Thought we could have a seperate thread to compare the differences of CNN and Transformer. I'll add my own comparisons when the app update comes out, down below. Youtube comparisons also welcome.
I did various testing with the new model while not perfect is better overall than the CNN model. There are still improvements to be made, surfaces, vegetation etc need to be improved.It's incrementally better but also slower (can also be a mixed bag in places).
They threw parameter counts at the problem (but Catarzano was at least honest about that).
tl;dr ML upscaling plateaued.
I did various testing with the new model while not perfect is better overall than the CNN model. There are still improvements to be made, surfaces, vegetation etc need to be improved.
One thing I did notice though some games do get major improvements while others do not. Ghosting and motion clarity improvements being the main ones that the transformer model improves on.
Should be great for lower end cards
path tracing is a bit much for a 4060. yeah and look at the wiring/fence and edges are too soft, no good.Does this reviewer not realize he's destroying the image quality for RT FOMO? It looks like playing games without prescription glasses.
View attachment 115599
With 20 series now being over 6 years old, this is a big win for owners.
You can actually run Indy on Linux on Vega 64/56 using software based Ray tracing. You can’t do that 5700 XT due to low compute.5700 XT due to Raytracing requirement
That's the coolest thing I will see today. Another clipYou can actually run Indy on Linux on Vega 64/56 using software based Ray tracing. You can’t do that 5700 XT due to low compute.
So it’s goes like this RDNA3 > RDNA2 > Vega > RDNA1. RDNA1 is the worst AMD architecture ever.
It's not similar. Over-sharpening was part of the criticism we had for both early DLSS and then early FSR, but nowadays the sharpening is considerably milder and often under user control. Sharpening will not solve AA artifacts.Isn't that similar blur on/off?
This is the way, this is why I keep pushing for DLAA and FSR Native AA to be included in comparisons with upscaling examples, the improvement in image quality is something more people should be aware of.
It doesn't. Transformer model is much closer to what it should look like. Previous model boosted normal map detail way too much.Here transformer looks worse
Maybe, but skin detail in CNN more looks like in real life. I'm also showed to friends, what is better, they said all CNN.Previous model boosted normal map detail way too much.
It doesn't. Transformer model is much closer to what it should look like. Previous model boosted normal map detail way too much.
You can actually run Indy on Linux on Vega 64/56 using software based Ray tracing. You can’t do that 5700 XT due to low compute.
So it’s goes like this RDNA3 > RDNA2 > Vega > RDNA1. RDNA1 is the worst AMD architecture ever.
In any case, previous ray reconstruction had that issue (IMO the left looks bad and also the faces looked bad in Cyberpunk). Now things that should be sharper generally look sharper without vaseline smearing. Nothing is perfect though.Yes, I also heard the explanation, but I don't agree.
What it's doing may theoretically be better, but practically, the transformer model looks too smooth and plastic, and the old CNN model looks better.
RDNA1 beat Vega while using considerably less power. It also beat the 2060 and had more VRAM. It aged better until now.