MTG / DCASE-models

Python library for rapid prototyping of environmental sound analysis systems
MIT License
42 stars 5 forks source link

Bug in FeatureExtractor.pad_audio #12

Closed pzinemanas closed 3 years ago

pzinemanas commented 3 years ago

We are using librosa.core.frames_to_samples to calculate the expected length after padding, but this is wrong because this function returns the center time (in samples) for each frame (the last one in this case).