As the title says…Topaz is busy doing a 1080p to 2160p upscale using Proteus. During processing, the UI simply closes. No error, no notification of anything. However, the ffmpeg process keeps running to completion successfully, it just doesnt do any post processing so I have to rename the output file manually. Logs attached, but the relevant one (2026-03-20-08-11-47-Main.tzlog) just ends in the middle of output logging. No error there either that I can see.
logsForSupport.zip (22.5 MB)
Can you share the system profile report as well for review to look into system crash errors?
How To Obtain Your System Profile (Windows)
The logs ending abruptly point to a system shutdown issue.
Open the Preferences Panel and lets try making a change to the following settings.
File>Preferences>Processing
Ensure that your GPU is selected for the AI Processor
Set the max memory to 80%
Set max processes to 1
Save the settings and close the panel.
Re-run your clip through the process that you were trying.
Here you go…
DxDiag.txt (141.9 KB)
Those are my default settings already. Again, the ONLY thing that shutdown was the UI. The actual video processing in ffmpeg continued and completed successfully an hour later.
When I restarted Topaz Video, it attempted to resume the upscale from the point where the UI died. BTW, I had previously had a Starlight Sharp process running that took 2 days to complete and Ui stayed up the whole time. So far, the UI has disappeared like this several times, and always when running any ffmpeg based upscale. it has never happened during a Starlight process.
Edited to add:
The machine was dedicated to Topaz…I did not use the PC for any other purpose while the Topaz process was running.
The DxDiag report is showing that the GPU card looks to be set to run with overclocking settings, let’s turn those off and lower the max memory to 80%. The previous command lines show it was running at 100% which could be part of the issue here as there is no overhead room for spikes in memory usage.
FFMPEG runs differently than the Starlight processes, and despite the Starlight models being the heaviest we have it might complete without issue while running FFMPEG and an older GAN model could be tripping the system and causing the app to hang and crash.
Honestly, I’d rather live with the UI crashing than turn off all overclocks on the GPU. That reduces the Topaz processing speed by about 20%. I have a Gigabyte RTX-5080 Gaming OC GPU and have monitored VRAM use closely during all processing. Starlight will use a max of 14GB out of the available 16GB. I have never seen VRAM usage exceed 4GB during a Proteus run. That is what I see when using HWInfo, GPU-Z, or HWMonitor. Starlight will push my system RAM use to about 40GB of system RAM when running Starlight (although that appears to be almost all cache, the CPU load is miniscule). The other models never use more than about 4GB of RAM or VRAM. I just can’t believe this is a resource problem.
However, I’ll give it a try and report back.
Okay Kyle..here is what happened:
Set the GPU to default and max memory in Topaz to 80%. Started a Proteus upscale run and the UI crashed after about 10 minutes. I manually stopped the lingering ffmpeg process and rebooted. I then tried to restart Topaz Video but now the UI crashed immediately. Assuming that something may have been corrupted by the crash I decided to start clean by uninstalling Topaz and the NVidia drivers. After reinstalling Topaz started fine and I was able to download both Starlight models. I selected a file for processing but when I tried to select Proteus as the enhancement model the UI crashed again. On a second try doing the same thing, it crashed again. Next I tried selecting Artemis instead of Proteus and that worked. Then I tried selecting Proteus again and it still worked. I am running the Proteus process now and everything seems to be working as expected. I forgot to change the max memory after the reinstall, so this run is still at 100% VRAM use in preferences. I’ll update this when I have more info.
Very odd behavior of the app and it being centered around the Proteus model. Keep me updated and send any logs, screenshots or crash reports to help@topazlabs.com so I can get them over to the devs for review into the matter if it crashes again on you.
The Proteus upscale completed and the UI stayed up the whole time. I’m away for the rest of the week, so I won’t be able to do any further tests until Tuesday next.
Had a tough time today with the 1.3 update. I downloaded and installed the update, after which the app window shutdown immediately upon launch. I rebooted the PC and tried again, same result. Uninstalled Topaz Video completely, rebooted and reinstalled. Now the Window stayed up, until I clicked on anything, and crashed again. I uninstalled again and this time did not ‘save settings’, rebooted, reinstalled. Now I was able to get into the program, download and activate Starlight Mini and Sharp. Started a SLS upscale, walked away and when I returned the UI was gone again with the Runner still working. I restarted Topaz Video without stopping the Runner and collected the logs that are attached. Afterwards, the upscale threw an error in the Export Queue but the Runner is still going.
DxDiag.txt (142.1 KB)
logsForSupport.zip (6.5 MB)
It looks like the overclocking issues are still occurring and crashing your app. I would go back and turn those off for a few test runs with the same sources and settings and see if that resolves the issues you are seeing.
My gpu was not overclocked. I set it the default settings for the card. The only other overclock is the memory controller, which is bumped from the base 2700Mhz to 3200MHz via the XMP settings attached to my DDR5-6400 memory. I may just rollback to version 1.1 since I never had this problem then.
Found the problem…it had nothing to do with overclocking or resource problems. The ASUS GPU Tweak program had its on-screen display setting turned on. Since Topaz Video activates the NVidia overlay, I assume it does the same for the ASUS overlay. The ASUS OSD never showed up, so I didn’t even realize it was turned on. I’ve ditched GPU-Tweak and switched back to Afterburner and I have an SLS session that has been running perfectly for a couple of hours.
My assumption is the GPU-Tweak was trying to access VRAM to add the stats and collider with Topaz Video’s access.
Interesting, I will have to remember this for the future with other users to check for this situation.
Thanks for sharing, and glad that you are back to working with the app and its running as expected.
