[ggo-discussion] Re: another sneak preview
- From: Peter Strempel <pstrempel@xxxxxx>
- To: ggo-discussion@xxxxxxxxxxxxx
- Date: Sat, 23 Aug 2003 09:29:42 +0200
On Fri, Aug 22, 2003 at 08:58:36PM -0400, Matthew Rohn wrote:
> Also, I am having an issue with shadows that I don't think has been raised
> yet. On my WinXP Pro/Geforce3 box, stone shadows appear patchy, as in this
> screen capture:
> http://mysite.verizon.net/mdrohn/images/gIGo/shadow.jpg. Interestingly
> enough the shadows are correct on my Win98SE box with an old Nvidia Riva
> 128 card, probably because it's running glGo in software mode.
Looks the same on my old Riva TNT 1, Im getting these shadow artifacts, too.
Don't know yet what causes this, after some fruitless research I gave up on
that matter for the time being and decided to bother about that another day.
As well, when I force software mode, it looks ok here, too.
glGo so far can definately run on graphiccards from around 1998. The games
of that time were not that bad at all. The only thing so far which does not
work on my card which I miss is the multisample fullscreen antialiasing. I
only get that with software mode, so its slow. There is also much room for
a big speed booster, right now I redraw the whole scene everytime something
changes. But when a stone was placed, it should be enough to redraw only the
particular part, even with lots of captures it should hardly be more than
20% of the scene which needs to be redrawn. Of course, when rotating all
must be drawn, but I guess when playing one wouldnt rotate all the time. At
least this was a way to speed up gGo/Java significantly.
I found some docu on Nvidia page about shadows, but it's pretty high-level
for me. I will go through that another day. :) NVidia has a couple of good
articles about 3D programming, but the target reader level is not beginner.
I am learning that stuff as I do it. So far I spent probably 80% of the time
on the OpenGL board. Getting the GTP and stone-capture/storage logic ported
from gGo was a breeze, mainly copy&paste and some adjustments from Java
storage classes to C++. I also expect the IGS code will be a breeze to port,
but that currently sounds too boring to do. :)
About the circle: I'm not pleased with it. It doesn't fit in the overall
look somehow. But for SGF file loading the various marks are needed, so I
guess getting the usual circle/square/triangle/cross markers somehow
displayed is necassary.
Right now the circle is drawn seperatedly on the stone, which is somewhat
tricky as it does not fit exactly on the stone surface. For a circle it was
not that bad, but I suppose a square will be a pain. I found no way yet to
overlay the circle into the texture and get that on the exact top of the
stone. Textures over spheres are non-trivial, there is some pretty
high-level math behind it I didn't understand so far. I suppose those markers
can be done in a better way, that will require some research.
Peter
Other related posts:
