AN IMPROVED ENDPOINT DETECTION ALGORITHM BASED ON ENERGY-ZERO-PRODUCT
To tackle the problem of the significantly reduced accuracy of the traditional endpoint detection algorithm based on short-term energy and short-term zero-crossing rate when the signal-to-noise ratio is less than 10 dB,an improved energy-zero-product endpoint detection algorithm based on the enhancement of the Sonogram feature is proposed.The spectral subtraction method was utilized to reduce the noise of the audio signal and thus improving the signal-to-noise ratio.Perform erosion,binarization and dilation were processed on the sonogram to achieve the enhancement of the spectrogram characteristics.The energy-zero-product feature was extracted and the dual-threshold endpoint detection algorithm was used to conduct endpoint detection on the audio signal.Experimental results show that the algorithm can effectively capture the fragments with ambient abnormal sound under different signal-to-noise ratio conditions,and has good robustness.