The AMD 9070 / 9070XT Reviews and Discussion Thread

[DegustatoR]

It does if you enable ‘game optimization’.

Which isn't "automatic".

 

[raytracingfan]

That’s not practical as scaling is gradually becoming the “normal” way to play some games. For a long time now we’ve essentially assumed that all hardware outputs the same image when that wasn’t always the case. One day we’ll assume the same for upscalers.

I think reviewers will just have to include both performance at native resolution and upscaled performance at equal image quality. Ultimately reviewers will need to make a somewhat subjective decision as to quality levels of one upscaler are equivalent to quality levels of another upscaler (e.g. DLSS 4 Performance = FSR4 Balanced) and perform their benchmarks at those levels on each card to do comparisons.

 

[Scott_Arm]

I think reviewers will just have to include both performance at native resolution and upscaled performance at equal image quality. Ultimately reviewers will need to make a somewhat subjective decision as to quality levels of one upscaler are equivalent to quality levels of another upscaler (e.g. DLSS 4 Performance = FSR4 Balanced) and perform their benchmarks at those levels on each card to do comparisons.

I think I'm leaning more towards them just providing the performance numbers for all of the quality settings and then I'll decide what I think looks good enough or not. The image quality comparisons are useful, and I think their subjective conclusions can be useful, but ultimately I'm the one that decides what's good enough for me. I don't really need to see Quality mode vs Balanced mode because a reviewer has decided they're a closest match.

 

[Albuquerque]

The really difficult thing about upscalers is the really severe artifacting is often only visible in motion. Smearing, blurring, persistence of shadow, ghosting, so much of it is hard to capture in a still. And of course, YouTube isn't much help with their 60fps cap and bitrate constraints / re-encoding shens.

And worst is, some of it you'd never really see unless you knew right where and what to look for, other times it's pretty glaring -- like the DF video I watched today where they were playing RTX Remixed HL2 and the moving shadows of the rotating blade thingies for murdering zombies created a horrible shadow persistence / smearing issue. A picture would've shown nothing, but the motion made it sorely obvious.

 

[Scott_Arm]

The really difficult thing about upscalers is the really severe artifacting is often only visible in motion. Smearing, blurring, persistence of shadow, ghosting, so much of it is hard to capture in a still. And of course, YouTube isn't much help with their 60fps cap and bitrate constraints / re-encoding shens.

And worst is, some of it you'd never really see unless you knew right where and what to look for, other times it's pretty glaring -- like the DF video I watched today where they were playing RTX Remixed HL2 and the moving shadows of the rotating blade thingies for murdering zombies created a horrible shadow persistence / smearing issue. A picture would've shown nothing, but the motion made it sorely obvious.

Capturing screens while the camera is moving is sometimes better than a youtube video because you can get higher quality. It's a pain to do consistently. Controller movement is probably better than mouse for that. But you probably need video anyway for some things.

Either way, I'd still like to see thorough breakdowns of performance gains from upscaling at each major resolution and each quality setting in reviews. Not sure if there's something like that for 9070XT yet, but Daniel Owen's video did a nice job of showing some of the upscaling gains and then showing the final performance comparisons for different use cases like everything included, ray tracing only, native only

 

[trinibwoy]

HUB breaks down the rebate situation.

 

[Florin]

It does if you enable ‘game optimization’.

Which, again, isn't the default