Real-time implmentation of low bit-rate wideband speech coders on ADSP-21065L

Amadea Paula Q. Unisa

(MS Graduated: 1st Sem 2008-2009)

Abstract

In this thesis, real-time Wideband Sinusoidal (WS) and Wideband Linear Prediction (WLP) speech coders were implemented using the ADSP-21065L EZ-Kit LiteTM Evaluation Board. These speech coders are based on the parametric speech models which are extensions of existing sinusoidal and LP models for narrowband speech. The parameters extracted from the input speech are quantized using vector and scalar quantization methods. A real-time implementation was achieved by programming all the subroutines of the speech coders in assembly language.

The average complexities of the WS analysis and synthesis blocks are 2.036 million instructions per second (MIPS) and 2.814 MIPS. The WS speech coder requires 785 words of program memory and 5,446 words of data memory. For the WLP speech coder, the analysis and synthesis blocks obtained average cycle counts of 28.179 MIPS and 11.249 MIPS. The WLP speech coder occupies 5,980 words of program memory and 9,355 words of data memory.

Two WS configurations and all of the WLP configurations achieved Mean Opinion Score (MOS) of greater than 3, indicating that the synthesized speech has noticeable but acceptable impairments. All the WS and WLP configurations achieve Diagnostic Rhyme Test (DRT) scores of greater than 80, thus the listener could discriminate between the rhyming pairs of consonant-vowel-consonant (C-V-C) words.

This research gives way to future programmers to implement the wideband speech coders in different platforms, a well as other applications involving the algorithms implemented in this thesis. This thesis can also help build better compilers by taking into consideration the issues raised in this study and by applying the implementation done in this thesis.