Hi, Kyle! So, interestingly enough, the last time I opened it, the error was not there. It did show up immediately after the latest update, but now it’s not. I’ll have to keep monitoring this. Thanks.
1 Like
When I have time, I’ll recreate the error using a video file that’s not labeled and send a report. But it’s ridiculous that there have been four updates since version 1.0.4, and all of them crash during rendering. I designed my rig for the sole purpose of using Topaz Video Ai. So it’s incredibly frustraiting I can’t use the new features I’ve been waiting to try.
Looking into this, while we do this can you uninstall the current installation and then download a clean installer and see if that works better? Could have been a corrupt installer or an issue with the download.
Depending on the hardware you are running and the models being used, with higher resolution files this could be a similar case to what we have seen with other users who run over clocking systems, or have XMP/EXPO settings enabled on some CPU’s which can be tripped and cause crashes and reboots. Cant say for certain without looking at the logs and system profile for more information to help diagnose though.
I’ve downloaded thrice (albeit during a time frame of about only 15 mins or so) - always the same.
The PKG installer worked.
I will try again later.
1 Like
Thank you as always.
I’ve been using it since the beginning, but this is my first comment.
1.2.0 works fine, but the app won’t launch in 1.2.1.
I tried reinstalling 1.2.1, but it still wouldn’t launch, so I installed 1.2.0 and it started up.
Can you email the support team with your system profile, any screenshots that show any error messages, if you are getting any, so we can try and sort out why the app is not launching for you.
How To Obtain Your System Profile (Windows)
How To Obtain Your System Profile (Mac)
@kyle.topazlabs @ananth.topazlabs
New NVIDIA STUDIO DRIVER Release:
ComfyUI Streamlines Local AI Video Generation | NVIDIA Blog
CUDA 13.2 Introduces Enhanced CUDA Tile Support and New Python Features | NVIDIA Technical Blog
Any news about performance improvements with the next updates of Topaz Video, especially for improved Starlight model with NeuroStream technology? 
5 Likes
Devs and the research team are working on improvements as they can. NueroStream will only be a part of Topaz Photo at this time.
I don’t mean to be disrespectful or rude, but they haven’t managed to make or optimize the Starlight Mini 1% faster in a year, yet they’ve developed 10 other models that are faster and better, but only for the cloud. The problem is their priorities, which is why I will no longer support them financially.
7 Likes
Disappointing to hear this. There are many different image upscalers out there. Topaz Photo is great, but there are many alternatives that are good enough.
Topaz is the leader in longform video restoration. That’s your fastball. I wish you guys would focus more on your best product. I have to imagine any movie studios using Topaz Video are requesting speed improvements, as the current speed is very slow.
9 Likes
They’re probably only faster in the cloud because the hardware can run them fast. They would still run slow or even slower than Starlight Mini if we were able to run them locally.
Yes, but they’ve been promising us for months that they would optimize the Starlight Mini model for us, and that’s been since last summer, and what has happened to this day? Nothing.
5 Likes
Since release no SLM progress. You can tell me whatever you want, for me it strongly suggests unwillingness because of the cloud + there are things you could do even without model improvements, such as more flexibility/free resolution selection for Starlight but it does not happen.
2 Likes
Still observing issues (as with past versions) with version 1.2.1 with the infamous “Recoverable Error”. You know, the one that never seems to recover no matter how many times you retry it. Pulled the log file on recent issue with this and extracted the following:
2026-03-11 08:17:46 28680 WARNING: Checking file “C:\ProgramData\Topaz Labs LLC\Topaz Video\models\audio-codecs.json” failed with error Unknown exception
2026-03-11 08-17-46.740 Thread: 31548 Info FF Process Output: 3
2026-03-11 08:17:46 28680 WARNING: ModelManager: Ignoring model: C:\ProgramData\Topaz Labs LLC\Topaz Video\models\benchmarks.json
2026-03-11 08-17-46.743 Thread: 31548 Info FF Process Output: 3
2026-03-11 08:17:46 28680 WARNING: Checking file “C:\ProgramData\Topaz Labs LLC\Topaz Video\models\model-recommendation-rules.json” failed with error Unknown exception
2026-03-11 08-17-46.745 Thread: 31548 Info FF Process Output: 3
2026-03-11 08:17:46 28680 WARNING: ModelManager: Ignoring model: C:\ProgramData\Topaz Labs LLC\Topaz Video\models\proxy.json
2026-03-11 08-17-46.745 Thread: 31548 Info FF Process Output: 3
2026-03-11 08:17:46 28680 WARNING: ModelManager: Ignoring model: C:\ProgramData\Topaz Labs LLC\Topaz Video\models\ref-1.json
2026-03-11 08-17-46.746 Thread: 31548 Info FF Process Output: 3
2026-03-11 08:17:46 28680 WARNING: ModelManager: Ignoring model: C:\ProgramData\Topaz Labs LLC\Topaz Video\models\thm-1.json
2026-03-11 08-17-46.747 Thread: 31548 Info FF Process Output: 3
2026-03-11 08:17:46 28680 WARNING: Checking file “C:\ProgramData\Topaz Labs LLC\Topaz Video\models\video-encoders.json” failed with error Unknown exception
1 Like
Please pull the full logs and send in to the support team so we can have the devs look into this further. help@topazlabs.com
1 Like
Hi guys, is it no way we can resume the output image sequence from where it stop? I didn’t see the pause button as in video output so i just hit stop the process in half way, then put the PC into sleep and hope to restart from where it left off on next day, but i notice it just start again from beginning although the PNG file that i exported still inside that folder. This is counter intuitive to output as an image sequence.
This PNG resume feature was working in earlier versions, but maybe not anymore. I also had an issue where the resume function when using a video codec (in my case FFV1) failed.
I tried it several times because some of my renderings crashed and stopped. Then Topaz Video offered to resume from the last checkpoint. I clicked it, and a few hours later the output video was unusable. The sequence from the end was placed at the beginning of the video, and half of the video consisted only of black frames. Because of this, I no longer use this function.
I also noticed that the Add Grain option in version 1.21 sometimes is not executed, and the rendering finishes without grain. I went back to version 1.20. This could also happen there, but less often.
1 Like
this is nightmare. Yeah, noticed on the video resume as well, i can’t see it work reliably since it 1st introduction.
1 Like
Stopping a process, ends it and clears the portion that was written to a temp folder. That is why it started back from the beginning of the file when you launched the app the next time.
Checking on the image sequence export aspect with the team to see what is going on there.
