search

tyiannak / pyAudioAnalysis

Python Audio Analysis Library: Feature Extraction, Classification, Segmentation and Applications
Apache License 2.0
5.88k stars 1.2k forks source link

Filtering um..., er..., breathing, smacking and slurping noises #69

Open ErfolgreichCharismatisch opened 7 years ago

ErfolgreichCharismatisch commented 7 years ago

Let's say, you say a lot of um..., er... and do breathing and smacking/slurping noises and say filling words like really. How would you go about removing them with pyAudioAnalysis in audio?

spnichol commented 7 years ago

Well, in theory you'd go about removing them the same way as anything else: collect as large a sample as possible of these undesirable noises and train a KNN or SVM (or whatever) model to classify their occurrence. Then you can use the mtFileClassification function to find when they occur in your test data, and you could then crop them out.