Chain Processing + Find Sound + Export Multiple

I have many (200+) sound files.
Each of those files contains 4-5 sounds that I want to split up into separate files.

The “Find Sound” thing works really well, and doing Export Multiple works fine as well.

Now my problem is that I’d have to do this step (Find Sound → Export Multiple) many times, once for each source file.

I’ve taken a look at the Chain processing methods, but I cannot find Export Multiple in the chain editor.

How do I apply the above steps to all my source files automatically?
If that’s not possible for whatever reason, maybe there’s an addon that does it?
Does anyone have any ideas what I could do?
Saving the labels that say where each sound in each file starts and ends would help as well, if all else fails I could write some tool to “hack” the files into their sound parts myself :slight_smile:

Export Multiple is not currently supported in Chains.

Labels can be exported
but that is not supported in Chains either.

How long are each of the sound files?

“Sound Finder” (and similar tools) has a reputation for not being 100% accurate (because it can easily be thrown off by background noise). How important is it that this process is 100% accurate for all 200+ files?

Ahh that is really sad to hear, Audacity is so close to solving the problem :frowning:

The sound files are between 2-8 seconds, each one containing 4-5 sounds (sometimes just 3)
They are already pre-processed so there is no background noise, I’m confident that the sound finder will be able to handle it since the files are well prepared.

So how about something like this:

  1. Import all 200+ files (about 20 mins total?)
  2. “Tracks menu > Align tracks > Align End to End”
  3. “Tracks menu > Mix and Render” (mix down to one track)
  4. “Analyze menu > Silence Finder”
  5. “File menu > Export Multiple”

Hmm, that would get rid of the original file names.
However I absolutely need the original names + number.

It may help us to help you if you tell us what you are actually doing. What are these 200+ files and what / why do you need to do to them?

They are sounds for a game I’m making, each file contains a few variations of the sound.
For example SwordHitLight.wav contains 4 different takes of the sound, so the game can chose to play one randomly.

The files are SwordHitLight.wav, AxeHit.wav, FootstepGrass.wav, FoostepSand.wav, …

I need to split them into their individual sounds, but I need to maintain the original names so they’re visible in the file-explorer and can be searched for.

There will be more sounds coming soon so I’ll likely have to redo the steps then for the new sounds (they get put into the same folder, so I’ll likely just do it again for all of them instead of just the new sounds).

OK, so a variation on the steps that I suggested previously, using a modified version of Silence Finder (which I have called “Silence Finder Mod…”)
ModifiedSilenceMarker.ny (4.31 KB)
Instructions for installing this plug-in:
I’ve only tested this version in Audacity 2.2.0 beta, but I think it should work in Audacity 2.1.3


  1. Import all 200+ files (about 20 mins total?)
  2. “Tracks menu > Align tracks > Align End to End”
  3. “Analyze menu > Silence Finder Mod…”
  4. “Tracks menu > Mix and Render” (mix down to one track)
  5. “File menu > Export Multiple”

The modification is that the labels have the name of the audio track that is being labelled. Export Multiple is able to use the label text for naming the exported files.
You may need to tweak the procedure to achieve what you want.

Seems to work really well.

However I’m surprised that this works. I thought I needed the sound finder, not the silence finder?
The silence finder perfectly places the markers directly where the sound starts.
But why? I mean I’m just confused by the wording. Why would I want to “find” the silence instead of the sound? :smiley:

Thank you very much for that script and idea, really saved me a ton of work.

But just out of curiosity, why can’t we use export multiple in chains?
It seems to be saving its serialized settings somewhere, so why not use that as input parameter (the parameter being the settings file-filename to load)?
I suppose it has to do with how the mods are generally made to interact?

Oh and also ,that mod you made seems to be really really useful for a lot of cases. When would you ever not want to keep the original name? Shouldn’t that be the “default” implementation of that plugin?

Anyway thanks again for solving this! :slight_smile:
I’ll give some more feedback once I have time to work on the full set of files. (I just tested with ~10 right now)

There is one problem.

Since the silence at the beginning of the file is too short, it doesn’t place a label.
Can I add silence in front of every track? It doesn’t seem to work when I have multiple tracks open (or maybe I just don’t know how to do it correctly).