Data
Long audio of chickens in the backyard with noise such as cars and birds. Source: Youtube poulty sounds with noise: https://www.youtube.com/watch?v=70IqKloH-mw
The following libraries were used; Noise Reduce and Bandpass Filter
--> First, install the following packages:
pip install librosa pip install ffmpeg to enable audio read. pip install noisereduce import noisereduce as nr to reduce noise from the audio file.
Next: : first convert from mp3 to .wav
Noisy Chicken audio: backyard_chickens_cut1.wav Denoised Chicken audio using Band Filtering: bandpass-filter_backyard_chickens_cut1.wav Denoised Chicken audio using Noise Reduce: denoised_backyard_chickens_cut1.wav
Preprocessing/Youtube/Predefined_noise : Contains wav files of predefined_noise (door sounds, footsteps sounds etc)
Preprocessing/Youtube/denoised_2: New denoised wav files after predefining the noise.
Preprocessing/Youtube/denoised_1 : Origininal denoised wav files without without predefining the noise.
uwunet_noise_predictor_100_epochs_linear_schedule.pt : pretrained model from kaggle: https://www.kaggle.com/models/baconerie/diffusion-model-noise-predictor
error (Denoise_1)
LibsndfileError Traceback (most recent call last) Cell In[56], line 6 4 # Remove batch and channel dimensions 5 cleaned_audio = cleaned_audio_tensor.squeeze().numpy() ----> 6 sf.write('cleaned_audio_tf.wav', cleaned_audio, sr)
File /opt/anaconda3/lib/python3.11/site-packages/soundfile.py:343, in write(file, data, samplerate, subtype, endian, format, closefd) 341 else: 342 channels = data.shape[1] --> 343 with SoundFile(file, 'w', samplerate, channels, 344 subtype, endian, format, closefd) as f: 345 f.write(data)
File /opt/anaconda3/lib/python3.11/site-packages/soundfile.py:658, in SoundFile.init(self, file, mode, samplerate, channels, subtype, endian, format, closefd) 655 self._mode = mode 656 self._info = _create_info_struct(file, mode, samplerate, channels, 657 format, subtype, endian) --> 658 self._file = self._open(file, mode_int, closefd) 659 if set(mode).issuperset('r+') and self.seekable(): 660 # Move write position to 0 (like in Python file objects) 661 self.seek(0)
File /opt/anaconda3/lib/python3.11/site-packages/soundfile.py:1216, in SoundFile._open(self, file, mode_int, closefd) 1213 if file_ptr == _ffi.NULL: ... 1219 # when opening a named pipe in SFM_WRITE mode. 1220 # See http://github.com/erikd/libsndfile/issues/77. 1221 self._info.frames = 0
LibsndfileError: Error opening 'cleaned_audio_tf.wav': Format not recognised. Output is truncated. View as a scrollable element or open in a text editor. Adjust cell output settings...
I learned that noise2noise is best for small samples