XRNS file size limit / bottleneck question

Hi, I use more and more hardware synths, drum machines and mic recordings etc in my projects these days, and my XRNSes often land on typically 300-400 mb. And I’m just curious - exactly what determines how large an xrns can be? Is it basically just my own system (ram, cpu, drive space) or does Renoise have its own bottleneck in this scenario? Using Win7 64-bits (NTFS), maybe even Windows has its limitations?

Would be interesting to know, especially before I start planning any really huge projects (like a live performance or a 80 min epic ambient journey or something :stuck_out_tongue:)

ok, guess it’s hard to tell… I’ll try to just simulate a ridiculously large project and see how it runs.

Regarding the xrns file format, I would like to suggest to offer an option that xrns files can be also saved unzipped. In macos, it is common that specific file structures come in kind of “bundles”. It looks like a single file, but actually is a directory containing lot of different files (e.g. the common .app. I think it is defined in the Info.plist, if a directory appears as file or directory). Since zipping huge flac files surely isn’t speeding up saving process and the amount of saved bytes surely is negligible. Also it would be much more easily manageable to apply song.xml fixes, plus you could directly index and load the song sample files into sample managers. Loading and saving songs could be quite drastically speeded up, esp. when using a SSD.

zipping song files + samples should be considered as outdated concept, IMHO.

1 Like

You can choose your preferred audio compression method that is used when saving:

Preferences > Files: https://tutorials.renoise.com/wiki/Preferences#Files

When using Flac, the already compressed audio is not recompressed a second time when saving the XRNS, rather it is simply added directly to the ZIP archive using the “store” method at compression level 0.

Likewise, uncompressed WAV will also be “stored” directly into the ZIP archive with no compression, and should in theory be the fastest method if you have a nice fast drive and system.

I personally haven’t done a lot of thorough testing to benchmark each method, but my songs do not tend to get very huge in most cases, so the default Flac method has always worked well enough for me.

As always, your mileage may vary!