Digital Signal Processing

What you'll learn

• Describe discrete-time signals and LTI systems by difference equations and convolution, express frequencies in rad/sample, and prepare a real-time cycle and latency budget for a DSP product.
• Determine alias frequencies, size anti-aliasing and reconstruction filtering for a chosen sample rate, apply bandpass sampling, and estimate the effects of zero-order-hold droop and clock jitter.
• Analyze quantization noise, dither, oversampling and sigma-delta noise shaping, interpret converter specifications such as SINAD and ENOB, and set the analog gain in front of a converter.
• Use the z-transform and its region of convergence to find system functions, poles and zeros, impulse responses and stability, and design resonators and tone generators.
• Compute and interpret frequency, phase and group-delay responses with the DTFT, and explain linear-phase, all-pass and minimum-phase filters and when each is appropriate.
• Apply the DFT to measure tones and noise correctly, choosing windows, block lengths and averaging, and relate circular and linear convolution.
• Explain and implement radix-2 FFT algorithms, count their operations, handle real inputs and fixed-point scaling, and use overlap-add or overlap-save for fast convolution.
• Design linear-phase FIR filters to a specification by the window method with Kaiser's formulas and by equiripple methods, and estimate the filter length needed.
• Design IIR filters from Butterworth, Chebyshev and elliptic prototypes with the bilinear transform and prewarping, and correct designs whose features lie near the Nyquist frequency.
• Select filter structures and fixed-point formats, and predict and control coefficient quantization, roundoff noise, overflow and limit cycles.
• Design decimators, interpolators and rational and multistage rate converters using polyphase structures and CIC filters.
• Apply LMS and NLMS adaptive filters and the Goertzel algorithm, and integrate and justify a complete DSP signal chain for a product.