Digital Signal Processing Reference
In-Depth Information
Appendix C
DSP Project Ideas
Below are some ideas for semester projects in digital signal processing. Some of
the projects are more complex than others and will take more time. Also, these
projects can be made more interesting by applying advanced techniques such as
neural networks, hidden Markov modeling, adaptive ltering, and other tools beyond
the scope of this topic.
Over the span of a couple of months, projects are not going to be industrial-
strength. Shortcuts should be included in the original project proposal, and ac-
knowledged in a nal report. For example, the number of test signals may be very
small, and the eect on the project's results should be discussed.
1. Image watermarker
Devise a way to put a simple watermark in an image that is invisible to a
human observer.
2. Noise ltering
Create a program to lter a noisy signal, such as cleaning up a speech signal
recorded in a room full of other conversations.
3. Compression
Given a music signal, store the information of both channels as eciently as
possible. What if your program subtracts one channel from the other? What
if you use the previous sample as the expected value for the next sample, and
only store the dierence?
4. Music distinguisher
Study songs representative of dierent music types, and write a program that
distinguishes between them. For example, one might expect rap music to have
more energy in the low frequency range (bass) compared to country music.
411
