Abstract:
An automatic acoustic feedback suppression algorithm based on Zoom-FFT (ZFFT) was proposed. To improve the analysis precision of the acoustic feedback frequency, ZFFT, which can distinguish the components of the distinguishable feature frequency when the original frequency resolution is constant and the main lobe is not overlapped, is used to refine the spectrum. In the comparison experiments, the proposed algorithm can improve the analysis precision of the acoustic feedback frequency, effectively suppress the acoustic feedback and improve the voice quality.