Not sure if this is a bug or issue but I’ve noticed rendering times are not very consistent.
For example; earlier today I upscaled a 1280x720 23.976 fps film at 1:20:56 length Using GAIA HQ to 1920x1080 and it took approximately 4 hours (about 7fps).
Now I have looked at doing 2 others same resolution and fps with times of 1:20:36 and 1:26:36 and it’s telling me for both of these using GAIA exactly the same as 1d 11 hours plus…
It’s the same PC using the same drives so can you explain why the massive time difference??? It seems bizarre. I flushed RAM and restarted the program but same thing. Tried another file same settings estimated 4 hours 36 minutes. Looks like some sort of bug to me.
Can you ensure that the frame count for these files are the same? That does seem quite odd since you also rebooted the machine in between processes.
Are there any other differences in the exports?
Didn’t reboot just closed the program and re-opened it, and the frame counts weren’t wildy different to justify the time difference. Obviously I closed all renders giving me those times after closing the program again and trying a 3rd file also around the same length with same frame rate it reported an estimate closer to the 4 and a half hours again.
Can you share the exact reproduction steps for each export as well as the files used and your logs?
You can securely submit your files(s) to our Dropbox using the link below. Please be sure to send me a note to let me know you sent something.
Submit File to Dropbox
Unfortunately not I’ve done multiple other renders since then. As I stated previously all setting for the files were EXACTLY the same: same fps, same upscale from and to and same model, trying different files with runtimes only varying by a few minutes. There were no other programs taking valuable resources in the background it just reported oddly weird render times 3 times in a row and I obviously cancelled al 3 attempts. The following day the files were back to taking 4-4.5 hours.
That is very odd!
Since we did not receive other reports of this, hopefully this was an isolated event that you will not run into again.