PS4 Pro Official Specifications (Codename NEO)

Delta color compression reduces render target (ROP) read and write bandwidth cost. It doesn't help anisotropic texture filtering.

Most games already use BC (DXT) compression for all textures that are anisotropic filtered. All DX11 compatible hardware supports BC compression.

 

Excuse me if this is a dumb question but isn't 2x1080 2160? Isn't 2160 4k? If so would render twice the pixels negate the need for the checker board upscale? Or are you implying that they would be upscaling in one direction only?

 

I don't know where exactly this came from, but it supposedly a side-by-side comparison between ROTR's 4K mode on the Pro and 4K maxed out on the PC:




Looks like the 4K mode will be lacking ambient occlusion, or at least it will use a form of AO that is substantially inferior to gameworks' VXAO in IQ.
Here's a close-up of Lara's face (sides were switched)










As for the resolution alone, for people sitting >1m away from a 4K TV it will be a bit hard to notice the difference between "real 4K" and 1440p + checkerboard upscale. Especially with things in motion.
Hopefully, we'll be putting the jaggies behind us completely within the next few years.

2*1080p is two 1080p screens side-by-side and the result would be a wider rectangle (32:9 instead of 16:9).
4k is two 1080p screens side-by-side on top of another two, so that aspect ratio is kept.
http://www.extremetech.com/wp-conte...-Digital_video_resolutions_VCD_to_4K.svg_.png

 

The other image has a very bad artefacts (arm).

 

alas HDR can't be represented using static framebuffer images, so I expect it will be 2007 again, back when people took off-TV photos so games looked better in screengrabs than they actually did in motion

 

Unfortunately tomb raider isn't exactly known for it's pristine image quality, so it's hard to judge anything from that.

 

4K is four times as many pixels as 1080p (twice as big on both axes). A checkerboard pattern fills in half the pixels of an image. Half the pixels of 4K is 2x the pixels of 1080p.

 

She shaves.

 

Welp. Something went wrong on PC here (her arm)

 

Maybe the PC screen was taken before they added the extra graphics options in a patch.

 

There are other objects that look sharper in the PC version. Whatever that object is to her right, for sure. Also in the PC one, the iq in her facial detail looks a bit cleaner.

But really, tomb raider is known for having f'ed up edges, so it's not the greatest game to use to judge image quality.

Edit: THe difference in quality in the strands of hair hanging to the right of her face is massive. Is that PS4 Pro image grabbed from video? I'm assuming it is. Just so hard to judge what's video compression and what's not.

 

Ps4 Pro secret sauce rendering at 8K!!!1!1

 

I don't understand what you are saying.

What specs are known for PS4 Pro?

 

The earlier leaked slides put it at 36 CUs, at least in Neo mode rather than the backwards compatibility mode. Leaked clock was 911MHz and the now public 4.2 TFLOP value fits.
Neo mode is described as activating shader engines, presumably when the extra 18 CUs are made active.

Since the PS4 already has 32 ROPs with its 2 shader engines, Neo would need 32 ROPs even when the GPU is half-off.

Going for Neo mode will turn on shader engines, and that in normal circumstances bring up a complement of extra ROPs--unless something was changed from business as usual.

 

What options are there?
Did they improve resolution of the atmospherics or the way it's composited to the image. (It was one of the most broken things in the game, 4x4 pixel pattern visible on door frames etc.)

 

Maybe 2160i (well the pattern is somehow like the i-resolution, at least the pixel-count) with 2xmsaa (and some blurring) vs 4k without any aa solution.
Or it it just the compression.

Interesting is the background, that looks sharper on the ps4p image. But this might be just happend because of some decreased depth-filter thing. Sharper is not better in this case.

 

There's also a very distinct reduction in the quality of the hair. Fewer strands and more blur. The blur could possibly be due to video compression, but still less hair overall. The DOF is also less strong in the PS4 image.

That said, the console version was never going to match all the quality settings of the PC version. The 4K upscaling doesn't really affect that, unless the quality settings are reduced compared to the 1080p PS4 version.

Regards,
SB

 

http://www.4gamer.net/games/990/G999024/20160908148/

http://pc.watch.impress.co.jp/docs/column/kaigai/1019021.html

Two informative reports.

4gamer reported the APU of PS4 Pro is fabricated in TSMC 16nm. As for the CPU side, an insider said "The basic architecture isn't different." "The composition of transistors is FinFET based, and the physical design is refined. Therefore you may not be able to declare strictly a Jaguar core and identical architecture."

The most interesting is GPU. It is not Polaris-based. In fact it is GCN 1.1 with some functions of Polaris. The reason is for backward capability. For example, PS4 Pro only has H.265 decoder, no H.265 encoder like Polaris.

PCWATCH also reported a TSMC 16nm APU. That report claims the latest "Tiger" core of Jaguar series. Both reports confirm 2.1 GHz clock rate.


The conclusions of the above are:

1. 2.1 GHz CPU (Tiger)
2. TSMC 16nm APU
3. 217~218 GB/s memory bandwidth
4. GCN 1.1 based GPU with some Polaris feature
(Is this why PS4 Pro consumes so much power because Polaris has better power efficiency in the same process?)

 

Didn't they say it was Polaris based at the conference? This seems not true.

 

Consoles usually consume half of PSU rating.