wilsjo2/OptiScaler-DLSSNR-PreSR-Multipass display-filter-v0.1.0-preview

Display Filter v0.1.0 preview

on GitHub

Standalone Display Filter preview for Windows 11. This is the latest locally built display_filter.exe, separate from the in-game OptiScaler releases.

Download and run

Only display_filter.exe is attached. No DLLs are bundled.

1. Save the executable in its own folder and run it. Administrator access is not required.
2. Select your own compatible nvngx_dlssnr.dll in the app. Alternatively, place it in a runtime subfolder beside the executable.
3. Select a screen or application window and start capture. Minimize the controls or switch back to the source to see the effect.
4. Press Ctrl+Alt+End to stop capture.

Private DLSS upscaling additionally requires your own nvngx_dlss.dll, beside the selected NR DLL or in the application's runtime folder. Bilinear upscaling does not require this additional DLL. The Microsoft Visual C++ v14 x64 Redistributable may be needed if the executable reports a missing MSVC runtime.

Included

• Whole-screen or window capture, SDR/HDR output, and a click-through overlay.
• One or two NR passes, adjustable reference white, model intensity/colour and skin controls.
• Matched-residual transfer, adjustable model resolution, and bilinear or private DLSS upscaling.
• Experimental motion estimation.
• Adjustable processing limit: 1–360 FPS; 0 means uncapped. Actual throughput still depends on capture rate and GPU time; desktop presentation follows display refresh.
• Removed the application's signature check on the selected NR DLL. Driver compatibility requirements still apply.

Settings and logs are stored in %LOCALAPPDATA%\OptiScalerNR.

Validation and limitations

Headless GPU checks passed on an RTX 5090, including zero-strength SDR/HDR identity, two passes, private DLSS, and motion estimation. Live presentation and game verification remain for user testing. Depth and motion guides differ from engine-provided data, and the reported image-quality gap versus in-game NR at matched settings remains unresolved.

This binary release does not update the existing OptiScaler game builds or source branches. GitHub's automatic source archives represent the repository tag; they do not contain the standalone local source used for this executable.

Executable SHA-256:

55D3B78AE6CF4CB369E224F53562D9B91A6F6DC2252A53E861A8A024B8D69FA5