Zhihao Du

@ZhihaoDU · User

GitHub profile ↗ · Compare

China152 followers26 repositories

Repositories

ZhihaoDU/speech_feature_extractor

Some useful features of speech process, such as MFCC, gammatone filterbank, GFCC, spectrum(power spectrum and log-power spectrum), Amplitude Modulation Spectrum(AMS) and so on.

★ 129PythonForks 42

ZhihaoDU/du2020kws

This a small footprint robust KWS system which is based on the multi-conditional training, retraining and joint-training. This system includes a small-footprint KWS system and a small-footprint speech enhancement model. We also investigate a compress method for CNN and LSTM.

★ 2PythonForks 1

ZhihaoDU/du2020dan

The implementation of our paper "Double adversarial networks for monaural speech enhancement" accepted by INTERSPEECH 2020.

★ 7PythonForks 0

ZhihaoDU/asteroid

The PyTorch-based audio source separation toolkit for researchers || Pretrained models available

★ 0Forks 0

ZhihaoDU/neos_speech_utils

The speech utils may be useful for speech separation, speech enhancement, speech synthesis researchers. Enjoy it.

★ 0PythonForks 0

ZhihaoDU/griffin_lim

Implementation of the Griffin and Lim algorithm to recover an audio signal from a magnitude-only spectrogram.

★ 0PythonForks 0

ZhihaoDU/FloWaveNet

A Pytorch implementation of "FloWaveNet: A Generative Flow for Raw Audio"

★ 0PythonForks 0