AMD Vs. Nvidia Image Quality - Old man yells at cloud

[Donnie Fisher]

Sorry Rofflay, we may be at cross wires here. I didnt make any video. My illustration which I referred to, and my points in general was from a couple of pages back where I had shown photos taken of the desktop as rendered by nVidia and AMD cards in my machine where there were differences.

It has got me looking at more of the settigns on the nVidia control panel though to see if there any any differences I can make. In particular, I would say that there is a difference within nVidia itself between RGB (full) and YCbCr444 with 444 being a little clearer for the desktop environment. (imho)

 

[4K8KW10]

My god read the thread he does not know how to setup Nvidia properly the whole test is null and void until someome with a 2080ti and a 5700xt runs side by side with the proper settings used on each vendor. His image there is laughably blurry and seems to lack Clamp and AF16. You do not need to watch the video look at the wall on the thumbnail. It looks like default Nvidia settings which they may alter for better default 3d mark scores. But who cares when you should be using the highest driver settings and turning down ingame settings. Honestly i think sometimes you guys dream of this being reality that finally AMD better Nvidia and you seem to completely jump facts when doing it?

It should be default Radeon Settings vs default nvidia control panel. Very apples to apples comparison.
I don't understand what you expect - with Radeon settings no mater what you tweak, you can not reach nvidia's quality.
Also, the default is what the users get, and I'm sure most of the users don't touch the control panel because they don't know what to change.
And even if they change the settings, they will still observe lack of details on the nvidia image.

 

[Rofflay]

It should be default Radeon Settings vs default nvidia control panel. Very apples to apples comparison.
I don't understand what you expect - with Radeon settings no mater what you tweak, you can not reach nvidia's quality.
Also, the default is what the users get, and I'm sure most of the users don't touch the control panel because they don't know what to change.
And even if they change the settings, they will still observe lack of details on the nvidia image.

But i thought AMD had higher quality??? You guys were saying this. No one should be using anything but the highest AMD versus highest Nvidia. It is only an fps at most maybe 2fps this is how it should be done. Later if you want perfprmance turn off ingame settings. This is like me painting something, And i start off with a low quality canvas and use expensive paint. It makes no sense i can do a sticky guide or you can look my history where i did a post on what actually should be on and off.

 

[Grim5]

Yeh ask anyone on OLED or with a 5k true contrast VA panel if Nvidia look like that. Not at all the blacks are the same on every vendor and they are superb.

OLED TV here, playing PC games with the 2080ti blacks don't look like nvidia footage that at all, they are look black like outer space.

The fact that those blacks look washed out means something is very wrong with the settings.

 

[Rofflay]

OLED TV here, playing PC games with the 2080ti blacks don't look like nvidia footage that at all, they are look black like outer space.

The wrong RGB does that, Indeed blacks are stunning i run 4k strobed at 120hz and the FG2421 has 5x the contrast of a 1000:1 IPS which is common post calibration on most non FALD IPS. Sometimes i find it absurd how that cost £350 from ocuk.

Nvidia have stunning image quality maxed out, Even on 1080p with 4xmsaa they never look like those videos.

 

[robfosters]

Recently swapped from a Vega 64 to a 2070 super and haven’t noticed any difference in image quality at all.

 

[4K8KW10]

Recently swapped from a Vega 64 to a 2070 super and haven’t noticed any difference in image quality at all.

Why would you do that? Those offer virtually the same FPS performance, not a real upgrade ?

Nvidia have stunning image quality maxed out, Even on 1080p with 4xmsaa they never look like those videos.

And now you have the opportunity to make a video so we know what exactly you are talking about.

 

[loftie]

Definitely interesting - though the difference is very minor, I noticed a much bigger difference than that just changing my panel from QLED to OLED.
Be interesting to see the settings used on each to see if they match, as well as what monitor was used since that also has an impact on what the GPU outputs - the colours on the AMD card do seem to have slightly more colour saturation.

I noticed immediately, then he mentions it on the video. At first I thought it was because the first game is BL3 and something to do with CAS, then when it switched games I realised it wasn't to do with CAS. Settings were stock out of the box IIRC, he does mention that you can tweak the nvidia settings to make it run correctly but since it's a CPU review not a GPU Vs GPU review he doesn't bother to do it.

What I wasn't sure on was if correctly setting the colour space on the 2080Ti would make the image appear sharper due to better shadows, or if the AMD just looked sharper. Was AMDs RIS automatically turned on?

 

[robfosters]

Why would you do that? Those offer virtually the same FPS performance, not a real upgrade ?



And now you have the opportunity to make a video so we know what exactly you are talking about.

So far, I get better FPS in every game, it actually completes Davinci resolve renders without crashing and it’s ridiculously quiet and cool compared the F-16 jet engine that was the Vega.

wasn’t keen on the vega at all tbh.

 

[Rofflay]

Why would you do that? Those offer virtually the same FPS performance, not a real upgrade ?



And now you have the opportunity to make a video so we know what exactly you are talking about.

Actually no because i do not install anything but games. Recording software is not something i habe even used once. I have no AMD card either and my 2080ti arrived so im going to be busy and then afk for 6months away from the forum as there is nothing tech wise i need anymore.

 

[bru]

It should be default Radeon Settings vs default nvidia control panel. Very apples to apples comparison.
I don't understand what you expect - with Radeon settings no mater what you tweak, you can not reach nvidia's quality.
Also, the default is what the users get, and I'm sure most of the users don't touch the control panel because they don't know what to change.
And even if they change the settings, they will still observe lack of details on the nvidia image.

Really, Ok.

Just remember those words next time a power debate comes up, or benchmarking. stock settings only. no tweaking or undervolting. Even the new RIS, which I believe the stock setting is off.

 

[4K8KW10]

Really, Ok.

Just remember those words next time a power debate comes up, or benchmarking. stock settings only. no tweaking or undervolting. Even the new RIS, which I believe the stock setting is off.

I have complained already that AMD does not good job with overvolting their cards out of the box. But that is a topic for another thread.

 

[Grim5]

I noticed immediately, then he mentions it on the video. At first I thought it was because the first game is BL3 and something to do with CAS, then when it switched games I realised it wasn't to do with CAS. Settings were stock out of the box IIRC, he does mention that you can tweak the nvidia settings to make it run correctly but since it's a CPU review not a GPU Vs GPU review he doesn't bother to do it.

What I wasn't sure on was if correctly setting the colour space on the 2080Ti would make the image appear sharper due to better shadows, or if the AMD just looked sharper. Was AMDs RIS automatically turned on?

for the washes out blacks and other colours - I can achieve the same look with any one of various settings which is why it's pointless comparison without knowing the settings.

for example if I set my Panels black level from low to high, it will look like that nvidia footage, same if I set brightness/gamma level over 50%. If I set the RGB color to Limited instead of full on my Nintendo switch it will look like that Nvidia footage. And lastly if I set RGB to limited in the Nvidia control panel it will look like that Nvidia footage.

and these are just the ones I know about

I don't own project cars - but I'm happy to run other games and take off screen pictures so you guys can see what it looks like on my 2080ti

 

[Rroff]

As far as default colour formats go you have to change them to whatever is applicable for your setup - default to full or limited range doesn't matter if the setup you have needs a different colour format/range - criticising nVidia over that aspect rather than taking a moment to set it up correctly is pitiful.

 

[Raelgun]

As far as default colour formats go you have to change them to whatever is applicable for your setup - default to full or limited range doesn't matter if the setup you have needs a different colour format/range - criticising nVidia over that aspect rather than taking a moment to set it up correctly is pitiful.

Sorry but, how much does Nvidia charge again?

If you are spending a ton of money on getting the 'best' surely the vendor should present your purchase at its best light out of the box?

Also it makes more sense for the image quality to be different from both vendors, for them to look exactly the same or close to is actually strange.

Both vendors have different processing pipeline and hardware baked features, so its logical and makes more sense for the output to be different.

I mean, the best case scenario for this is digital cameras, take a look at reviews between the best canons and sony cameras and you will see a difference and this applies the same here, they both have their own sensors/processors to output an image, again GPUs are not CPUs.

 

[Rroff]

If you are spending a ton of money on getting the 'best' surely the vendor should present your purchase at its best light out of the box?

AMD and nVidia might have different defaults but that doesn't mean they will support all setups correctly or the best out of the box.

 

[Grim5]

Sorry but, how much does Nvidia charge again?

If you are spending a ton of money on getting the 'best' surely the vendor should present your purchase at its best light out of the box?

Also it makes more sense for the image quality to be different from both vendors, for them to look exactly the same or close to is actually strange.

Both vendors have different processing pipeline and hardware baked features, so its logical and makes more sense for the output to be different.

I mean, the best case scenario for this is digital cameras, take a look at reviews between the best canons and sony cameras and you will see a difference and this applies the same here, they both have their own sensors/processors to output an image, again GPUs are not CPUs.

The settings you need to use depends on the screen that you plug in, it's that simple. The driver can't adjust itself perfectly for the million different panel models out there, don't be silly.

Even the screens themselves need tweaking - TV's can cost $5000 but you still need to change settings - doesn't mater how much they charge you for the panel.

 

[Raelgun]

The settings you need to use depends on the screen that you plug in, it's that simple. The driver can't adjust itself perfectly for the million different panel models out there, don't be silly.

Even the screens themselves need tweaking - TV's can cost $5000 but you still need to change settings - doesn't mater how much they charge you for the panel.

When doing comparisons between 2 processing units, you don't change anything else outside that, we are talking about AMD vs Nvidia when it comes to image processing and these are the only 2 variables required.

Image processing units from different vendors should produce different results, this is actually expected, they should not be producing the exact same image.

But regards to the topic of drivers, yes, I think the default should be at its best, not several notches down, we have in game settings for this.

Also who says TV vendors get a pass on this? same thing applies to them.

 

[Drumroll]

Never noticed. Possibly because I usually adjust colours and other settings when upgrading monitors/GPUs.

 

[Rroff]

When doing comparisons between 2 processing units, you don't change anything else outside that, we are talking about AMD vs Nvidia when it comes to image processing and these are the only 2 variables required.

For an objective test you'd set them to the same equivalent, optimal, settings. Anything else is silly and doesn't prove anything.