Empty Project is ~400mb big. How?

Hey people,
I have some Cubase projects, which got really big over time. But even when deleting all tracks with all plugins, it still has nearly the same size. This tells me, that there must be any issue somewhere else inside the project, which causes this size. In my example, which you can download down below, the project got nearly 400mb (!!!) big.

Generally I wouldn’t care that much, but the increased size leads to incredible long saving times, export/import times and makes the entire project feel way slower, than smaller ones.

Has anybody an idea perhaps, what could cause this and help?

UPDATE: The Audio Pool is causing this it seems, as if I begin to empty it, the size gets smaller, but it doesn’t seem to get smaller linearily, it seems to be certain links to files causing more of an increase than others, but I have a hard time getting to know, which ones especially. But still there is no link to any file in the project, as it is completely empty. How can even a lot of links, cause a project file to get that big?

PS: The zip is 90mb, if you extract it, it will be around 370-380mb.

Link to file: 89.94 MB file on MEGA

Did you use ARA in any way? That is the only time my projects blow up in size. Even removing any audio that is associated with the ARA files didn’t make the project smaller.

Try emptying the Pool’s Trash - it has almost 8000 files referenced. It may take a while (several minutes on my NVMe SSD) and Cubase may seem frozen, although it’s not.

1 Like

Nah unfortunately not. Using Variaudio only!

This was just an example project, in reality I use most of the referenced audio files in the pool, so I am trying to understand, which of those increases the file size that much, and how such an increase is possible, as these are just “links”?

by comparison: a text file containing 1,000,000 letters is about 1mb big.

I don’t have inside knowledge about Steinberg’s proprietary file formats. And I also doubt that they would be eager to share that info.

But I would assume that it’s not just a list of 8000 links, but a list of 8000 database entries containing anything that Cubase may need to know about the audio files in the pool, including anything that is a non-destructive edit. And that would seem to have to include every detail of VariAudio data. So each of those 8000 database table rows, might effectively have to include many rows (for example, one for each VariAudio operation and/or datapoint). And whatever other non-destructive audio edits there may be on a file. Like when an audio FX plugin has been applied to an audio event.

If you’re genuinely curious enough, you could look at the CPR file with an editor that is capable of handling very large files (I used VS Code) and see how much binary data seems to be around each of the regular text (thus readable) file names.

And/or run simple tests, starting with an otherwise empty project and a single wave file and then measure how much each (or several) VariAudio edits add to the .cpr file size. And maybe also try other non-destructive type of edits that Cubase would need to keep track of.