Batch split at silence and interval

I am trying to split a huge number of files (more than 100) each into a set of small mp3 file. Thus for an app I am building that will download those files and rebuild thwm on thw device.

I tried several tools (mp3splitter from pristol, cool mp3 cutter and mp3 splitgtk) each had it’s own shortcomings or bugs.

I am trying to cut each file at silence point of 2sec and in a way that each chunk is at least 2 minute long. So silence points within 2minutes after previous silence will be disregarded. Reason I am using 2sec silence is sothat the rebuild of the file from the chunks will not be noticeable to the listener. Reason for the 2min chunk length is to reach optimal number of mp3 chunks per mp3 original (less than 2min interval will generate huge number of files per original)

I resorted to audacity and found I could use silence finder to locate silnce points. But as you cam guess they will be numerous. Mp3splitter gtk had option to limit minimum length of a split.

I could also insert the interval labels (at 2 min) ; but the follow on work to delete unwanted labels is too laborious. Is there a way (or a trick) to introduce silence labeling with interval boundaries in audacity?

You could try the attached Nyquist plugin. It’s Steve’s, not mine and it is experimental so use at your own risk. You are labelling the sounds. There is a control for minimum label length.

See here for how to install Nyquist plugins: You can add it to an Audacity batch processing “Chain” but it won’t help much because you can only export single files using a Chain, not export multiple. You would have to apply the Sound Finder to a file then export multiple manually.

In any case Audacity re-encodes MP3’s so you will degrade them. You will be better to search further for direct MP3 editors. Did you try mp3DirectCut? It is supposed to have “Auto cue with minimum length”. See here or read their documentation.

soundFinder-New3.ny (5.98 KB)

Thanks for your answer.
MP3DC was one of the first tools I tried (among mp3splt-gtk, cool mp3 cutter, and Direct WAV MP3 Splitter from Pristol).
My issue with MP3DC as well as others is that the splits they generate do not have a stable file length measure. I am not sure how this is done; but I play the output files and the the audio time at the end of the seekbar keeps changing (e.g. 5:42, 5:36, 5:27 ) as if the player (VLC) is unable to tell the length of the audio. It gets even funnier: The same file is reported by windows media player to be 13:44 while by itunes 4:58

I reassemble the splits at the smartphone device once they’re downloaded into my audiobooks app; the app panics as it cannot tell the full chapter length. This is because chapter is formed of around 10 splits; so app sums up “what it thinks is” the file length of each file.

Maybe I am doing something wrong with MP3DC; but I did not see any relevant option in the pause detection box or the split save options.
After your reply, I read abit more about MP3DC but did not see someone complaining of this side effect

And thanks for re-confirming my concern with using Audacity, I almost resorted to it out despair.
Not sure where to go next

Did you try mp3TrueEdit? The free version is feature-limited and I don’t think it has the exact feature you want but I know the person who develops it - he might be prepared to add the feature you want as some kind of commercial arrangement.