Oh yeah.. that that has now been called dx9.0c
DirectX9 vs DirectX10 *again*
- Thread starter alexsok
- Start date
Oh yeah.. that that has now been called dx9.0c
True, neither side has any prove, but as always when comparing new to old, they tend to compare new to very old and not previous generation.
Berek
Regular
Does anyone have another link to those images? I attempted to access them just now and cannot. I would be very interested in seeing how more realistic water is when enhanced by D3D10 effects. Hopefully we'll be seeing more physics related reactions now for water, wind(smog, smoke, etc.), and other elements.
What's in those screenshots is not any better than what can be done today with D3D9 effects.
I think you're missing the point of the difference. Its all about what's allowed speed wise. I've never gotten the impression that DirectX 10 would bring in a ton of new effects, I was under the impression that DirectX 10 allowed use of previous effects that degraded performance to much to use.
I agree. Take FarCry Instincts Predator (Xbox 360) for example... Keeping in mind that it was a port and also ran fairly well.
http://screenshots.teamxbox.com/screen/47531/Far-Cry-Instincts-Predator/
http://screenshots.teamxbox.com/screen/47532/Far-Cry-Instincts-Predator/
video:
http://xboxmovies.teamxbox.com/xbox-360/2905/Far-Cry-Instincts-PREDATOR-Water-and-View-Distance/
There was never going to be a "dx9.1", sm3.0 was there all along, .0c just added the hlsl profiles for it
Last edited by a moderator:
I think I've made similar statements numerous times on B3D, but what you've posted is pretty much the way I see it. At least in the "early days" before developers really have a solid grounding in new hardware/technology.
Simple example: You could have cube-map effects for every single object in a game under D3D9, but to have them dynamically updated would require more performance than it'd be worth. Not necessarily the raw GPU performance, but the API overhead and the general software complexity. Under D3D10 we can do this much more intuitively and efficiently via GS-based single-pass rendering, suddenly the same effect becomes a lot more feasible and developers can indeed throw these sorts of effects everywhere. But if you were to grab screenshots then the D3D10 stuff wouldn't actually look much better/different from something D3D9 could do...
hth
Jack
Yeah. I don't think I believe that. I mean, I understand the arguement. But I think the proposal was on the table, considered, and finally rejected. And that the driver for making the proposal was primarily marketing reasons, not technical ones --the technical ones against just won in the end.
Yeah, only the differences in the screenshots Microsoft provided are inaccurate and would most certainly not require DX10 for decent performance.
And definitely unavailable to 99%+ of the target audience. It's silly to even consider.
As it was meant in the “port but add nothing“ way I couldn’t agree more. But as soon as you start to use D3D10 specific functions it begins to make sense. The first games that use D3D10 will get much buzz for free.
DmitryKo
Veteran
http://www.theinquirer.net/default.aspx?article=26220, http://www.theinquirer.net/default.aspx?article=26220
etc.
Yeah, sure... why speculate if you can simply take a look at WinHEC presentations below?
etc.
Last edited by a moderator:
Because Aero needed D3D9Ex (I guess mainly the cross-processing of shared surfaces part?)
loekf
Regular
Ah got it, it's in the slides, Aero has been build upon D3D9Ex. Thought M$ used only DX10.
Not suprising considering the usual spaghetti-style SW architecture from Microsoft ;-)
