FFT Spectrum Analyzer
Mix up to three tones with noise, window it, and watch the spectrum respond.
Reading a spectrum honestly
The FFT shows the signal through the lens of its window. A rectangular window leaks energy into sidelobes — a strong tone can completely bury a weak neighbour (try tone 2 at 0.1 amplitude next to tone 1 with Rect, then switch to Blackman). The bin spacing is
and a tone that falls between bins spreads across several of them (scalloping). Windows trade main-lobe width against sidelobe level: Hann is the everyday default, Blackman buys ~25 dB more rejection for a wider lobe, and flat-top sacrifices resolution to make peak amplitudes accurate — that's why instruments use it for calibration. The noise floor you see also depends on N: every doubling of the FFT length drops it by 3 dB (processing gain), which is how lock-in-style measurements pull signals out of noise.