Robust spectral representation using group delay function and stabilized weighted linear prediction for additive noise degradations

D. Gowda, J. Pohjalainen, P. Alku, M. Kurimo

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract / Description of output

In this paper, we propose a robust spectral representation using the group delay (GD) function computed from the stabilized weighted linear prediction (SWLP) coefficients. Temporal weighting of the cost function in linear prediction (LP) analysis with the short-term energy of the speech signal improves the robustness of the resultant spectrum. The additive property of the group delay function provides for better representation of weaker resonances in the spectrum, and thereby improving the robustness of the representation. The SWLP provides robustness in the temporal domain, whereas the GD function provides robustness in the frequency domain. The proposed SWLP-GD representation is shown to be robust against different types of additive noise degradations, compared to the popularly used discrete Fourier transform (DFT) or LP based representations. In a small-scale closed-set speaker recognition experiment, the cepstral features derived from the proposed SWLP-GD spectrum perform better than the traditional mel-cepstral features computed from the discrete Fourier transform (DFT) spectrum under conditions of mismatched degradations.
Original languageEnglish
Title of host publicationSpeech Technology and Human - Computer Dialogue (SpeD), 2013 7th Conference on
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages1-7
Number of pages7
ISBN (Print)978-1-4799-1065-6
DOIs
Publication statusPublished - 1 Oct 2013

Keywords / Materials (for Non-textual outputs)

  • cepstral analysis
  • delays
  • frequency-domain analysis
  • prediction theory
  • signal representation
  • speaker recognition
  • speech processing
  • GD function
  • SWLP-GD representation
  • SWLP-GD spectrum
  • additive noise degradation
  • additive property
  • cepstral feature
  • cost function
  • frequency domain analysis
  • group delay function
  • robust spectral representation
  • spectrum robustness improvement
  • speech signal processing
  • stabilized weighted linear prediction
  • temporal domain analysis
  • temporal weighting
  • Degradation
  • Delays
  • Discrete Fourier transforms
  • Predictive models
  • Robustness
  • Signal to noise ratio
  • Speech
  • frequency weighted segmental SNR
  • robust spectrum estimation

Fingerprint

Dive into the research topics of 'Robust spectral representation using group delay function and stabilized weighted linear prediction for additive noise degradations'. Together they form a unique fingerprint.

Cite this