Robust Fundamental Frequency Estimation in Coloured Noise
(2020) 2020 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2020 In ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings 2020-May. p.741-745- Abstract
Most parametric fundamental frequency estimators make the implicit assumption that any corrupting noise is additive, white Gaus-sian. Under this assumption, the maximum likelihood (ML) and the least squares estimators are the same, and statistically efficient. However, in the coloured noise case, the estimators differ, and the spectral shape of the corrupting noise should be taken into account. To allow for this, we here propose two schemes that refine the noise statistics and parameter estimates in an iterative manner, one of them based on an approximate ML solution and the other one based on removing the periodic signal obtained from a linearly constrained minimum variance (LCMV) filter. Evaluations on real speech data indicate that... (More)
Most parametric fundamental frequency estimators make the implicit assumption that any corrupting noise is additive, white Gaus-sian. Under this assumption, the maximum likelihood (ML) and the least squares estimators are the same, and statistically efficient. However, in the coloured noise case, the estimators differ, and the spectral shape of the corrupting noise should be taken into account. To allow for this, we here propose two schemes that refine the noise statistics and parameter estimates in an iterative manner, one of them based on an approximate ML solution and the other one based on removing the periodic signal obtained from a linearly constrained minimum variance (LCMV) filter. Evaluations on real speech data indicate that the iteration steps improve the estimation accuracy, therefore offering improvement over traditional non-parametric fundamental frequency methods in most of the evaluated scenarios.
(Less)
- author
- Jaramillo, Alfredo Esquivel LU ; Jakobsson, Andreas LU ; Nielsen, Jesper Kjar and Grasboll Christensen, Mads
- organization
- publishing date
- 2020
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- keywords
- coloured noise, fundamental frequency, LCMV filter, least-squares, maximum likelihood, pre-whitening
- host publication
- 2020 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2020 - Proceedings
- series title
- ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
- volume
- 2020-May
- article number
- 9053018
- pages
- 5 pages
- publisher
- IEEE - Institute of Electrical and Electronics Engineers Inc.
- conference name
- 2020 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2020
- conference location
- Barcelona, Spain
- conference dates
- 2020-05-04 - 2020-05-08
- external identifiers
-
- scopus:85089237075
- ISSN
- 1520-6149
- ISBN
- 9781509066315
- DOI
- 10.1109/ICASSP40776.2020.9053018
- language
- English
- LU publication?
- yes
- id
- 78621d4f-cd2f-4a5a-a3b3-13348cb571b6
- date added to LUP
- 2020-08-19 08:42:55
- date last changed
- 2022-04-19 00:13:10
@inproceedings{78621d4f-cd2f-4a5a-a3b3-13348cb571b6, abstract = {{<p>Most parametric fundamental frequency estimators make the implicit assumption that any corrupting noise is additive, white Gaus-sian. Under this assumption, the maximum likelihood (ML) and the least squares estimators are the same, and statistically efficient. However, in the coloured noise case, the estimators differ, and the spectral shape of the corrupting noise should be taken into account. To allow for this, we here propose two schemes that refine the noise statistics and parameter estimates in an iterative manner, one of them based on an approximate ML solution and the other one based on removing the periodic signal obtained from a linearly constrained minimum variance (LCMV) filter. Evaluations on real speech data indicate that the iteration steps improve the estimation accuracy, therefore offering improvement over traditional non-parametric fundamental frequency methods in most of the evaluated scenarios.</p>}}, author = {{Jaramillo, Alfredo Esquivel and Jakobsson, Andreas and Nielsen, Jesper Kjar and Grasboll Christensen, Mads}}, booktitle = {{2020 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2020 - Proceedings}}, isbn = {{9781509066315}}, issn = {{1520-6149}}, keywords = {{coloured noise; fundamental frequency; LCMV filter; least-squares; maximum likelihood; pre-whitening}}, language = {{eng}}, pages = {{741--745}}, publisher = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}}, series = {{ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings}}, title = {{Robust Fundamental Frequency Estimation in Coloured Noise}}, url = {{http://dx.doi.org/10.1109/ICASSP40776.2020.9053018}}, doi = {{10.1109/ICASSP40776.2020.9053018}}, volume = {{2020-May}}, year = {{2020}}, }