assessing the capability of a diverse processing, multi-microphone sub-band adaptive signal processing scheme for improving the intelligibility of speech corrupted with automobile noise. Results from formal listening tests demonstrate a significant improvement in the intelligibility and quality of the processed speech. Spoken digits corrupted with automobile noise at a low and a high signal-to-noise ratio were used with two commercial speech recognisers. The results obtained with the recognisers did not demonstrate any statistically significant improvement due to processing in sub-bands.