pyAudioAnalysis icon indicating copy to clipboard operation
pyAudioAnalysis copied to clipboard

Problem encountered when doing mid term feature extraction

Open lxrswdd opened this issue 5 years ago • 1 comments

I did a `directory_feature_extraction_no_avg' with mid_window =1.0 and mid_step = 1.0 on a wav file with length of 1s. However, I got a (2,136) result instead of (1,136) as i expected. This happens to a 2s wav file too as the (4,136) is extracted instead of (2,136)

lxrswdd avatar Oct 09 '20 03:10 lxrswdd

directory_feature_extraction_no_avg is for directories. Also it does not extract long-term averages, so it extracts a matrix of mid-term features for all segments of all wav files in the folder.

You probably need mid_feature_extraction() instead

tyiannak avatar Oct 09 '20 09:10 tyiannak