Open mattberjon opened 7 years ago
It could be nice to have a function for audio resampling.
It could be nice as well to have program that could check the samplerate of the audio database for conflicts according to the samplerate in the config file or just samplerate outliers in the database.
The output would be the outliers paths and the possibility of resampling them.
At the moment I take extract the samplerate of the audio file and use it to perform the computing. The issue is that if you have several samplerate in your audio batch it can lead to some artefacts for the machine learning algorithm. Moreover, we specify in the config file a samplerate.
It would be good to check every time the samplerate from the config file and the audio file. If different, raising a flag would be advised. Several possibilities are offered: