General Information[]
Introduces statistical signal processing. Signal representation and manipulation are covered via correlation and using the DFT/FFT to estimate other transforms; applications of these topics are then covered, including quantization, quantization effects in digital filters, multirate DSP, filter banks, delta-sigma modulation, power spectrum estimation, and introductions to Wiener and Kalman filtering and image processing.
Prerequisites[]
Prerequisite: ECE 3250
Outcomes[]
- Be able to draw Fourier spectra in both discrete-time frequency and continuous-time frequency, while undergoing common operations such as filtering, upsampling, downsampling, A/D and D/A.
- Given a finite-duration, discrete-time signal, be able to estimate the discrete-time Fourier Transform and the original frequency content in the signal (e.g., the continuous-time Fourier Transform) and give bounds on the accuracy of the estimate.
- Given a continuous-time signal with certain frequency- and time-based characteristics, be able to design a (real-world, non-ideal) system to appropriately sample the signal such that desired characteristics are maintained to within given tolerances.
- Given a digital audio (1-d) or image (2-d) signal, be able to select an appropriate frequency- based transform suitable for the desired or specified processing.
- Quickly prototype and debug signal processing algorithms using Matlab.
Workload[]
SP17 - Light. 2 problem sets. 2 exams. 1 project (+report+presentation).
Advice[]
SP17- Be careful not to fall behind on material as there is virtually no formal work for this class.
The past 4 offerings of the course (SP15, 16, 17, 18) have all had different instructors, so be warned that the class is changing.
Past Offerings[]
| Semester | Time | Professor | Median Grade |
|---|---|---|---|
| Spring 2017 |
TR 11:40-12:55 |
David Delchamps | A |
| Spring 2018 |
TR 11:40-12:55 |
Mert Sabuncu |