Original scientific paper
https://doi.org/10.2498/cit.1001767
Gaussian Mixture Model-based Quantization of Line Spectral Frequencies for Adaptive Multirate Speech Codec
Tihomir Tadić
; Research & Development Center, Ericsson Nikola Tesla d.d., Zagreb, Croatia
Davor Petrinović
orcid.org/0000-0003-3950-7864
; Department of Electronic Systems and Information Processing, Faculty of Electrical Engineering and Computing, University of Zagreb, Zagreb, Croatia
Abstract
In this paper, we investigate the use of a Gaussian MixtureModel (GMM)-based quantizer for quantization of the Line Spectral Frequencies (LSFs) in the Adaptive Multi-Rate (AMR) speech codec. We estimate the parametric GMM model of the probability density function (pdf) for the prediction error (residual) of mean-removed LSF parameters that are used in the AMR codec for speech spectral envelope representation. The studied GMM-based quantizer is based on transform coding using Karhunen-Loeve transform (KLT) and transform domain scalar quantizers (SQ) individually designed for each Gaussian mixture. We have investigated the applicability of such a quantization scheme in the existing AMR codec by solely replacing the AMR LSF quantization algorithm segment. The main novelty in this paper lies in applying and adapting the entropy constrained (EC) coding for fixed-rate scalar quantization of transformed residuals thereby allowing for better adaptation to the local statistics of the source. We study and evaluate the compression efficiency, computational complexity and memory requirements of the proposed algorithm. Experimental results show that the GMM-based EC quantizer provides better rate/distortion performance than the quantization schemes used in the referent AMR codec by saving up to 7.32 bits/frame at much lower rate-independent computational complexity and memory requirements.
Keywords
Gaussian mixturemodels (GMMs); Karhunen-Loeve transform (KLT); line spectral frequency (LSF); Adaptive Multi-Rate (AMR); speech coding; transform coding; vector quantization (VQ); entropy constrained scalar quantizer (ECSQ)
Hrčak ID:
71049
URI
Publication date:
30.6.2011.
Visits: 1.893 *