Hi, I really enjoy the web audio experiments you made and was playing around some useful functions you wrote. And noticed a few problems/ possible improvements:
So, as it says in the title - removeSilence() function breaks when feeding it with larger AudioBuffer (also with bigger silence chunks). Well it doesn't technically break but I did a little investigation and found out that the code works up until here:
return signal.filter((v, i) => { const frameIndex = Math.floor(i / numSamplesInFrame); return !longGroups.some(group => group.includes(frameIndex)); });
then it just continues working but making the CPU go above 100% for a way too long time. I believe it could be improved but my skills are limited here. How could I go around this?

Also, how can I make it return stereo buffer, I mean I could technically just fill the second channel with same audio data but would it be the right approach to do like this? Considering I don't want to loose the quality.

Thanks a lot! :)