Microsoft today via a devblog announced their push to make Variable Rate Shading an industry-wide adoption in the search of increased performance that can support the number of pixels and quality of those pixels in future games. The post starts with what is likely the foremost question on the mind of any discerning user that hears of a technique to improve performance: "does it degrade image quality?". And the answer, it seems, is a no: no discernible image quality differences between the Variable Rate Shading part of the image, and the fully rendered one. I'll give you the option to speak on your own perception, though: analyze the image below and cast your vote on the poll.

As resolution increases, so does the amount of work that any given GPU has to process to generate a single frame - and compare that to the amount of additional work that goes from rendering a 30 FPS, 1080p game to a 60 FPS, 4K one, and... It stands to reason that ways of squeezing the highest amount of performance from a given process are at a premium. Particularly in the console space, where cost concerns require the usage of more mainstream-equivalent hardware, which requires creative ways of bridging the desired image quality and the available rendering time for each frame.We've already spoken at length regarding Variable Rate Shading, regarding both its NVIDIA Turing debut, and the AMD patent application that aims to implement a feature that's akin to that. In this case, this is Microsoft that's saying they're adapting Variable Rate Shading onto DX12, allowing developers to easily take advantage of the feature in their games. According to Microsoft, integration of VRS technology through DX12 should take developers no more than a few days of work, whilst enabling some 14% performance gains for other rendering efforts, such as resolution, target frames per second, or even more relevant image quality improvements.Microsoft details three ways for developers to integrate the technology into their rendering engines (per draw; within a draw by using a screenspace image; or within a draw, per primitive). This should enable developers to mix and match the implementation that best applies to their engine. At the time of announcement, Microsoft said that Playground Games and Turn 10 (Forza), Ubisoft, Massive Entertainment (The Division, Avatar), 343 Industries (Halo), Stardock, Io Interactive, Activision, Epic Games and Unity have all shown their intention of adding VRS to their game engines and upcoming games. That most of these are affiliated with Microsoft's push isn't surprising: remember what we said early in that the console development space (and VR) is where these technologies are needed most.

PS: The left part of the image is the fully rendered one, the right part of the image is the VRS-powered one, rendered at a 14% increased performance.