If you’re outputting image sequences (ie, one image per frame), Topaz Video will typically write a filename in the format of [source filename]+[model]+[number].
You would expect that [number] represents a frame number from the video source, but it has absolutely no relationship. It is simply an incrementing number from 0 that applies at the start of every export, regardless of the frame.
Now - one of the reasons I use image sequence exports is because on my 12GB VRAM system, video exports from Starlight Precise 2.5 can take a LONG time to export (days for multi minute videos). That’s ok, but sometimes I want to do something else on my system. With no pause functionality for SP2.5, that’s not an option. If I’m 95% through a video export that’s taken 2 days and I want to use my GPU for something, I have to cancel the video and start again from scratch.
The workaround is to render out image sequences - then I can just stop the export and resume it later from the last frame exported. Let’s say I stop at frame 500. I can set the Mark In to frame 501 and start the export again. BUT if my last filename was “example_slp_00050.png”, the first frame to be generated with the new export in the same folder will try and write again to “example_slp_00000.png”, find that that file already exists and instead, write it as “example_slp_00000.png_Copy(1)”. Ok, that’s now running but if you have something crash and you need to resume again? If you mange to figure out what frame you were up to and kick it off from there, you’ll now have “example_slp_00000.png_Copy(2)” in your folder as well. You end up with a complete mess that makes it very hard to stitch the video back together.
Topaz Video should handle this automatically - it should be writing filenames based in the frame being rendered - this would completely remove the problem.
I imagine there’s an argument before about whether this is an “improvement”, “bug” or “issue” - frankly, as a user, i don’t care. I just want to see it fixed. It’s maddening.
