When to Trigger
Activate this skill when the user mentions:
- FFT, DFT, spectral analysis, power spectral density
- Digital filters (FIR, IIR), Butterworth, Chebyshev
- Time-frequency analysis, STFT, wavelets, spectrograms
- Signal denoising, SNR, noise floor
- Sampling, Nyquist theorem, aliasing, ADC/DAC
- Modulation (AM, FM, QAM), demodulation, baseband
- Convolution, correlation, matched filtering
Step-by-Step Methodology
- Signal characterization - Identify signal type (continuous/discrete, deterministic/stochastic, stationary/non-stationary). Determine sampling rate, duration, and bit depth. Check Nyquist criterion (fs > 2*fmax).
- Preprocessing - Remove DC offset (mean subtraction). Apply windowing (Hann, Hamming, Blackman) to reduce spectral leakage. Handle missing data or outliers. Normalize amplitude if needed.
- Spectral analysis - Compute FFT with appropriate zero-padding for frequency resolution. Estimate power spectral density (Welch's method for noise reduction, periodogram for snapshot). Identify dominant frequency components and harmonics.