Graphics card is so expensive, but the game still has to play, the performance of the old graphics card or barely affordable low-end graphics card, how to play the latest and best games? Graphics card manufacturers are certainly afraid of players completely desperate, so the update of frame rate optimization technology and even the push of new ones have accelerated, the new year has just begun, N card A card and each launched a new frame rate optimization technology, compared with the previous technology, what is the difference between them? The result? If the graphics card is not strong enough, let's take a look.
This time, NV's DLDSR technology and AMD's RSR technology invariably chose a way, that is, no matter how you set the game, in short, I use a lower resolution to process the picture, and then stretch it into a high-resolution output. In this way, players only need to set it up in the driver, do not have to look for function switches everywhere in the game, and do not need manufacturers to modify the game engine to support related technologies.
Of course, this way is also a bit simple and rough, although the processing method is closer to the bottom layer of image generation and input, so the frame rate may be higher, but the picture effect is probably not comparable to the game of cooperative development and optimization effect. Of course, NV will still use the RTX 20/30 series series of RTX GPU Tensor tensor core operations to assist in processing, or "brain supplement", which should improve clarity, but there may also be strange image problems with brain compensation errors.
Correspondingly, the typical application scenario announced by AMD is 4K resolution, which should be significantly different from the actual processing resolution (estimated to be about 1080P). NV is ready to use at 1620P (about 2K resolution). In terms of picture quality, in the comparison chart released by AMD, the texture of the background such as the wall after the RSR technology is turned on is obviously a little more mushy, but because it is only the original picture stretch, the picture is certainly impossible to make a mistake.
So, why is there DLSS, FSR, etc., and are still in a hurry to introduce these two more adaptable frame rate optimization technologies? Xiaobian believes that the current graphics card supply is too disconnected from the needs of players and game developers, and it has made people a little impatient, and they must hurry up so that players can play games with excellent results. Otherwise, players do not have high-energy graphics cards, and developers have to use cartoon rendering or even 2D graphics to develop playable games, and the rapid development of game graphics cards will be in trouble.