 |  |
Abstract: We introduce a new speaker independent method for reducing wind noise in single-channel recordings of noisy speech. The method is based on non-negative sparse coding and relies on a wind noise dictionary which is estimated from an isolated noise recording. We estimate the parameters of the model and discuss their sensitivity. We then compare the algorithm with the classical spectral subtraction method and the Qualcomm-ICSI-OGI noise reduction method. We optimize the sound quality in terms of signal-to-noise ratio and provide results on a noisy speech recognition task.
Demonstration: Here are two examples of the results of the proposed method. For comparison, also the result of the Qualcomm-ICSI-OGI method is provided. The proposed method clearly does a good job at removing the sudden gusts of wind compared with the Qualcomm-ICSI-OGI; however, there is a little more distortion in the speech.
| Example | Noisy Speech | Qualcomm-ICSI-OGI | Proposed Method |
|---|---|---|---|
| 1 |
|
|
|
| 2 |
|
|
|
- See also:
Reduction of Non-stationary Noise using a Non-negative Latent Variable Decomposition- Wind Noise Reduction using Sparse Non-negative Coding
- Files:
imm5258.pdf
wnr_jl.ppt
- Links:
- Cite:
- Mikkel N. Schmidt, Jan Larsen and Fu-Tien Hsiao, Wind Noise Reduction using Non-negative Sparse Coding, Machine Learning for Signal Processing, IEEE Workshop on (MLSP), 2007
- BibTeX:
- @inproceedings{schmidt07mlsp,
title = "Wind Noise Reduction using Non-negative Sparse Coding",
author = "Mikkel N. Schmidt and Jan Larsen and Fu-Tien Hsiao",
booktitle = "Machine Learning for Signal Processing, IEEE Workshop on (MLSP)",
month = "Aug",
pages = "431--436",
year = "2007"
}