FunscriptDancer - generate Funscripts from music beat/pitch/energy, cross platform

I ran into this problem. I had a command window open with admin privileges, since running juliaup.

Somehow, I ended up installing sonic-annotator and the vamp plugins after that point, then added c:\Program Files\sonic-annotator to the path. Ran the funscriptdancer.bat file, watched it grind away for 30 minutes, open the app, and it failed at the same point.

The environment variables only get loaded in that instance of the command prompt when that window opens. I opened another command prompt window, and checked the path, and saw the sonic annotator was there, so that left me trying other stuff for a while.

It even ran sonic-annotator -h from a command window, so I knew it was installed and the path working.

I typed ‘path’ in the window that funscriptdancer.bat was waiting, and it didn’t show, as well as the attempt to run sonic-annotator from that folder. Opening a new command prompt (run as admin), and then it worked fine.

Another thing: Make sure there’s a trailing backslash on the path entry. Sometimes it will fail if it’s just C:\Program Files\Sonic-Annotator
and not C:\Program Files\Sonic-Annotator\

Ignore my previous edits to this page. Has anyone found a good way to make different kinds of intensity to the scripts? So like good basic settings for low intensity or high intensity?

best regards!

Julia updates:

Every so often when it runs it says there’s an update to Julia. Just dropping out to a command prompt and running “juliaup update” like it asks seems to do the trick, but then it warns that other things were compiled using a 1.0.9 dependency. Are there ways of updating the other files to the current rev of Julia?

Scary warning aside, it does run, so, good enough for now, I guess.

Trying to follow along what it’s doing in the command window is a little weird as is garbles some of the text after it has FFMPEG extract the audio to a .wav file. This seems to be harmless, because eventually it dumps the waveform back into the FunscriptDancer window to manipulate and export.

I get an error with larger files, and I’m trying to determine if there’s a way to allocate more resources to handling the file, or if there’s a known upper limit. (Maybe by reducing the sample rate by half, or simply splitting the video file into smaller pieces first?)

The error happens after FFMPEG has extracted the .wav. Somewhere around 99%, the Vamp-aubio:aubiotemp:beats" processing fails.

"CodedAudioFileReader: Caught bad_alloc when trying to add 131072 elements to buffer. Error Failed to process file “tmp\Decendant [strothy 2023].x264.wav”.

The next process, involving “vamp:vamp-aubio:aubiopitch:frequency” seems to pass without a problem, or at least it gets to “Extracting and writing features … Done”

Also: you need to go in and clean out the funscriptdancer\tmp folder every now and then. Funscriptdancer removes the audio files (if it doesn’t crash) but the other .CSV files will accumulate.