Issues syncing projects between machines

I have a reasonably big project (40 tracks) that I work on from two locations. Each location syncs the file set (Project.aup, everything in Project_data folder tree) to the same cloud drive. Nobody else works on it, and it is supposed to work in this sequence:

  • go to work
  • sync from cloud to work machine
  • edit the project at work
  • sync to cloud
  • go home
  • sync (from) cloud to home machine
  • edit the project at home
  • sync to cloud
  • rinse and repeat…

All has been running OK, but when I load the project sometimes I get warnings about orphan files that it has found and what do I want to do with them? Because the warning is worded in a warm and fuzzy way I just allow it to delete these files and afterwards the project seems to load as normal. I figure the orphan files result via Audacity’s use of the .au files - when data inside file is edited, the edited data is written to a new-named .au file and the old one is deleted. Because the old one is only deleted locally and not on the cloud copy of the project, the next sync operation reinstates it. Over time the cloud copy will become full of orphan files and I will need to consent to deleting more and more orphan files at startup.

Is this how it works? because I have just had a worrying glitch :frowning: ?

Today I sync’ed from the cloud to work and my sync utility (Syncovery) reported several conflicts where the files had been modified on both sides since the last sync. All the files on the cloud were newer, so I copied them from the cloud to my work PC.

When I started Audacity it reported ophan files as usual which I cleaned out, however, it then reported lots of files missing and unrecoverable. Now it appears my work version of the project is broken, and my home version is not.

What would be considered best practice for implementing this “work-from-home/work-from-work” arrangement?
Is the file set completely defined by the filenames in the XML .aup file? I.e. can an image of Project.aup and Project_Data*.au be considered as a complete backup of a project?

I’m coaching a bigband with 20 individual instruments on 20 tracks. We also record on several locations and share files on an server.

As we do not trust the stability of sending the AUP-format with thousands of files, we decided to EXPORT the tracks instead of sending the whole project. This means that the main files is on one computer and the cloud only keeps a copy and the changes.

we do it this way:

Work “at home”
Save the AUP-Project as “Lossless Copy Of Project”. This will produce a folder with 40 WAV-files, one for each track. Now you can zip all WAV-Files into one ZIP file and upload them. disadvantages: In this case all ENVELOPES you allready set in the tracks will be rendered to real signal changes. A track with a lot of part recordings will become one long track with silence between the part.

Work “outside”
Download the ZIP and open a temp. project. Add new tracks for new work. At hte end only export the new tracks, where you worked on and upload it. At home you add the files to the main project.

Thanks for your input. What you are outlining might be suitable, I’ll need to give it some thought.

With my current setup (when it works properly!) I have scheduled tasks on the 2 machines that perform the sync to/from the cloud, so that I can pick up from where I left off on either machine without any manual steps - i.e. I edit the project at home and overnight the project (changed files only) is saved to the cloud from the home machine and restored from the cloud to the work machine. When I get to work I open the project and it is just as I left it at home.

The major side effect of this scheme is you seem to get orphan .au files accumulating, which I gather are files in the Project_data folder that aren’t referenced in the .aup file anymore. When the issue occurred it seems that I had files that Audacity detected as orphaned, but which were referenced after all. Maybe this was a result of the machine having a .aup file that was freshened without the .au files being freshened also - I don’t know.

Your method is essentially defining the project in terms of a set of WAV files, but as you point out, you lose amplitude profiles (I don’t use these in general) and any other project-related stuff - effectively you are recreating the project each time. If Audacity had a reasonable set of command-line options it would be more feasible to attain my goal of hands-free project sync between the home and work machines.