We are developing a new class, ``Introduction to Real-Time Digital Systems,'' that combines signal processing and computer architectures in a laboratory setting to excite sophomores about communications, signal processing, and controls. The students experiment with sampled speech, music, and image signals to gain experience in analyzing, enhancing, and performing real-time processing on them. They learn about Fourier analysis at an intuitive level as the key technique to unlock the composition of a signal, but also experience its usefulness in practice. The students explore many different views of signals and systems as they build signal processing algorithms using high-level and assembly languages. They also gain practical experience developing algorithms in MATLAB [1] and embedded applications on Texas Instruments TMS320C50 boards [3][2]. In demonstrations, students are also exposed to the visual block diagram programming environments SIMULINK [4] and Ptolemy [5].

The students for the course are sophomores majoring in Electrical Engineering and Computer Sciences (EECS). At Berkeley, the fields of electrical engineering, computer science, and computer engineering are taught in a single EECS Department. We have integrated these fields in the EECS courses so that they look seamless to the students. Thus, students are able to combine these fields in different ways, which is particularly appropriate in the context of modern technology. This new sophomore course offers students an exposure to a combination of communications and signal processing with computer science and engineering. The course comes at a time when the students are deciding on their areas of specialization from among electronics, systems, and computer science and engineering.

Regardless of their final area of specialization, all EECS students can benefit from the course. They gain an appreciation for real-time discrete systems and a digital style of implementation. Their understanding of discrete-time systems complements the analog and digital circuit design and desktop software programming they are learning in their other lower-division classes. Because we introduce systems by way of applications, interesting concepts, and digital computing, we hope to motivate students to study communications, signal processing, and controls. For those students who choose the systems area as their specialization, their practical understanding of concepts such as the frequency domain, sampling, aliasing, and quantization will give them better motivation to study the theory in later systems classes, because they have a greater appreciation of the application of the theory beforehand.

The initial offering of this real-time DSP course is during the Spring 1996 semester. This two-credit course runs for sixteen weeks, as shown in Tables 1 and 2 on the next page. The course consists of one hour of lecture, one hour of discussion, and four hours of laboratory work each week. Throughout the course, concepts and applications are interwoven. Each lecture demonstrates applications to illustrate concepts being presented. In each laboratory, students further their understanding by developing signal processing systems. The students spend 3-5 weeks on each of the following application areas: computer music and digital audio, speech, digital communications, and image processing. Within each application area, students build simulations and real-time systems. In developing the laboratories for this course, we found several books on MATLAB [8][7][6] and the family of Texas Instruments fixed-point DSP processors [10][9][3] to be very helpful.

Brian L. Evans, 211-105 Cory Hall, Berkeley, CA 94720-1772