News Crysis DX9 and DX10 Comparison

hehe, yea, i thought that as well.

edit- holy poop that was fast! 1mb/s download for me! I just went with number 2 and number 4 servers.

 

These sorta looked "staged" like they removed or turned off features in the DX9 version just to make DX10 look much better.

 

I don't really beleive that, considering all the effects done in dx10 could have been done in 9 without a performance hit.

Take the Lost planet demo's for instance.

 

Well obviously not, or they'd have done it via DX9.

 

They could have just disabled it to make DX9 look worse off, besides the jungle fight in DX9 all he did was slowly move around and shoot. While in DX10 he throws a grenade simply to make DX10 "look better" with having an explosion in it. I want a TRUE video comparison where they don't pull random stuff like this to make DX9 look worse off.

 

I'll admit right off that I don't know much about GPU architecture or the real differences between DX9 and DX10....but when videos like this are presented, I think it's only fair to off the public some precise explanation of exactly how the videos were created, so here's my 2cents:

...so maybe I missed the full explanation of the videos somewhere, but I'd be interested in knowing how they chose settings (and what video cards they used) for the DX9 and DX10 tests.

For example, if they're trying to show the greatness of DX10, I'd ask:

1) Did they use the same video card for both tests?

If not, they may have chosen a "top of the line" DX9-only card (I guess this would be an Nvidia 79xx or an ATI 19xx model) and used max possible settings to keep a given framerate for the DX9 test , and then used a "top of the line" DX10 card (ie: 8800GTX Ultra) in a similar way for the DX10 test-- obviously, this would result in much better video for the DX10 test, both because of the differences in cards/settings and because of the use of DX10....This sort of test sounds "fair" on first glance, but ends up really being a test of the hardware, and not of the software differences.

2) If they used the same video card in both tests, was the GPU utilization the same in both tests?

So, say they used an 8800GTX Ultra, and they used the max settings possible while keeping 30fps while forcing either DX9 or DX10. I'd wonder if the DX10 version might use ~100% of the GPU, while the DX9 version might only be able to use, say, ~75% since it can't utilize DX10-specific hardware. On one hand, this would show that with DX10 hardware, you get much better performance on DX10 games....but on the other hand, maybe you could have done the same if you just made it all DX9 hardware so parts of the GPU weren't left idle during the DX9 test -- in which case, couldn't games and hardware companies (and consumers) have saved some time and money by just advancing DX9 technology instead of making a bigger revamp for DX10?

3) As others have pointed out, might they have used different AA, etc. settings to make the video look different as well?

 

the jungle scene when he throws the grenade and the palm tree falls of the guys head.. rofl!

 

The only real difference i could see in both videos was that in the DX9 video the shadows of the plants and tree were just dark patches on the ground, in DX10 they looked like what a actual shadow looks like. And also the lighting was much better in DX10, the rays of light filtering through the high tree's. It looks awesome!

 

Rich has kindly mirrored these for us

 

These videos originate from NVIDIA's nZone site: http://www.nzone.com/object/nzone_crysis_videos.html

Also, the footage on those videos is very old. Both scenes were shown in May 2006 during E3. I don't know why NVIDIA put the videos online. I mean, the development of Crysis has gone forward so much since those videos were released that, as far as I know, it's highly likely that they don't even reflect the state Crysis is in now and, heck, the differences might even not be that great as shown in those ancient videos.

I'm willing to bet that the "proper" comparison videos will be released when Crysis is nearing completion and that, I think, is when we'll see the true differences.

 

If you look at the Hunter video there is quite an obvious difference in the way the smoke and flames are rendered, it looks much more detailed and realistic in the DX10 video. Also when the Hunter roars the smoke and flames are blown out the way, in the DX9 video they don't move.

 

"Meh".

 

When it all comes down to it: DX10 > DX9 and that is where the future of gaming is going and there are many reasons why. The 2 top reasons why are:

a) More graphical enhancements than DX9
b) Better use of rendering these enhancements leading to improved framerates overall

Of course, a lot of those gains come in terms of visually stunning effects that you can only see/experience when you are actually gaming, not just looking at a screenshot. Of course, once the internet speed increases it'll be easier to download videos that will help show off these enhancements. I mean just d/l the DX10 SDK and you'll see of the enhancements for yourself and why DX10 is on its way.

Anyways, just wait for Crysis to hit the shelves before you start judging whether DX10 is worth it or not, or else you're doing yourself a disfavor by basing it on 'aged' E3 Demos.