Now you claim that rarely adds things that modifies the original image on an artistic level, but it is a fact that an AI fills in all the “missing spaces” that were created by the upscaling process.  This is, by its very nature, adding things that modifies the original image.  The developers certainly did not fill those “missing spaces”, an AI did.  

Do you accept this much to accurately describe what happens during DLSS upscaling?

Yeah, but the problem is that FSR 4.1 is still being behind DLSS 4.0 in some aspects. PS6 will struggle with both path tracing and neural rendering enabled. The console is supposed to have the TDP of 170W according to the rumours, which is not helping. While something like Switch 3 might use even more advanced version of neural rendering. Given Nvidia huge advantage at ray tracing and frame generation, I believe only S3 will be able to render something better than presented examples.

This is on a 5090.... and it has jotter and lag.... you notice it with the electric poles on the side of the road as he drives by.

Supposedly dlss5 cost more than half the frame rates, you normally have, to run it.
If a 5090 cannot run Forza Horizon, at high frame rates with this on, yeah, Its of no use to me.

IF you go to 1:53 (again at 3:24) in the video and watch him turn "intensity" of dlss5 to 0% and back up to 100% (I assume that is on/off)...
the visual difference really is not that big.  It changes the lighting of the scene and that is about it.

The best use and worse of it too, is still faces of characters imo.
As you can see towards the end of the video.
Not that that stuff matters much of anything in a racing game.

The changes to characters arn't always for the better though.... and the demand on the gpu, is not worth it imo.

FSR 5 and PSSR 3 shouldn't be. A neural rendering quality mode can just leave path tracing off if both would be too much. I was thinking the PS6 and Xbox Helix would have 30fps path tracing modes but that could also be done for neural rendering too so multiple quality mode options. The GPUs in the RDNA 5 lineup should be capable of neural rendering unless AMD drops the ball.

But the same can be said about other rendering techniques, such as memory compression, which can reduce or change the textures in the way not intended. This technique is not using AI, but with next generation consoles there will be AI based memory compression too. There are always compromises and DLSS upscaling is not made to change the intent, but to provide more performance for slightly lower quality or less correct image. What if in the future AI can provide 3 times more performance for the image that is 98% correct, will you like it by then?

I’ll take it that you accept what I wrote is an accurate description of what happens during DLSS upscaling.

Do those other rendering techniques you mentioned, such as memory compression, use the same methodology as what I described for DLSS upscaling?

If so, I am not a fan of them.  If not, then you really can’t say the same about them.  Either way, my opinion of DLSS remains unchanged.

As for your question about improved performance, it depends how the AI “improves” performance.  If it imposes its own changes to the art independently of the developers, then my stance on this hypothetical future AI is no different than my stance for the current methodology that DLSS frame generation/upscaling uses.

Ok, when it comes to typical 3D graphics, there are a lot of compromises and fakeness or masking in order to provide realistic graphics. When it comes to memory compression, the logic is to throw away or use less precise parameters to show most important parts of the graphics, while the rest that is not visible is not in the image. So for sure in humans eyes it can be seen as almost the same image, but not truly the same and as intended by the artist.

There's only so much compute available in PS5, so a lot of things aren't rendered as intended anyway. So, based on what you wrote I don't think you specialize in computer graphics. You can have your subjective opinion sure. But if AI can provide a better looking image than native with 98% correctness like the developers intended it to be displayed, with higher performance, I would gladly turn it on.

Technically RDNA 5 supports neural rendering, but the quality of image and performance is more AI model dependent. Knowing AMD and the reality of AI research talent pool, the majority of people who are geniuses, work either on their own projects or for the biggest companies in the AI industry paying millions per year. So, I doubt AMD will be competitive with Nvidia next generation. If Switch 3 has something beyond Rubin architecture, then the quality of image should be even higher, while the performance penalty lower than now.

I do expect AMD to remain behind yeah so RDNA 5 will be notably behind Rubin in this area. I expect the PS6 and Xbox Helix to handle neural rendering about as well as the PS5 and Series X handles ray tracing but even the first generation of this stuff is insane so it'll still be a significant selling point. DLSS is so incredible already that the 2030's are gonna get truly wild.

I do expect AMD to remain behind yeah so RDNA 5 will be notably behind Rubin in this area. I expect the PS6 and Xbox Helix to handle neural rendering about as well as the PS5 and Series X handles ray tracing but even the first generation of this stuff is insane so it'll still be a significant selling point. DLSS is so incredible already that the 2030's are gonna get truly wild.

I doubt they can be as behind as they were with RT since ANN inference performance is something the PS5 Pro is focusing on already, for instance. AMD was clearly caught off-guard by Nvidia's RT cores, and it was basically an afterthought in RDNA2. With this stuff, not so much.