Often the implementation of super sampling technologies such as NVIDIA’s DLSS and AMD’s FSR in video games are perceived by users as an attempt to solve problems resulting from a lack of optimization by developers.
Speaking in a podcast, Jacob Stutton – co-founder of Infuse Studio – however provided a different explanation, placing the responsibility directly on the game engine, in particular for titles based on Unreal Engine 5.
According to his words, Epic’s engine would not be able to reach native 4K resolution with good performance.
Unreal Engine 5 doesn’t like 4K?
Speaking on the podcast “Moore’s Law is Dead”, the co-founder of Infuse Studio, Jacob Suttonexpressed some interesting opinions on the use of super sampling technologies in video games and the nature of graphics engines.
According to the developer, the idea that DLSS (which we talked about in depth in our special) or FSR are used as a sort of “crutch” to adjust game performance penalized by lack of optimization would be completely specious. Indeed, for Stutton, the use of upscaling is now a sort of prerequisitein particular for games based on Unreal Engine 5: the Temporal Super Resolution included in the engine, for example, would therefore be an obligatory choice and not a possibility to be resorted to based on need.
Even with built-in upscaling, at the moment it wouldn’t be possible to run a game at native 4K resolution anyway, while maintaining decent performance. The use of “third party” technologies, developed by NVIDIA, AMD and Intel, is therefore necessary due to the very nature of UE5.
Essential technologies?
That current games based on Unreal Engine 5 are quite demanding is certainly nothing new: titles that have used the full range of features granted by the engine, such as Nanite And Lumensstruggle to run smoothly, especially on consoles.
The latest tech demos are also pushing current GPUs, like the NVIDIA GeForce RTX 4080, to their limits, both in terms of performance and resolution.
The evolution of upscaling and super sampling technologies is therefore necessary to guarantee optimal performance in the near future and the big names in the sector have long traced a precise path in this sense. What do you think about it? Will DLSS, FSR and XeSS be increasingly central to achieving high performance? Will they become essential features for video game development?
