Robust Source-Filter Separation of Speech Signal in the Phase Domain

Loweimi, Erfan; Barker, Jon; Torralba, Oscar Saz; Hain, Thomas

doi:10.21437/interspeech.2017-210

Cited by 21 publications

(31 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1) instead of (2). As shown in [23], arg[X M inP h (ω)], contrary to the wrapped phase (ARG[X(ω)]), no longer exhibits a chaotic and trendless shape. In fact, it could be imagined as a superposition of two components, one changing slowly (Trend) and the other one oscillating quickly (Fluctuation).…”

Section: Phase-based Source-filter Separationmentioning

confidence: 99%

“…As explained in [23], Trend and Fluctuation correspond to the vocal tract (VT) and excitation (Exc) components, respectively. Fig.…”

Section: Phase-based Source-filter Separationmentioning

confidence: 99%

“…For evaluating the efficacy of the suggested approach, a feature (named BMFGDVT [23]) was extracted from the vocal tract (filter) component of the phase spectrum and tested for its usefulness in ASR (Fig. 2).…”

Section: Phase-based Source-filter Separationmentioning

confidence: 99%

“…In [23], we proposed a novel framework for source-filter separation in the phase domain. The method was successful in segregating the vocal tract and excitation elements and its superiority in comparison with the magnitudebased approach was discussed and illustrated.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Statistical normalisation of phase-based feature representation for robust speech recognition

Loweimi

Barker

Hain

2017

2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

In earlier work we have proposed a source-filter decomposition of speech through phase-based processing. The decomposition leads to novel speech features that are extracted from the filter component of the phase spectrum. This paper analyses this spectrum and the proposed representation by evaluating statistical properties at various points along the parametrisation pipeline. We show that speech phase spectrum has a bell-shaped distribution which is in contrast to the uniform assumption that is usually made. It is demonstrated that the uniform density (which implies that the corresponding sequence is least-informative) is an artefact of the phase wrapping and not an original characteristic of this spectrum. In addition, we extend the idea of statistical normalisation usually applied for the magnitudebased features into the phase domain. Based on the statistical structure of the phase-based features, which is shown to be super-gaussian in the clean condition, three normalisation schemes, namely, Gaussianisation, Laplacianisation and table-based histogram equalisation have been applied for improving the robustness. Speech recognition experiments using Aurora-2 show that applying an optimal normalisation scheme at the right stage of the feature extraction process can produce average relative WER reductions of up to 18.6% across the 0-20 dB SNR conditions.

show abstract

Section: Phase-based Source-filter Separationmentioning

confidence: 99%

“…As explained in [23], Trend and Fluctuation correspond to the vocal tract (VT) and excitation (Exc) components, respectively. Fig.…”

Section: Phase-based Source-filter Separationmentioning

confidence: 99%

Section: Phase-based Source-filter Separationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Statistical normalisation of phase-based feature representation for robust speech recognition

Loweimi

Barker

Hain

2017

2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

show abstract

“…Specifically, it can only be used for MFCC or log filterbank energies (FBE) and can not be directly applied to representations which -instead of log compression -apply a power transformation to the FBEs. Examples of such features are generalised-MFCC [12], PLP [3], PNCC [5] and phase-based features [4,[13][14][15][16][17][18].…”

Section: Introductionmentioning

confidence: 99%

Use of Generalised Nonlinearity in Vector Taylor Series Noise Compensation for Robust Speech Recognition

2016

Self Cite

View full text Add to dashboard Cite

Designing good normalisation to counter the effect of environmental distortions is one of the major challenges for automatic speech recognition (ASR). The Vector Taylor series (VTS) method is a powerful and mathematically well principled technique that can be applied to both the feature and model domains to compensate for both additive and convolutional noises. One of the limitations of this approach, however, is that it is tied to MFCC (and log-filterbank) features and does not extend to other representations such as PLP, PNCC and phase-based front-ends that use power transformation rather than log compression. This paper aims at broadening the scope of the VTS method by deriving a new formulation that assumes a power transformation is used as the non-linearity during feature extraction. It is shown that the conventional VTS, in the log domain, is a special case of the new extended framework. In addition, the new formulation introduces one more degree of freedom which makes it possible to tune the algorithm to better fit the data to the statistical requirements of the ASR back-end. Compared with MFCC and conventional VTS, the proposed approach provides upto 12.2% and 2.0% absolute performance improvements on average, in Aurora-4 tasks, respectively.

show abstract

Exploring Multimodal Features and Fusion for Time-Continuous Prediction of Emotional Valence and Arousal

Kumar

Choi

Pandey

et al. 2022

Intelligent Human Computer Interaction

View full text Add to dashboard Cite

Robust Source-Filter Separation of Speech Signal in the Phase Domain

Cited by 21 publications

References 19 publications

Statistical normalisation of phase-based feature representation for robust speech recognition

Statistical normalisation of phase-based feature representation for robust speech recognition

Use of Generalised Nonlinearity in Vector Taylor Series Noise Compensation for Robust Speech Recognition

Exploring Multimodal Features and Fusion for Time-Continuous Prediction of Emotional Valence and Arousal

Contact Info

Product

Resources

About