Key Features and Components
The NVIDIA Real-Time Denoisers is a fast solution that uses input signals and environmental conditions to deliver results comparable to ground truth images.
Release NotesNRD V. 2.9:
- DLSS integrated into the sample with dynamic resolution scale support
- Introduced new denoisers for ambient and specular occlusion
- Introduced stand-alone ReLAX diffuse-only and specular-only denoisers
- Added new functionality - advanced pre-accumulation filtering pass
- Added new functionality - user-driven antilag (user provided history confidence)
- Added helpers for better many RPP support
- Added optional material ID aware spatial filtering support
- Improved white-box integration support (UE4, consoles, custom engines)
- Meters to units conversion now happens inside the library
- ReBLUR and ReLAX: IQ improvements
- ReBLUR and ReLAX: now compute local curvature to improve specular tracking
Already accepted into the NVIDIA NRD access program? Click here to visit the GitHub repo.
NOTE: We will be asking for your GitHub handle so we may grant you access to the RT Denoiser Source code via directly adding your Github Account to the code repo. We are asking for your email address so we may follow-up with any questions during the approval process and potentially to discuss the project you’re working on with the RT Denoiser. By submitting your request, you consent to the collection and use of your GitHub handle and email address for the purposes previously specified in this paragraph.