This paper describes a tool developed at the Universities of Alcalá and Vigo whose purpose is to provide undergraduate students with a friendly way to learn the basics of sound-coding systems, their implementation, and how to adjust them. In contrast to the classical approach, in which the students are presented, in a theoretical way, the basic principles of sound coding, the proposed system avoids the need for the students to program, enabling them to focus their attention on those issues strictly related to the coding problem itself. The system has been designed to allow for the possibility to observe all of the information present in the sound coder in great detail, thus making it easier to understand all of the processes that take part in it. In addition, it is also possible to recreate pre-configured scenarios simulating different kinds of failures in the system.

Since only 20 hours were available for the topic of sound coding, it was necessary to improve the efficiency of the learning process at maximum. A first approach to overcome this limitation would be an all-theoretical course, which would allow to make a complete study of a wide range of coding algorithms. The main drawback is that, specially for a doctorate course, a more practical approach would be preferable.

We propose the introduction of a more interactive approach, where some hours are devoted to explaining the theory on the blackboard, leaving enough time to make some practices on these topics. This approach however has the limitation of forcing the teacher to reduce the amount of information given to the students. It is necessary to focus the attention on a limited number of coding structures rather than explaining lots of them. Given their popularity is was decided that in order to explain source coders, a CELP structure should be used, while the family of MPEG audio coders would be used as an example of perceptual audio coders. At this point a second problem arises. With only 20 hours it is quite difficult to make the students assimilate all the concepts necessary to understand the principles behind each one of the proposed coding structures. Most courses employ a classical approach consisting on giving the students some source code corresponding to parts of a given audio coder and ask them to implement in any programming language some other blocks from that encoder. The students usually complain about this approach, arguing that too much time is wasted solving debugging and programming problems rather than in purely audio coding issues. An alternative is therefore proposed, consisting on the use of the tool described in this paper so that no programming efforts are asked to the students, and they can focus their attention on understanding what it going on inside the coder, and how to interpret and diagnose the results obtained.