We will see and you'll be amazed by how "spot-on" those predictions will turn out to be. I am very well used to people dismissing those claims and predictions outright, only to see in a couple years, that those predictions were very close to the actual results, and when I remind them of my prior predictions, then they will snap up telling me that "it was obvious man!". It's funny how the human brain can delude itself to believe in different directions at different points in time. I am just saying, those claims/predictions WILL turn out to be right. Note it down, and see for yourself, for an "I told you so" moment. This happened pretty much always with my predictions. =D

I came across this piece of info comparing the GPUs of XB1 and X360, which seems to confirm my theory of about 6x horsepower the XB1 has over X360 .

Quote:

Its GPU is only about 5-6 times faster, absolute tops

Xenos Xbox 360 GPU: 240 Gflops
Xbox One GPU: 1310 Gflops

Xenos Pixel fillrate: 4,000Mpix/s
Xbone Pixel fillrate: 13,648Mpix/s

Xenos Texel fillrate: 8,000Mtex/s
Xbone Texel fillrate: 40,944Mtex/s

I think you are kind of forgetting that each part has a knock on effect here so it's not really how things can work in the real world. to suggest that adding 4xAA would cost 1.5times the ps3 power but this is clearly just based on doing it to a 640p game, if you are talking about 4 more pass's per frame at 1080p you are talking about first multiplying the number of pixels by 3 times, then having to apply 4x AA to the new figure which would then consume 4.5x the power of a ps3 to apply.

Same thing exactly with the resolution and frame rate.

simple example

if 720p x 30 fps = 1million pixels x 30 fps means shifting 30 million pixels per second

if 1080p x 30 fps = 2 million pixels x 30 fps is indeed shifting 60 million pixels per second

but 1080p x 60 fps = 2million pixels x 60 fps is 120million pixels per second.

what I'm saying is you can't just add up that double the resolution is 3x the ps3 and double the pixels is 2x and add on some AA for 1.5x you would need to multiply each factor by the next to get how much power is needed so you are talking closer to 12x here with these 3 rather than the 8x you think it would equal.

10x the power of the PS 3 .It's called marketing 101. Get people to buy the PS 4.
Both consoles have ultra book graphics nothing more. No secret sauces on either console, (more marketing 101).
1080p,4xAA,60fps on high settings with Crysis3,Metro, etc,graphics ? Seriously? Be prepared for disappointment.

The reason I said x3 for the upping in resolution is because I was going from 640p to 1080p, which is roughly 3x pixelsand the reason I assume AA punishes the performance about 20-30% is because they usually do not directly apply it but optimize it conditional on the scene / game etc. On a PC scale, I'd agree, but on a low level console specific API level, things are more down to a direct match between bandwidth and Compute Units, rather than PC-like abstractions, which are highly diverse and unreliable. Even then, all those are gross simplifications. I am not claiming any specific statement is accurate, but I am claiming that the overall results will be very much like what I am expecting.

You don' get twice the graphics detail buy cutting the frame rate in half. It doesn't work that way. Many other factors are involved.

Edit, The PS4 has approximately 40% to50% more GPU power XB1. The games won't look 50% better either. Doesn't work that way. Many factors involved also.

That is just not true, Mark Cerny is anything but marketing gimmicks. Yes he'd try to give a good PR impression to boost the trust on the console and try to be and seem as gaming focused as possible, but this guy is not about marketing gimmicks. He did not and will not say one thing which is  not at least theoretically correct. Now take a look at this comparison between X360 and XB1 GPUs;

As I quoted above, XB1 is around 6x times performance of X360. Assuming platform parity this past gen between X360 & PS3, and given that PS4 is readily 50% faster than XB1 (the gap is larger than HD7870 vs 7790), it's logical that PS4 is roughly 10x the performance of PS3. It has 16x memory to say the least.

As a reference:

As a comparison, XB1 GPU is an underclocked 7790 with much slower memory (despite the fast esram). So the gap is larger than the simple comparison below.

http://www.hwcompare.com/14298/radeon-hd-7790-vs-radeon-hd-7870/

 http://www.videocardbenchmark.net/gpu.php?gpu=Radeon+HD+7870&id=324

http://www.videocardbenchmark.net/gpu.php?gpu=Radeon+HD+7790&id=2502

I like your analysis, even if it is simplified as you quite rightly state. It explains why next-gen graphics will look only marginally better than prev-gen - the increase in frame rate and resolution will eat away a lot of the PS4/XB1's extra power.

There are some notable exceptions though. With hardware support for things like tessellation, improved shaders and other effects incorporated into the next-gen GPU's, you can get additional improvements for some things without requiring additional power. So there will be "free" benefits like improved lighting and more realistic surfaces. Whether this is going to provide the "WOW!" factor like Directx 9.0c shaders did for last-gen is going to be interesting to see (my guess is it will only be notable for graphics geeks, not so much for the average gamer).

10x the power doesn't mean 10x the graphics. The law of diminishing returns and other factors prevent that. Will the graphics look better? Sure. But not by a factor of 10. And not like the PS1 to PS2 to PS3 graphics differences either.

I estimate it at 9-15, and 5 times is laughably low...