With the launch of Xbox One X on November 7, many games will be upgraded to take advantage of the new Microsoft console. We already have a list with more than 130 titles that will incorporate various improvements through a free patch. Such improvements will in many cases imply an increase in resolution. But not always the games will reach a native resolution of 4K (although logically is the goal and what we would like to see), but there will be other cases in which that resolution is achieved by a technique of checkerboard rendering.
What is the Xbox One X checkerboard rendering?
A term we’ve begun to see more regularly over the last few months, mainly because it’s the way PS4 Pro, Sony’s console, employs to achieve 4K resolution. Although some confuse checkerboard rendering with a rescaling, it is not the same, but something more elaborate and offers much better visual quality. It is a rendering technique for processing images at higher resolution. We explain the difference between both concepts so that you have a better idea of how in some cases your new Xbox One X will work.
Let’s start with the rescaling, which is a more basic technique. Rescaling an image at a higher resolution is stretching it. To do that, the color of the same pixel of the image at lower resolution is interpreted in the image rescaled by a group of pixels, the amount will depend on the variation of resolution. This makes the image lose definition, something a priori negative. But there are different methods to alleviate this loss of definition, and make the rescaled image become preferable to the original.
In any case, the results obtained by checkerboard rendering are much better , since here we do not talk about a stretching of the image, with its corresponding loss of definition, but of transforming the image to the new scale. What is done by checkerboard rendering is to predict missing pixels by increasing the scale of the image. That is, the original or native pixels remain the same, but the others are calculated to simulate the original image but at higher resolution, filling in the missing gaps.
For this process an analysis of the previous image is performed that is to the resolution that is intended to reach, which allows to predict the next one. Therefore, the more native pixels are in the original image and the smaller the difference between the native resolution and the one being searched, the fewer pixels will have to be filled, and the better the prediction will be. However, not necessarily a difference of smaller number of pixels, between native and predicted, is a better result of the technique, since starting from an image to a higher resolution can imply a better quality of the result, since it will be done a prediction based on much more information.
To clarify a little more this last one, suppose that a game yields to native 1800p (about 5 million pixels), and through checkerboard rendering it is going to transform its resolution to 4K. Missing pixels up to 4K (about 8 million pixels) will not be native, but a prediction made from the native pixels of the previous image. That is, the graphics engine is saving by the way natively produce 3 million pixels. In return those 3 million will offer them through an algorithm that requires much less effort. So that saved effort can go to something else, like performance (framerate).
Now, that algorithm may be different. The Ubisoft employee at Rainbow Six Siege , is different from the one proposed by Valve, and has made itself known for delivering very good results. I will not go on to explain the differences between the two, as it gives for long, and perhaps it is better to do it another time, but it seems that the method used in Rainbow Six Siege, which has already begun to spread to other games like Watch Dogs 2, and that is similar to the one used in PS4 Pro, it will be the technique that will begin to employ the majority of games in Xbox One X that work by means of chekerboard rendering. The issue becomes even more complex if we consider that the checkerboad rendering can be combined with other techniques, such as antialiasing (edge smoothing) or dynamic resolution.
To summarize everything commented. The demands to offer a native image at 4K, approximately 8 million pixels, are greater than those of offering a smaller image, say 1440p, and then take that image to a larger scale by some technique. And that’s where checkerboard rendering comes in, a process that offers better results than with a simple rescaling. This is not to say that the checkerboard rendering is going to be something usually used in Xbox One X. The Microsoft console has shown enough capacity to take games like Forza Motosport 7 to the 4K natively. But it is a good resource for developers in those cases where it costs to reach the native 4K, in exchange for achieving a good performance. This is a technique that has already been used in PS4 Pro and some PC games, such as Rainbow Six Siege or Watch Dogs 2, with great results.
The result achieved by checkerboard rendering does not reach the quality of a native resolution to 4K, but it is quite satisfactory, and Xbox One X should be able to make great use of it in those cases where the developers do not reach get a native resolution in your titles. In fact, on paper, the greater technical resources of Xbox One X could make even more advantage of this method. Whether starting from a higher native resolution, or combining it with other methods.