Xbox Upscaler... Any Details and Information?

So things should be getting better! I would frankly much prefer to play with a gamma slider than mess around with the brightness and contrast on my TV. Sacrilege! :yep2:
Precisely. The TV should show a good picture. If it's set up to show a good picture from the usual sources, the same settings should apply to game content. A photo on the PS3's viewer looks excellent on my display, and a greyscale range is perfectly balanced (to my eye, which is the only calibration that matters ;)). Ergo the game should work at the same settings. Yet invariably the darks are pretty much black at the recommended 'can't see the logo' gamma and the game is a muddy mess. Slide the in-gamma up so the logo is visible and the game looks great.
 
I've never known a gamma slider work! I just ignore them and slide it so I can see the darks which black out otherwise.

Well, I roughly calibrated my tv so I usually don't need to use the sliders at all, but if you saw the results from this one it's more broken than any slider has ever been. When games have them I usually take a quick look just to see if I'm in the ballpark of what the game is expecting. This one was way off and I put the slider all the way to the darkest setting, which made the game incredibly dark to look at, but the target symbol in the slider was still as visible as it was on the default position.
 
To be clear, the xbox one (and 360 for that matter) supports both full range and limited rgb out just fine to match whatever your set is expecting. The problem with Sunset Overdrive is that it is "always" applying a 2.35 gamma curve, correct for limited rgb btw, but doesn't correctly switch to the standard 2.2 gamma curve when the console is set to full rgb. While some may disagree, limited is the best setting in general since it will be correct almost always. The term "limited" implies a negative connotation in that it doesn't carry the same range of values, which is also incorrect since the standard supports the notation of blacker-than-black and whiter-than-white so the actual results are just as good as if everything is correct using full without the risk of it being wrong.
 
I think the divergence here is that the Sony platforms are defaulting to full and 2.2 gamma, which is better suited to PC displays and capture cards, where Microsoft platforms are defaulting to limited and 2.35 gamma which is better suited for the typical HDTV display.
 
To be clear, the xbox one (and 360 for that matter) supports both full range and limited rgb out just fine to match whatever your set is expecting. The problem with Sunset Overdrive is that it is "always" applying a 2.35 gamma curve, correct for limited rgb btw, but doesn't correctly switch to the standard 2.2 gamma curve when the console is set to full rgb. While some may disagree, limited is the best setting in general since it will be correct almost always. The term "limited" implies a negative connotation in that it doesn't carry the same range of values, which is also incorrect since the standard supports the notation of blacker-than-black and whiter-than-white so the actual results are just as good as if everything is correct using full without the risk of it being wrong.
Thanks! This is what I tried to say many many times but never found the right words to do so. You pretty much nailed it.
 
"Most scalers" are good enough to scale 720p to 1080p without noticeable blur. As taisui said, even something as simple as Bicubic would be sufficient. Unless the XB1's scaler is now worse than the X360's, then there shouldn't be any noticeable blur being added from scaling.


I'm not saying that people don't prefer the old scaler. I'm simply saying that there's probably not any added blur, it's simply blurrier because it's lower res than native 1080p.

I am not sure. I compared several upscaled XB1 games to the native PS4 version (after the sharpening filter has being removed). We can notice a slight blur even on identical medium resolution textures.

From memory I noticed this slight blur with notably BF4, Thief, Metro Redux, Shadow of mordor and COD AW (somebody posted an example of CODAW on another thread). But it may not be noticeable to everyone, it's very slight.

Here's an example of a medium resolution texture found in Shadow of Mordor. PS4 native 1080p VS XB1 upscaled 900p. No need to even label both versions as the blur is easily noticeable:

enhb.gif


This kind of added blur could come from a bilinear interpolation upscaling.
 
Here's an example of a medium resolution texture found in Shadow of Mordor. PS4 native 1080p VS XB1 upscaled 900p. No need to even label both versions as the blur is easily noticeable:
Wow, I'm actually quite surprised at that difference but it's jarring because it's flicking back and forth. No doubt yours eyes would adjust quickly to the upscaled 900p.

But still. Wow. I'm shocked.
 
I am not sure. I compared several upscaled XB1 games to the native PS4 version (after the sharpening filter has being removed). We can notice a slight blur even on identical medium resolution textures.

From memory I noticed this slight blur with notably BF4, Thief, Metro Redux, Shadow of mordor and COD AW (somebody posted an example of CODAW on another thread). But it may not be noticeable to everyone, it's very slight.

Here's an example of a medium resolution texture found in Shadow of Mordor. PS4 native 1080p VS XB1 upscaled 900p. No need to even label both versions as the blur is easily noticeable:

enhb.gif


This kind of added blur could come from a bilinear interpolation upscaling.

Could the blur not come from the image being sub native in the first place?
If you really want to test the difference between both systems scalers you would need to test
a game that is sub native on both systems. Maybe you could try AC unity if it doesnt use software scaling. The results would be interesting. I admire your pixel counting abilities Global.
 
Could the blur not come from the image being sub native in the first place?
If you really want to test the difference between both systems scalers you would need to test
a game that is sub native on both systems. Maybe you could try AC unity if it doesnt use software scaling. The results would be interesting. I admire your pixel counting abilities Global.

Yes, you may be right, that's why I wrote I wasn't sure.

Like you said AC Unity at 900p in both versions should add some really interesting new materials. Funny you bring AC Unity because I already thought about that game for helping us in those blurred matters. :sly:

But one thing is for sure, the upscaling used in almost all upscaled XB1 games (except a few exclusives notoriously using software upscaling) is not very good and very, very similar to the one seen in the 900p PS4 games.
 
AC Unity being upscaled from 900p on both PS4 & XB1 we can now safely assume that both hardware upscaling solutions are virtually identical. I have detected no differences of clarity or blur, even minimal, between both.

http://images.eurogamer.net/2013/articles//a/1/7/1/9/4/6/4/1PS4_013.bmp.jpg
http://images.eurogamer.net/2013/articles//a/1/7/1/9/4/6/4/1XO_013.bmp.jpg

http://images.eurogamer.net/2013/articles//a/1/7/1/9/4/6/4/1PS4_008.bmp.jpg
http://images.eurogamer.net/2013/articles//a/1/7/1/9/4/6/4/1XO_008.bmp.jpg
 
Last edited:
So here is what I want to know. Are there extra goodies built into the upscale chips these days? I noticed several places on BF4 that just flat out look better on the Xbox One. Bash me all you want, but I know what I see. Several areas the game just looks more detailed and prettier on Xbox One.. even if it it upscaled.
Can you find a screenshot for comparison? The face-offs show little difference - both games are upscaled.

So is the upscaler doing some extra processing and adding to the visuals to make them look better.
No. The upscaler can interpolate so added pixels get close to what their natively-rendered counterparts would look like, but they operate at a level that cannot add detail to the graphics. Again, you're possibly looking at a sharpening filter in effect making the visuals 'pop'. Have you compared the game on XB1 and PS4 on the same screen, or in different circumstances introducing more variables?

Just Googling, BF4 doesn't even use XB1's hardware scaler.
Digital Foundry said:
The pre-final code utilised the Xbox One hardware scaler and featured an unnatural gamma shift, while the release version saw the implementation of a software scaler created by DICE itself that resolved the colour balance issue in addition to removing the sharpening effect.
Hence whatever improvements you feel you are seeing are difficult to justify logically. First guess is that you've seen XB1 on a TV with sharpening turned up or something, while you saw PS4 on a less than ideal display. Second guess, you just prefer the blurrier look of 720p upscaled to the slightly less blurry look of 900p upscaled. Third guess, it's entirely placebo effect and in a blind test you wouldn't be able to tell the difference. ;)
that would eleviate the need to run at 1080p and free up resources.. insuring that the hardware always stays at 60fps even as the machine ages.
o_O You reference BF4. BF4 fails to hit 60 fps on XB1 now. It's slower than PS4 at 900p (30% lower res than 1080p which is considerably sharper than 720p). So how do you get from "it's rendering 720p at a lower framerate in BF4" to "it can achieve stable 60 fps in future while looking just as good as1080p"?
 
The other day I noticed something curious. I took a screenshot of Life is Strange and I don't have Kinect connected to the console, so I must double tap the Guide button -the large X button in the middle of the gamepad- and press Y to take a screenshot.

Thing is... once you tap the guide button just once, the game goes to the pause screen. So by the time you hit the guide button a second time and press Y you are in the pause screen already. (this doesn't happen with Kinect 'cos you just use your voice)

So the message "Xbox screenshot saved" -or whatever it is- prompted and I was very disappointed thinking that the screengrab I got was from the pause menu. In the end it wasn't, :smile2: it was from the frame that the console was displaying before going to the pause menu. So I took the screenshot just fine.

It's like the pause menu and the game itself are two different display planes.
 
Or they buffered the last frame in memory.
That sounds very plausible, and maybe details like the guide button method are what made taking screenshots a bit more technically challenging than expected, according to Phil Spencer.
 
Press once, take screenshot. Press twice, save it out.

Possible, but with the current method, pressing once sends you to the Home screen & the game/app continues to run in a window. So there's no way of knowing if the screenshot was taken or not on the first press.

Tommy McClain
 
Yep, but there's no need to know. The system can grab it. It needs that slice of RAM reserved for if you do want a screenshot anyway, so may as well implement it that way. darned sight easier than any other solution! ;)
 
Yep, but there's no need to know. The system can grab it. It needs that slice of RAM reserved for if you do want a screenshot anyway, so may as well implement it that way. darned sight easier than any other solution! ;)
The method works as @AzBat has explained. The games irremediably go to pause once you press the guide button, which you MUST press, and I thought it was an issue but it isn't, the screenshot is taken as is even if you are in the pause menu. Guess you never ever can take a screenshot of the pause menu nor the dashboard (this only works for those with that joined the Xbox preview program).

These are the screenshots I got as of late when the game was paused. It's from Killer Instinct, 900p native, but the game makes up for that with flawless 60 fps, it's so particles happy :smile2: and it looks good on the TV.

Took these screengrabs 'cos I like the lighting of the scenery.

edit: original images here, Tinypic compressed them.

https://onedrive.live.com/redir?res...42&authkey=!ABAs473lcbZRYNA&ithint=folder,png
 
Last edited:
Back
Top