1. Multi-Microphone Spatial Beamforming
Simulates dual end-fire omnidirectional microphones (inter-element spacing d = 12 mm). Implements fractional delay-and-sum FIR beamforming, steering nulls toward back-hemisphere noise sources (DI = +4.8 dB).
2. Normalized Least Mean Squares (NLMS) Adaptive Feedback Cancellation
Suppresses electro-acoustic feedback whistling (leakage from receiver back to microphone inlet) by synthesizing an adaptive FIR filter W(z) tracking the dynamic feedback path F(z) via Normalized LMS gradient descent, yielding >12 dB Added Stable Gain (ASG).
3. Multi-Channel Wide Dynamic Range Compression (WDRC)
Compresses acoustic speech dynamic range across octave bark bands using syllabic attack/release envelope detection (tatt = 5 ms, trel = 50 ms), restoring intelligibility for sensorineural hearing loss profiles.