Minimum Processing Beamforming

Most of the well-known classic beamformers have resulted from optimization problems that minimize a cost function such as mean-square error (MSE) between noisy speech and reference clean speech. The rationale behind these formulations involves speech-versus-noise dichotomy, where anything branded noise shall be suppressed much possible. While leading to simple closed-form solutions reasonably practical beamformers, this has its own limitations, for instance, when ambient provides context is therefore not entirely undesirable. In article, we offer new rationale, output beamformer minimally processed with respect certain signal, long given performance criterion fulfilled. We provide case study inspired by Speech Intelligibility Index (SII), processing penalty MSE. Regarding consider two cases. first case, signal set unprocessed recording microphone, giving rise limits minimum necessary fulfilling intelligibility requirement. For second an aggressive beamformer, yielding essentially eliminates unless concomitant distortion violates Through simulation studies, demonstrate some benefits each cases in relevant contexts.