{"id":1674130,"date":"2026-08-25T15:30:00","date_gmt":"2026-08-25T15:30:00","guid":{"rendered":"http:\/\/oPKWLPLiAYvWTuLfsT5jY5"},"modified":"2026-08-25T15:30:00","modified_gmt":"2026-08-25T15:30:00","slug":"dlss-4-5-ray-reconstruction-is-a-definite-improvement-but-path-traced-fidelity-is-still-not-a-100-solved-issue","status":"publish","type":"post","link":"https:\/\/arcader.org\/news\/dlss-4-5-ray-reconstruction-is-a-definite-improvement-but-path-traced-fidelity-is-still-not-a-100-solved-issue\/","title":{"rendered":"DLSS 4.5 Ray Reconstruction is a definite improvement but path traced fidelity is still not a 100% solved issue"},"content":{"rendered":"
\n

Nvidia’s new DLSS 4.5 Ray Reconstruction feature is laser-focused on trying to fix the second most annoying thing about ray traced or path traced games: visual artifacting. Yes, the lighting effect does deliver the most accurate-looking lighting and shadows in games, but in motion there is a definite visual weirdness added into scenes. And that weirdness generally comes from the denoising pass designed to turn the low-fidelity path traced image, with its limited number of rays, into a fully realised frame.<\/p>\n

Leave a game’s viewport still and the scene will pretty rapidly come together to make a sharp, convincing image. Shift that around, or have that scene filled with moving objects or characters and you’re going to see the sort of visual trails you might only have been used to in the warehouse parties of your early twenties. Just me?<\/p>\n

The issue is that it’s way too expensive in terms of raw compute to trace every single photon\/pixel that would otherwise make up a fully traced scene. Even with the current method of tracking a far smaller representative sample of rays bouncing around the geometry of your game world it’s still incredibly demanding of your PC’s hardware.<\/p>\n

This is, by the way, the most annoying thing<\/em> about ray traced or path traced games: the amount of graphical horsepower the tech demands over pre-baked lighting techniques. The huge performance hit it involves is punitive for all but the most powerful GPUs, and we are still no closer to solving that right now.<\/p>\n

So we still have to use small numbers of tracked rays to create an approximate scene which looks something like this:<\/p>\n

\n
\n
\n

\"Nvidia<\/a><\/p>\n<\/div>\n<\/div>

(Image credit: Nvidia | CDPR)<\/span><\/figcaption><\/figure>\n

But that doesn’t look great if you’re trying to make a convincing world, though I guess there’s something to that aesthetic in a sort of 3D ASCII kind of way<\/a>. Anyway, to realise this image into a production ready frame you need to use a denoiser to fill in the blanks between the sampled rays, and the hand-tuned denoisers track and group pixels across multiple frames, to give them access to more samples. They then use these neighbouring pixels to interpolate and blend the scene together.<\/p>\n

As I said, that’s fine for a static image, but those pixels don’t stay representative for long and in motion they break down, adding in artifacting to a scene and smearing detail away from textures. This is where the original Ray Reconstruction feature<\/a> came in, replacing the hand-tuned denoisers with an AI-based denoiser working alongside Nvidia’s upscaling models, to generate more accurate pixels in between those sampled rays. <\/p>\n

And now we’re finally getting an updated version of Ray Reconstruction which uses the sharper second-gen Transformer model initially brought into DLSS 4.5 Super Resolution back in January<\/a>. <\/p>\n

\n
\"Nvidia
Nvidia<\/small><\/figcaption><\/figure>\n
\"Nvidia
Nvidia<\/small><\/figcaption><\/figure>\n<\/figure>\n

And it’s about time, because for ray or path traced games, even if you were using the improved transformer model of DLSS 4.5 Super Resolution, Ray Reconstruction was still using the old model, leaving you those tell-tale artifacts.<\/p>\n

Nvidia is touting four main improvements with DLSS 4.5 RR:<\/p>\n

    \n
  • Efficient Denoiser: <\/strong>The new model delivers 35% more compute capability, and processes 20% more parameters, while maintaining similar performance to the previous model.<\/li>\n
  • Enhanced Super Resolution:<\/strong> Building upon the advances from DLSS 4.5 Super Resolution, the model has deeper spatial awareness across every scene, and more intelligently uses game engine pixel sampling and motion data. The result is improved lighting accuracy, better temporal stability and clearer motion in ray-traced and path-traced content.<\/li>\n
  • Expanded Training Dataset: <\/strong>Trained on a larger dataset, the new model is even better at image reconstruction. This intelligence gives the model even better awareness to pick the most accurate engine data to reconstruct scenes closer to ground truth.<\/li>\n
  • Finer Developer Control:<\/strong> The new model provides developers with finer control for temporal accumulation, providing precise tuning of model response for even better image quality.<\/li>\n<\/ul>\n

    The new version of the feature is releasing today, and jamming the second-gen transformer model into supported games is done via the Nvidia App<\/strong> rather than from inside your games’ settings screens. You just need to select a title from within the Nvidia App itself, navigate to the ‘DLSS Override – Model Presets’<\/strong> entry and select ‘Preset F’<\/strong> from the ‘Ray Reconstruction’ dropdown menu<\/strong>. Apply it and you should be set.<\/p>\n

    \n\n\n\n\n\n\n\n\n\n\n\n\n
    DLSS 4.5 Ray Reconstruction supporting games<\/caption>\n
    \n

    Alan Wake 2<\/p>\n<\/td>\n

    		\n

    EVERSPACE 2<\/p>\n<\/td>\n

    		\n

    NTE (Neverness to Everness)<\/p>\n<\/td>\n<\/tr>\n

    \n

    Avatar: Frontiers of Pandora<\/p>\n<\/td>\n

    		\n

    F1 25<\/p>\n<\/td>\n

    		\n

    Portal with RTX<\/p>\n<\/td>\n<\/tr>\n

    \n

    Backrooms: Escape Together<\/p>\n<\/td>\n

    		\n

    FBC: Firebreak<\/p>\n<\/td>\n

    		\n

    PRAGMATA<\/p>\n<\/td>\n<\/tr>\n

    \n

    Call of Duty: Black Ops 7 (Coming August 20th)<\/p>\n<\/td>\n

    		\n

    Half-Life 2 RTX<\/p>\n<\/td>\n

    		\n

    Resident Evil Requiem<\/p>\n<\/td>\n<\/tr>\n

    \n

    Crimson Desert<\/p>\n<\/td>\n

    		\n

    Hogwarts Legacy<\/p>\n<\/td>\n

    		\n

    Samson<\/p>\n<\/td>\n<\/tr>\n

    \n

    Cyberpunk 2077<\/p>\n<\/td>\n

    		\n

    I Am Jesus Christ<\/p>\n<\/td>\n

    		\n

    Star Wars Outlaws<\/p>\n<\/td>\n<\/tr>\n

    \n

    Death Relives<\/p>\n<\/td>\n

    		\n

    Incursion Red River<\/p>\n<\/td>\n

    		\n

    Subliminal<\/p>\n<\/td>\n<\/tr>\n

    \n

    Directive 8020<\/p>\n<\/td>\n

    		\n

    Indiana Jones and the Great Circle<\/p>\n<\/td>\n

    		\n

    Sword of Justice<\/p>\n<\/td>\n<\/tr>\n

    \n

    DOOM: The Dark Ages<\/p>\n<\/td>\n

    		\n

    Marvel\u2019s Spider-Man 2<\/p>\n<\/td>\n

    		\n

    The First Descendant<\/p>\n<\/td>\n<\/tr>\n

    \n

    Enlisted<\/p>\n<\/td>\n

    		\n

    NARAKA: BLADEPOINT<\/p>\n<\/td>\n

    		\n

    War Thunder<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n

    While complete image fidelity across all ray or path traced games is by no means a 100% solved problem, it’s clear that DLSS 4.5 Ray Reconstruction is able to deliver a more stable image when it’s in use. And it’s not just about giving you a more stable image with less ‘boiling’ textures and trails, either, it’s adding back in more detail to a scene that would have otherwise been lost.<\/p>\n

    You can see that especially in the below video for Resident Evil Requiem’s opening scene. The textures on the white car at the beginning are still shimmering behind the rain effects using the older model, but are almost entirely stable with DLSS 4.5 RR enabled. But it’s the extra detail on the ground that highlights what the new second-gen transformer model is capable of here.<\/p>\n

    DLSS 4.5 Ray Reconstruction on left, DLSS Ray Reconstruction on the right<\/h4>\n