Digital Foundry Article Technical Discussion Archive [2013]

[Solarus]

DF makes clarity about BF4 face-off.

i do video editing for my school, and i question this entire thing. for one thing from what ive learned you always set your hardware according to the source and even when you do, you always check to make sure everything is how it's suppose to be. in the comparison article he talks about xbo having deeper blacks; but even to new video editors they would have at least checked a histogram to see that one was producing such crushed blacks against the other two platforms. thats kind of standard practice for any video editor.

looking at ign, gametrailers and various youtube videos it really just seems that digital foundry completely dropped the ball for both xbo and ps4 capture.

 

[Arwin]

It's not rocket science Solarus. DF did the following:

- set up capturing for Xbox One, using Full RGB (0-255) at 1080p60fps and capture footage
- set up capturing for Playstation 4 using Full RGB (0-255) at 1080p60fps and capture footage
- set up capturing for PC using Full RGB (0-255) at 1080p60fps and capture footage

There is limited time, as this is done on location ....

The Xbox One footage comes out broken, because simply, the Full RGB out on the Xbox is 'broken', possibly on purpose, as this is not the first time - they did something similar with the 360. I made histograms of screenshots back then, and the 360 always crushed colors at the dark end of the spectrum, leading to irrecoverable color information loss.

Considering that ign, gametrailers and all the other footage have come from the 30fps capture hardware supplied by the Battlefield 4 team, and apparently was set to Limited RGBcolors (old fashioned TV range, which is something like 22-235 or so?), Digital Foundry did anything but drop the ball, if you ask me, being the only one who have given us full range color captures, and the only ones giving us captures at 60fps.

I'm tempted to say go back to school but then I'm sure there's nothing there about capturing video games for performance and feature analysis.

I'm sure that if DF could have realised at the time that the 360 had diverging output, they may have done some more capturing with Limited RGB settngs, but pff. I'm sure we'll get that once DF gets their hand of retail hardware.

 

[Jay]

What I don´t understand in this screenshot one the Xbox-part is this:
There are 3 display-panes, one "designated" for Hud-Display (if I remember correctly). So it should be possible to draw the texts (i.e. above the minimap) in 1080p because it is seperate from the game-rendering.

How come all texts are so blurry?

I thought that also, I put it down to time, it's a launch title (on multiple platforms no less), and it's ported from PC (I assume).
For us as consumers it is a small thing that they could have just implemented, but for them it's one of many things small and large on a list to do. It all adds up time wise.
Would it have made a difference, yes, was it worth spending time doing, maybe not if it meant not having time to do something else.

 

[BoardBonobo]

No they dont. All of them are on yt, both are heavily compressed.

He also hasnt released any direct feed screenshots to really compare.

Stupid conspiracy theories should be bannable imho.

Just check his 1080p feed. The difference is very apparent. But it's not just his capture, DF is the odd one out in all this. Unless of course, PS4 MP looks considerably better than SP...

 

[taisui]

Just check his 1080p feed. The difference is very apparent. But it's not just his capture, DF is the odd one out in all this. Unless of course, PS4 MP looks considerably better than SP...

Didn't they say that DICE supplied all the capturing device except DF, who used their own...?

 

[Solarus]

It's not rocket science Solarus. DF did the following:

- set up capturing for Xbox One, using Full RGB (0-255) at 1080p60fps and capture footage
- set up capturing for Playstation 4 using Full RGB (0-255) at 1080p60fps and capture footage
- set up capturing for PC using Full RGB (0-255) at 1080p60fps and capture footage

There is limited time, as this is done on location ....

right but are you saying that in a blind eye test "experts" wouldn't notice black crush immediately? Especially if this was done on location in which everything was setup by dice for limited RGB, it would have been noticeable to anyone that hey...Full RGB doesn't look right at all on this platform in comparison to the others. id question anyone who calls themselves and expert but doesn't spot that. doubly so if they just spent 8 years seeing it a generation before

that and you don't simply set it and forget it, you set it, capture a bit, make sure everything is correct and continue capturing. kinda like...shooting film, set your camera's various balances and levels, test to make sure everything is as you set and then continue recording.

The Xbox One footage comes out broken, because simply, the Full RGB out on the Xbox is 'broken', possibly on purpose, as this is not the first time - they did something similar with the 360. I made histograms of screenshots back then, and the 360 always crushed colors at the dark end of the spectrum, leading to irrecoverable color information loss.

right and wouldn't you have immediately noticed the samething happening yet again? its not like its hard to spot.

Considering that ign, gametrailers and all the other footage have come from the 30fps capture hardware supplied by the Battlefield 4 team, and apparently was set to Limited RGBcolors (old fashioned TV range, which is something like 22-235 or so?), Digital Foundry did anything but drop the ball, if you ask me, being the only one who have given us full range color captures, and the only ones giving us captures at 60fps.

limited is 16-235

While they might have gotten 60fps (i don't think anyone is complaining about that) the gamma is completely bonked; which is just as important in their tests as framerate, so i stand by my opinion that they dropped the ball. especially when they pride themselves on being expert and being above using every day consumer grade hardware, theyve been capturing for like 8 years it really seems like an amateur hour mistake.

my other problem is in the original article they absolutely notice the black crush to the point they were going to increase the brightness setting (this too should have been a lightbulb moment), but decide not to do anything to it? so theyre most definitely looking at all of them side by side, at that point a lightbulb should have went off in their heads immediately and realize...something is definitely wrong with Xbox and full range. i mean how are they an expert but they cant add two plus two?

I'm tempted to say go back to school but then I'm sure there's nothing there about capturing video games for performance and feature analysis.

no but most video editing classes teach you about capturing footage and checking for errors in preproduction, its not like capturing videogames is rocket science thats so far removed from how you would capture any other kind of footage.

 

[Strange]

I think the issue should be "why does Xbox One have issues displaying full range" instead of "DF f**ked up because they captured full range.

 

[Delta9]

A few gifs from gaf with all three. Xbone is going to look so bad in those big maps.

http://abload.de/img/bf4-3oma55.png
http://abload.de/img/bf4-2t5ap0.png
http://abload.de/img/bf4-18fbpa.png

 

[McHuj]

A few gifs from gaf with all three. Xbone is going to look so bad in those big maps.

Why would the size of the map matter?

 

[taisui]

X1 version looks the worst, but not sure why I catch people sharpening the PS4 image for comparison GIF, if the PS4 version already look way better.

 

[Jay]

X1 version looks the worst, but not sure why I catch people sharpening the PS4 image for comparison GIF, if the PS4 version already look way better.

Probably because they don't believe it looks _way_ better, otherwise as you said what would be the point.
It would make more sense to doctor the X1's otherwise.

 

[Brad Grenz]

X1 version looks the worst, but not sure why I catch people sharpening the PS4 image for comparison GIF, if the PS4 version already look way better.

Because there are people claiming the XO version is "more detailed" so they are adding a sharpen filter to simulate cranking the sharpen setting on your TV for people to see what that would look like.

 

[taisui]

Because there are people claiming the XO version is "more detailed" so they are adding a sharpen filter to simulate cranking the sharpen setting on your TV for people to see what that would look like.

That doesn't make any sense, the PS4 version already look better with higher rez, SSAO, and smoother fps.

 

[Jay]

That doesn't make any sense, the PS4 version already look better with higher rez, SSAO, and smoother fps.

Trouble is, people have dithering views/perceptions.
Your happy with believing that (rightly or wrongly), other people aren't so have been putting doctored images out.
To try and make it look closer to the X1, but obviously it is just to simulate that people can go and mess about with their tv settings for one game.

 

[taisui]

Trouble is, people have dithering views/perceptions.
Your happy with believing that (rightly or wrongly), other people aren't so have been putting doctored images out.
To try and make it look closer to the X1, but obviously it is just to simulate that people can go and mess about with their tv settings for one game.

You mean people actually believe in one thing but say another? I'm mind-blown

 

[Delta9]

Why would the size of the map matter?

Because the bigger the distance the more low resolution falls apart.Just look at anything at even 30 feet away and it looks garbage in comparison.

X1 version looks the worst, but not sure why I catch people sharpening the PS4 image for comparison GIF, if the PS4 version already look way better.

If you mean the ones I posted they are not changed at all.

 

[taisui]

Because the bigger the distance the more low resolution falls apart.Just look at anything at even 30 feet away and it looks garbage in comparison.

So adding distance fog will make it look better...by hiding it?

 

[blakjedi]

Does it actually look better? Or is it better because certain technical boxes were checked that support a theory of superiority? The only real difference I noted was in edge aliasing that I wouldn't have noticed without the on-screen magnifying glass...

 

[djskribbles]

I notice the differences in aliasing pretty clearly in the city maps.

 

[shredenvain]

A few gifs from gaf with all three. Xbone is going to look so bad in those big maps.

http://abload.de/img/bf4-3oma55.png
http://abload.de/img/bf4-2t5ap0.png
http://abload.de/img/bf4-18fbpa.png

I would be weary of any images coming from gaf they are altering the images in Photoshop
and are blurring one version and sharpening the other they are also changing the color depth of one version to a lower bit depth.