A simulation tool for introducing algebraic celp (ACELP) coding concepts in a DSP course

Abstract

This paper presents an educational tool 1 for introducing Code Excited Linear Prediction (CELP) coding concepts in senior undergraduate and graduate DSP-related courses. The tool consists of a user-friendly graphical interface along with a complete MATLAB 2 realization of all aspects of the Algebraic CELP G.729 algorithm [2]. This simulation software is accompanied by a series of computer experiments and exercises that can be used to provide hands-on training to class participants. The exercises designed based on the simulation tool may be used by instructors in a class setting to demonstrate key signal processing concepts associated with the processing of telephone-based speech. The MATLAB ACELP tool is being used in Arizona State University undergraduate DSP courses as well as in a graduate course on speech coding and in a continuing education short course. Evaluation of the tool and the exercises is being performed by an educational software assessment specialist. In the last ten years we have witnessed a series of breakthroughs in speech coding followed by several standardization efforts [1]. Most of the standardized algorithms are based on CELP coders. Although speech coding researchers and practitioners are well aware of the fundamental ideas used in CELP, students do not get much of an opportunity in courses to study these algorithms. A software simulation tool, implementing the ACELP algorithm has been developed for the purpose of introducing speech coding and the associated signal processing concepts to both undergraduate and graduate students. We choose an Algebraic Code Excited Linear Prediction (ACELP) algorithm as a basis for this educational tool because of the wide proliferation of algebraic codebooks in cellular standards. The algorithm was coded in a modular manner and in its entirety using MATLAB. The tool is based on a user-friendly graphical user interface (GUI) that allows the student to study and verify through graphics the various aspects of the algorithm such as: the LP analysis, the open-loop pitch search, the adaptive codebook search (pitch search), the fixed codebook search, and the bit allocation patterns. We choose MATLAB as the implementation platform because it allows the user to easily understand the complex parts of the algorithm whose function is not