SparSig Sparse Signal Processing in Wireless Communication

Real-time Implementation of Sparse Linear Prediction for Speech Processing

Project participants:


Employing sparsity criteria in linear prediction of speech has been proven successful for several analysis and coding purposes. However, sparse linear prediction comes at the expenses of a much higher computational burden and numerical sensitivity compared to the traditional minimum variance approach. This makes sparse linear prediction difficult to deploy in real-time systems. In this paper, we present a step towards real-time implementation of the sparse linear prediction problem using hand-tailored interior-point methods. Using compiled implementations the sparse linear prediction problems corresponding to a frame size of 20 ms can be solved on a standard PC in approximately 2 ms and orders faster than with general purpose software.

Paper published at International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2013.

A Matlab mex interface and the proposed algorithms are available. If you are interested in using the algorihtms outside Matlab then it should also be possible (an example Makefile is provided for inspiration).