Acoustic analysis of speech in pure Ruby: WAV decoding, radix-2 FFT and power spectra, generalized cosine windows, mel filterbanks and MFCC, the YIN fundamental frequency estimator, linear predictive coding by the Levinson-Durbin recursion with formant extraction from the all-pole envelope, spectral moments, Butterworth biquads, windowed-sinc decimation and short-term energy measures. No dependencies, no native extensions, deterministic output.