Hi Robert,
I appreciated your details explanation. My program was using coin3D for
volume rendering previously and it was working fine on Intel graphics machine,
so I expected osg can work the same. I absolutely agree with your point of
view. I have set the minimum requirement for using osg. It is worth to
replace coin3D with OSG for the future use. Thanks again for your help and
suggestion.
Regards,
Clement
________________________________________
From: osg-users-boun...@lists.openscenegraph.org
[osg-users-boun...@lists.openscenegraph.org] On Behalf Of Robert Osfield
[robert.osfi...@gmail.com]
Sent: Monday, 19 March 2012 7:54 AM
To: OpenSceneGraph Users
Subject: Re: [osg-users] Incorrect volume image
Hi Clement,
On 18 March 2012 09:46, <clement....@csiro.au> wrote:
> Many thanks for your suggestion. osg is working perfectly on my dell e6410
> laptop, but others people can't run osg properly on their machines which are
> standard dell desktop machine with Intel graphics hardware. I want to
> overcome this problem on standard dell desktop machines. Last time you
> mentioned how to use texture2D, but it seems quite complicated and not using
> standard osgVolume api. I would like to know the minimum system requirement
> for osg especially running osgVolume. Thanks.
I can specify a minimum requirement for the OSG as it's the minimum
requirements will depend upon each applications needs, so it's really
up to the application developer what limits they choose. In the case
of osgVolume it requires a least 3D texture support, and preferably
GLSL shader support to go along with it. Practically this means
OpenGL 2.0 with proper hardware and driver support for 3D textures.
The only hardware and drivers that reliably provide this functionality
are ATI and NVidia GPU and associated drivers.
Intel hardware and drivers so far have shown them selves to be
inappropriate for supporting osgVolume, and I very much doubt this
will change in the mid term. Perhaps one day Intel will buy out
NVidia and final gets some decent graphics know how, but for now I
think it's a pretty safe bet to assume that all Intel graphics will
not be suitable and would probably be best to be detected at run-time
and an appropriate report provided to your end users.
The 2D texturing fallback you could implement, but it will never be
able to provide the same quality of results and the performance will
be so poor compared to a full osgVolume implementation rendered on
decent hardware I really don't think it would be worth the effort. As
a point of reference 3 1/2 years ago when I first start work on
osgVolume I looked at the various routes we could take with the
implementation and it was pretty clear that it didn't make much sense
to try and implement a rendering fallback using 2D texturing as the
cost of implementing and maintaining it would be highly
disproportionate to the actual capability achieved. 3 1/2 years it we
already had low cost, low end graphics card capable of doing 3D
textures, it was only the Intel embedded graphics that wouldn't be
able to handle, but they don't have the performance to do it anyway so
it's rather pointless. Back then anyone who really wants to do volume
rendering would be using a mid to high range card, one wouldn't
convenience of doing high end graphics like volume rendering on
anything less.
Fast forward to today and still doesn't make any sense to adept to
support hardware that doesn't have the feature set or performance to
do volume, and it's so cheap to purchase hardware that can handle it
it really is far far cheaper just to buy someone a new card. As for
laptops that are stuck with Intel graphics, well sorry, they have
crappy graphics, one shouldn't ever purchase low end graphics for high
end graphics tasks. Volume rendering *is* high end graphics, it's
one of the few tasks that can push modern graphics cards.
I must admit, I'm kinda surprised that I should have to telling you
this. You are the first engineer I've come across that has ever
expected volume rendering to work properly on Intel graphics.
Robert.
_______________________________________________
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
_______________________________________________
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
