Abstract | ||
---|---|---|
All-pole spectral envelope estimates based on linear prediction (LP) for speech signals often exhibit unnaturally sharp peaks, especially for high-pitch speakers. In this paper, regularization is used to penalize rapid changes in the spectral envelope, which improves the spectral envelope estimate. Based on extensive experimental evidence, we conclude that regularized linear prediction outperforms bandwidth-expanded linear prediction. The regularization approach gives lower spectral distortion on average, and fewer outliers, while maintaining a very low computational complexity. |
Year | DOI | Venue |
---|---|---|
2008 | 10.1109/TASL.2007.909448 | Audio, Speech, and Language Processing, IEEE Transactions |
Keywords | Field | DocType |
prediction theory,spectral analysis,speech processing,all-pole spectral envelope estimation,regularized linear prediction,speech signal,Bandwidth expansion,envelope estimation,linear prediction (LP),regularization | Speech processing,Speech coding,Spectral envelope,Pattern recognition,Outlier,Speech recognition,Linear prediction,Bandwidth (signal processing),Regularization (mathematics),Artificial intelligence,Mathematics,Computational complexity theory | Journal |
Volume | Issue | ISSN |
16 | 1 | 1558-7916 |
Citations | PageRank | References |
20 | 1.13 | 7 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ekman, L.A. | 1 | 20 | 1.13 |
W. Bastiaan Kleijn | 2 | 1110 | 106.92 |
Murthi, M.N. | 3 | 70 | 8.27 |