Nvidia announced DLSS 5 back in March and it received, let’s say, mixed reactions. A lot of people weren’t happy about how, judging by the demos, it seemed to essentially just slap a ‘yassifying’ AI filter on top of a game scene. Nvidia itself confirmed that all the tech has to work with for each AI frame is a traditional 2D frame and motion vectors, essentially confirming the “filter” idea. Now, at SIGGRAPH 2026, the company has explained that developers will get two main sliders to control DLSS 5’s intensity.
My initial reaction was probably the same as yours: two whole sliders, whoop-de-doo. It’s hardly a massive improvement from the one on-to-off masking slider we previously thought devs would have access to. But let’s take a look at what they’re actually controlling and how it works before we jump to conclusions.
The two sliders are for ‘structural intensity’ and ‘tone intensity’. Despite what the word “structure” might seem to imply, Nvidia is keen to point out that “it is not changing the geometry in any way, and it is uplifting the image.”
The structural intensity slider in fact is “modulating the high-frequency detail of the frame.” This includes things such as ambient occlusion, subsurface scattering, and reflections. The tone intensity slider is “modulating the low frequency detail of the frame”, including “expressive” things such as lighting and, well, overall tone.