Microsoft is highlighting a certain technology on its future Xbox Series X, which should improve game fluency while maintaining a 4K definition. Its name: Variable Rate Shading. Probably also used at Sony on PS5, the VRS is mainly a question of software optimization. Explanations.
The new generation of consoles, coming at the end of the year, brings its share of acronyms and technical terms that are more or less obscure, but important to understand what distinguishes it from the previous generation. So let’s continue to explore the imaging technologies that will determine future machines and look at Variable Rate Shading or VRS. Less in the limelight than ray tracing, the VRS is nevertheless essential in the context of game optimization. Because of rotate Halo Infinite or Godfall in Ultra HD and at 60 FPS, some “tips” will necessarily be requested from the developers. VRS is one of them.
Confirmed for now on the Xbox Series X, however, the VRS will likely also be present on PlayStation 5 (PS5), as the two consoles are based on the same RDNA 2 architecture from AMD, which will use this rendering technology. However, Microsoft uses its own “version” of VRS integrated into DirectX 12. In addition, remember that VRS is not new since it has been used at Nvidia for some time, under the name Adaptive Shading. The first to take advantage of this was Wolfenstein II: The New Colossus.
Cover this shading that I can’t see
But enough, it’s time to get to the heart of the matter: what is the Variable Rate Shading? In a few words, we could define it as optimization of the calculations needed to display a 3D image, depending on its complexity.
Let’s get into detail: every pixel displayed on the screen needs a shading, which gives it its shape, transparency, light, etc. In an Ultra HD screen, with 3840 x 2160 pixels, the number of shaders to calculate at the same time is enormous, the number of pixels displayed on the screen reaches nearly 8.3 million. The VRS trick is to keep in mind that these 8.3 million pixels don’t necessarily require the same display level: they can barely be visible, because in the dark or at the edge of the player’s gaze (this is also called foveal rendering for virtual reality). In the context of very fast movements (for example in an FPS), it is the same: why claim the ability to accurately calculate a pixel that will fade?
A particularly telling image is this below: in Blair Witch from Bloober Team, we spend much of the game in almost total darkness. In this case, is it necessary to calculate the representation of the forest that is in the dark and therefore invisible to the eyes of the player? The Variable Rate Shading finds its interest here because it will adjust the calculation of each shading in real-time to what is lit. In the context of Blair Witch and our sample image, nearly half of the player’s “sight” is plunged into darkness, which equates to as many fewer calculations to perform for the graphics card.
What’s the point of calculating all the representation of dark areas?
Another example, from Nvidia documentation: on a car racing game like Forza Horizon 4 it is not necessary to apply the same level of shading to the whole image. Here, only the car, horizon, and sky are calculated with a level of detail “at 100%”, since it is in this area that the player’s gaze is focused while driving. In the green-colored area, a single shading applies to 4 pixels and even 16 pixels in the red part, little observed in the game.
Of course, not everything is “black or white” and the VRS has to constantly adapt to what is displayed on the screen. If we take the example of Adaptive Shading at Nvidia, the technique is used in different ways and takes over the three above-mentioned instructions: Depends on the brightness of the area to be calculated, the complexity of the texture, the position relative to the player’s gaze and the movement speed causing the blur effect (the famous “motion blur”).
All this makes for a custom VRS, which saves valuable computing resources. According to tests conducted by Tom’s hardware on Wolfenstein II, the use of VRS makes it possible to achieve between 5 and 15% extra FPS in the game. It may sound weak, but that’s what can make all the difference in achieving the 60 FPS in Ultra HD that was so hoped for on the following consoles.
The VRS that Microsoft plans to use in the Xbox Series X comes from the PC and is part of the DirectX 12 API. Originally presented in March 2019, theoretically, it allows up to 20% more performance, with an almost invisible visual loss. The company also provides an example with Civilization VI (below): only the red areas benefit from “full” shading. If a small handful of games use this technology today, several publishers and developers plan to use it for their next productions. Among them Ubisoft, Turn 10, Playground Games, 343 Industries, or Epic Games.
With ray tracing or SSD, the Variable Rate Shading is one of the important points to analyze when the Xbox Series X and PlayStation 5 (PS5) are released. If the second isn’t yet officially using this technique, its presence on AMD’s rDNA 2 architecture – which will be on both consoles – strongly suggests that the VRS will be on your TV no matter what machine is on the throne in your living room in a few months. In any case, the Variable Rate Shading is more proof that the next-gen will be both a matter of optimization and rough performance.