Well, that's kind of interesting result given all the recent rumbling about this.
A final result that's essentially a tie in experience although technically it sounds like the 360 slightly comes out better overall, mainly down to the higher resolution although it's clear that with the way other affects are applied you're rarely aware of the difference and in places the PS3 is actually going to look better - such as the better close up shadowing and texture filtering in places.
PS3 overall a bit smoother - a steady slightly lower fps being preferable to a much more variable one that can drop lower - which actually surprises me, but unfortunately the article doesn't contain enough detail to give clues as to why the 360 manages to have a better fps most of the time (which implies the code is running just that bit better on it) but then handles some spots worse.
And as expected rough edges with a first console game. I don't believe you can crown a graphics king as there's no way to compare different engine types (for example how do you compare this to GT5 engine or Forza? Asnwer, you can't really and the whole graphics king concept is a sham) but even if I did the engine is clearly too uneven in performance, particularly - and rather oddly I'd say - on the 360.
If I had to pick an FPS winner on tech alone it'd probably still be Killzone 3. Solid fps, better 3D and less rough edges all round - the AI on Crysis 2 for example can be a lot flakier.
Still, I'm impressed. Crytek clearly took the time to actually alter and balance to each console as best they could given what they were trying to push through the hardware. The results on both PS3 and 360 are very good overall, and while I had doubts Crytek could do it, they've produced a - mostly - graphical powerhouse on consoles.
Personally, I think it's fitting that, just like Crysis before it on PC, there's the clear sense the game is trying a bit too hard to push a bit too much through the hardware of the time, as well as the sense that here and there Crytek just couldn't quite optimize the code enough.
For me it joins GT5 in the "totally amazing in parts but just a bit too uneven in places to really consistently call the best" category.
Try to be reasonable... its easier than you think...
