Second International Conference on Spoken Language Processing (ICSLP'92)

Banff, Alberta, Canada
October 13-16, 1992

Statistical Recovery of Wideband Speech from Narrowband Speech

Yan Ming Cheng (1,2), Douglas O'Shaughnessy (1), Paul Mermelstein (1,2)

(1) INRS-Telecommunications; (2) Bell-Northern Research, Quebec, Canada

We present an algorithm to generate wideband speech from a narrowband version of the same. The main body of the algorithm is a Statistical Recovery Function (SRF), which predicts the highband spectrum based on the narrowband spectrum. Assuming that bandpass portions of the speech are generated completely by a fixed number of random sources, the SRF explores the dependency among the random sources. The performance of the algorithm has been measured both in terms of spectral distortion and spectral signal-to-noise ratio (SNR). We obtained a 3 dB gain in SNR for the reconstructed wideband speech as compared to the narrowband speech. Informal perceptual experiments indicate a significant preference for the reconstructed speech.

Full Paper

Bibliographic reference.  Cheng, Yan Ming / O'Shaughnessy, Douglas / Mermelstein, Paul (1992): "Statistical recovery of wideband speech from narrowband speech", In ICSLP-1992, 1577-1580.