Recoverable Error during Rendering

Video AI Bugs and Issues
1

Description:

I am experiencing recurring recoverable errors during rendering.

The issue occurs primarily when I try to upscale footage that was previously upscaled to 1440p using Starlight Sharp, and then upscale it further to 2160p (custom resolution).

For this workflow, I am using:

  • Iris Medium model

  • Chronos Fast at 24p

  • Hyperion for SDR to HDR conversion

System Specifications:

  • CPU: AMD Ryzen 9 9950X3D 16-Core Processor (4.30 GHz)

  • RAM: 64 GB

  • GPU: ASUS GeForce RTX 5090 ROG Astral Gaming OC

  • Driver Version: 595.97 (Game Ready)

  • OS: Windows 11 Home 64-bit

  • Version: 25H2

  • OS Build: 26200.8037

Topaz Settings:

  • Maximum processes: 1

  • VRAM usage: 100%

  • Export format: ProRes 422 HQ

  • Container: MOV

  • Audio: Copy

  • Crash Recovery: Fully enabled

Troubleshooting Steps Already Attempted:

  1. Tried different models (GAIA, Proteus, Artemis, Apollo)

  2. Changed output formats (e.g., H.265, converting audio instead of copy)

  3. Performed a complete reinstall of Topaz and GPU drivers (including full registry cleanup)

  4. Installed Topaz on drive C:, while all other data is stored on drives D: and T:, each with separate subfolders

  5. Additional minor adjustments and tests

The Windows 11 installation is also relatively fresh (around 2 months old).

Additional Information:
No log files are being generated for these errors.
In the corresponding folders, I can only find “Main.tzlog” files.

2

I’m having the same issue i think it’s a General Bug. Hope they fix it soon.

3

The log files include the date followed by the main.tzlog so what you are looking at seems to be correct. Could you share those log files so we can take a look?

To gather logs, please open the app and select Help > Logging > Get Logs for Support and attach the zip file to your reply.

Feel free to email in at help@topazlabs.com

4

After extensive testing, I was able to pinpoint the root cause of the issue.

The recoverable error occurs when the “Hyperion” model is used alongside any other models. It doesn’t matter whether it’s Proteus, Iris, Gaia, or others—the problem consistently appears in mixed usage scenarios.

Hyperion only works reliably when used completely on its own. In fact, it’s so unstable that errors can even occur if more than one project is present in the export queue.

TL;DR: Since at least version 1.3.0, Hyperion has been highly unstable and should only be used in complete alone.