search

madhavmk / Noise2Noise-audio_denoising_without_clean_training_data

Source code for the paper titled "Speech Denoising without Clean Training Data: a Noise2Noise Approach". Paper accepted at the INTERSPEECH 2021 conference. This paper tackles the problem of the heavy dependence of clean speech data required by deep learning based audio denoising methods by showing that it is possible to train deep speech denoising networks using only noisy speech samples.
MIT License
172 stars 42 forks source link

Saved audio file length #5

Closed joeoct93 closed 1 year ago

joeoct93 commented 1 year ago

Hello, I've been trying out your noise2noise denoising on jupyter notebook, and I found that the audio saved is locked to 3 seconds, even if the audio processed is more than 3 seconds long. Is there a way to control this? Thank you.

madhavmk commented 1 year ago

Hello @joeoct93, sorry for the late reply Yes, we are cutting/padding audios in the datasets to 165000 numpy array length, which equates to 165000/48000 = 3.4 seconds on a 48khz sample rate. This fixed length can be modified by setting self.max_len attribute in the SpeechDataset class.