Speech Processing Project D

Aalto University

The goal of the course is to familiarise the student with the design, implementation, evaluation and documentation of a selected signal processing -oriented speech technology algorithm, building a low-bit-rate speech codec. First, an introduction lecture is given to recap the necessary background methodologies which are needed to implement the selected algorithm. Second, by working in teams of 2-3 members, the student conducts the project by progressing through the following steps:<br /><br />(1) The design and the signal processing -based implementation of the speech codec using, e.g., Matlab. <br />(2) The analysis and objective evaluation of the implemented codec. <br />(3) The informal, small-scale subjective evaluation of the implemented codec. <br />(4) The documentation of the entire project. The teams work in the project according to a pre-selected schedule consisting of a few milestones. <br /><br />At each milestone, the progress done is reported to the lecturer in short face-to-face meetings between the team members and the lecturer.

Back