A General Monophonic Real-time Noise Reduction Method

Abstract

The invention relates to a universal single channel real-time noise-reduction method. The universal single channel real-time noise-reduction method comprises the following steps that noisy voice of an electronic format is received and comprises voice and non-human-voice interference noise; a short-time Fourier magnitude spectrum is extracted frame by frame from the received voice to serve as an acoustic characteristic; a specific value film is generated frame by frame through a deep recurrent neural network with long-and-short-term memories; the magnitude spectrum of the noisy voice is sheltered through the generated specific value film; and the sheltered magnitude spectrum and an original phase of the noisy voice are used, and through inverse Fourier transform, a voice waveform is synthesized again. According to the universal single channel real-time noise-reduction method, voice noise reduction is conducted through a supervised learning method, and the ideal specific value film is estimated through the recurrent neural network with the long-and-short-term memories; and the recurrent neural network provided by the invention is trained through the large amount of noisy voice, various realistic acoustic scenes and microphone impulse responses are included, and finally universal voice noise reduction independent of background noises, speakers and transmission channels is achieved.

Description

technical field [0001] The present invention relates to a general monophonic real-time noise reduction method, and more particularly, to a new method for mask estimation using a long short-term memory (LSTM) recurrent neural network (RNN). Background technique [0002] The speech noise reduction task refers to the separation of speech signals from noisy speech signals, and this technique has a wide range of applications, such as robust automatic speech recognition (ASR) and mobile communications in everyday environments. Speech noise reduction or separation has been studied in the field of signal processing for decades. Among them, monophonic speech noise reduction is a very challenging topic, because monophonic speech noise reduction only relies on a single microphone recording signal, and cannot utilize the commonly used spatial information of microphone arrays. On the other hand, monophonic noise reduction can be applied to a wider range of acoustic scenarios than microp...

AI Technical Summary

Problems solved by technology

[0005] The previous state-of-the-art speech noise reduction scheme is a feedforward deep neural network (Deepneural network, DNN) trained with a large amount of data. Although this scheme can separate specific human voices from untrained noise, the model is not effective for non- Noise cancellation isn't great for certain vocals

In order to improve the noise reduction effect of non-specific human voices, the most effective method is to add the voices of multiple speakers in the training set. However, this will cause DNN to confuse voices and background noise, and tend to misclassify noises as voices.

Images