Capstone: Digital Signal Processing Toolkit

Build a cohesive DSP toolkit and apply it to real signals in your own repository.

Over 4–6 weeks, build a DSP toolkit in your OWN git repo: spectra, filtering, modulation/demodulation, and system analysis, applied end-to-end to a real signal (audio, ECG, communications…).

Spectral analysisFilter designModulation/demodulationSystem analysisTesting

The Mission

Choose a real signal and a task (e.g. denoise audio, detect ECG peaks, or demodulate an AM/FM signal). Build the DSP blocks — spectral analysis, filter design, and modulation — apply them, and deliver the toolkit + a report with spectra and results.

Milestone schedule

Week 1Repo + spectral analysis

›Set up repo, package, tests.
›FFT-based spectrum + windowing.

Deliverable: Tested spectral-analysis module.

Week 2Filtering

›IIR/FIR filter design + application.
›Validate response and denoising.

Deliverable: Filter module with a denoising demo.

Week 3Modulation / task-specific processing

›Modulation/demodulation or detection blocks.
›Apply to the chosen real signal.

Deliverable: End-to-end processing on real data.

Week 4+Integration, docs & report

›Package the pipeline.
›Docs + report with results.

Deliverable: Documented release + report.

Grading rubric

Correctness35%

DSP math is right; passes the harness.

Completeness25%

Spectra + filtering + modulation + application.

Code quality20%

Structure, tests, docs, reproducible.

Analysis & report15%

Spectra and results are sound and well presented.

Going beyond5%

Real-time processing, adaptive filters, or ML on features.

Submission

Git repo URL + tagged release.
Acceptance-test output.
Report (≤ 15 pages).
Optional processed-signal demo.