DLSS 5: Nvidia's Controversial AI Technology Gets a Second Chance

Nvidia's DLSS 5 technology has been met with criticism for its ability to make game characters look terrible, but the company is re-introducing it with a new message about preserving artistic intent and giving developers full control.

DLSS 5 uses generative AI to analyze and modify real-time streams of images after they've been rendered by the game engine. It can enhance light, shadow, color, and detail, making characters look more photorealistic. However, it can also be applied poorly, leading to unnatural-looking results.

Nvidia claims that DLSS 5 is deterministic, meaning games should look the same every time, and that developers have control over how it's applied. They can adjust sliders to change color and light, and choose from three different AI models: default, cinematic, or natural.

Developers can also decide which parts of the scene to apply DLSS 5 to, such as just a character's face or skin. Nvidia says that game studios may leave character faces alone, but this is unlikely given the potential benefits to gameplay experience.