Find the duration of speech

i have a video or audio with 2 persons speaking... Is it possible to get the

 1.   duration of speech by person1 and person2 separately 
 2.   duration of overlapping speech 
 3.   duration of silence also

from the video or audio input?

Answer 1

[Image Analyst]

Why can't you just threshold the signal and then use regionprops() to find out where the silent regions start and stop? I know you've used regionprops() before.

[silentRegions, numRegions] = y < someThreshold;
labeledRegions = bwlabel(silentRegions);
props = regionprops(labeledRegions, 'PixelList');
for k = 1 : numRegions
    fprintf('Silence region %d is between element %d and %d.\n',...
      k, props(k).PixelList(1), props(k).PixelList(end));
end