Signal Computing: Digital Signals in the Software Domain

Spring 2020

Michael Stiber
Bilin Zhang Stiber
University of Washington Bothell
18115 Campus Way NE
Bothell, Washington 98011

Eric C. Larson
Southern Methodist University
Lyle School of Engineering
3145 Dyer Street
Dallas, TX 75205

A digital signal processing (DSP) textbook, and other materials, aimed at computer science students. Why one specifically for CS students? DSP has traditionally been taught as part of the electrical engineering curriculum and, as such, is accessible to CS students only to the extent that they have taken its prerequisites. In some engineering schools, that may be the case, but it is often the case that the CS math curriculum lacks some of the expected prerequisites. Moreover, DSP is most often taught as an upper-level EE course, and therefore, practically speaking, inaccessible to most non-EE majors.

But it's not just these practical matters that argue for a DSP textbook for CS majors. DSP really means something different to CS majors, who will be focusing in their professional lives more on the use of such systems as modules in larger projects, or on collaborating with EEs in product design. To a computing professional, DSP means building multimedia systems, or embedded systems, or ubiquitous computing devices.

It's that motivation that led us to write a book that:

• assumes a mathematical background likely to be common for the vast majority of CS majors,
• structures its mathematical content as a progression through different abstractions; a process that should be familiar to CS majors,
• emphasizes algorithms and computational complexity over things such as filter design, and
• relates DSP to applications in multimedia and data analysis.

The Spring 2020 release PDF of our book is available here. We maintain a Github page and repository, and encourage you to make your own contributions to this ongoing work; send us pull requests. Certainly, some of our goals are perhaps still more aspirational than actual!

This textbook is also indexed at College Open Textbooks and The Open Textbook Library.

Let us know if you use this book, or part of this book, in your class, or if you otherwise find it useful. Thank you! We also have a draft Instructor's Guide, available upon request at that same email address to faculty using all or part of the textbook or lab book. It's incomplete, but includes solutions to some of the problems at the end of each textbook chapter and to most of the relevant lab steps.

Other Resources

• A Laboratory Manual. This has been rewritten as of summer 2015 to use Matlab.
• A Matlab class that simulates an analog signal and provides various operations on analog signals, including sample and hold.
• A Matlab function to quantize a one-dimensional signal.
• A Matlab function to plot a discrete signal in "stair step" style.
• Audio files to demonstrate filtering using the sum of four sinusoids: sine4.au and filtered_sine4.au.
• A Matlab file to illustrate beating, plus audio files 1000Hz, 1005Hz, and beating.
• A Matlab .m file for demonstrating quantization using tones is available as quantdemo1.m.
• A Matlab .m file for demonstrating quantization using a more interesting sound is available as quantdemo2.m, along with a data file, amoriole2.mat.
• A Matlab script that illustrates the operation of the discrete cosine transform (actually, the magnitude of the 2D FFT) on simple sinusoidal images.

This material is based upon work supported by the National Science Foundation under Grant No. 0443118.