openvrml-develop — OpenVRML developers mailing list

[openvrml-develop] SimVRMLLookat Viewer

[John R. <ric...@sp...>]

Hello,

I will be at Macworld trying to show off a modified MacLookat VRML viewer.
It is called SimVRMLLoookat. It is MacLookat modified to accept Apple
Events from SuperCard and Labview. It consists of the SimVRMLLookat viewer.
The SimVRML RPT (Rapid Prototyping Tool), a user guide and test virtual
environments. Mostly architecture.

SuperCard is functionally equivalent and I feel superior to Macromedia
Director 8 and this API goes a long way to making it equivalent to Director
8.5 with shockwave 3D. The reference implementation of the API is created
in SuperCard. It is really cool, if I must say so.....

Labview is for real time data acquisition and for making scientific UI's.
I'll be at the Lego Mindstormes BOF. Mindstormes uses a sort of Labview
lite called robolab.

The modifications have an API. It is speech input enabled. Speech requires
a decent microphone. The SuperCard development environment is cheap (~
$100). Robolab is cheap and can be found in toy stores (pisco Lego-Dacta
makes it). The full Labview by National Instruments is $2000 or more.

The viewer is anticipated to be up on source forge in mid January.

It is not evident in this release, but the goal is to add generic core
simulation functionality to OpenVRML. The API gives more information.

ACKNOWLEDGEMENTS: Michael Louka who made MacLookat. Mike has been very
helpful. Also, all the others involved in OpenVRML shuch as Chris Morley,
Branden McDaniel, S.K. Bose,.......

John F. Richardson

[openvrml-develop] Re: LOD processing

[Damian M. <da...@er...>]

On Mon, 7 Jan 2002, Damian McGuckin wrote:

> During such processing, let's assume we have a single VRML for most of
> the VRML description, but that the various levels of any LOD nodes are
> stored as some Inline URLs.

I hit send before I meant to.

> Does the Parser expand/read all the URLs at the time the primary file is
> being initially parsed or when you are close enough to the object
> described by the node that a particular level of that node is rendered
> within VrmlNodeLOD::render.

The former appears to be the case.

However, ideally in some scenarios, you would like only the basest level
of a NODE to be automatically loaded and then as you got close enough to
an object, import the more details VRML description to give you some sort
of on-demand processing facility.

Any ideas on how to go about this?

Thanks - Damian (McGuckin)

Pacific Engineering Systems International, 22/8 Campbell St, Artarmon NSW 2064
Ph:+61-2-99063377 .. Fx:+61-2-99063468   | unsolicited email not wanted here !
Views and opinions here are mine and not those of any past or present employer

[openvrml-develop] Re: Installation of GTKLOOKAT

[Damian M. <da...@er...>]

> > Damian McGuckin had the following problem building GTKLOOKAT from
> > sources
> >
> > aclocal: configure.in: 9: macro `AM_PATH_GTK' not found in library
> >
> > It won't be found in /usr/local/share/aclocal because it came with the
> > release of the O/S.
> >
> > What's the problem?
> >

> Thomas Flynn had another problem
>
> I don't get this problem on my system but instead get the following
> error when running configure
>
> checking for gtk-config... /usr/bin/gtk-config
> checking for GTK - version >= 1.2.0... yes
> ./configure: line 2599: syntax error near unexpected token `OV_PATH_GTKGL(,'
> ./configure: line 2599: `OV_PATH_GTKGL(, { { echo "$as_me:$LINENO:
> error: Missing GtkGLArea. See 'config.log' for details." >&5'
>
> GtkGLArea is correctly installed on my system under /usr
> [tflynn@cr471342-a gtklookat]$ rpm -q gtkglarea
> gtkglarea-1.2.2-10

My results are

	$ /usr/bin/gtk-config --version
	1.2.10
	$

and also

	$ /usr/bin/gtk-config --cflags
	-I/usr/include/gtk-1.2 -I/usr/include/glib-1.2 \\
	-I/usr/lib/glib/include -I/usr/X11R6/include

and
	$ /usr/bin/gtk-config --libs
	-L/usr/lib -L/usr/X11R6/lib -lgtk -lgdk -rdynamic \\
	-lgmodule -lglib -ldl -lXi -lXext -lX11 -lm
$

where I have deliberately added in the \\ to imply that the line is really
wrapped.

And because GTK-1.2 is preinstalled ( rather than RPM'd ) on the latest
release of Mandrake Linux, the next command will give a very predicatable
result.

	$ rpm -q gtkglarea
	package gtk is not installed

Obviously the installation process is relying on some macro set-up by the
rpm.  Any ideas what it is?  Tom, do you have some m4 macro related to gtk
in the area /usr/local/share/aclocal, or anywhere else for that matter?

In the end, I just copied the 2 Makefiles from the previous version of
gtklookat-0.9.1 and the executable compiled and built fine.  However, this
expediency is not the answer. Something seems non-robust. We shouldn't
need to install some new release of GTK just to get OpenVRML working.  I
just can't pin-point it. Maybe we need an m4 macro for those who don't
have got GTK rpm'd on their system.  Maybe I'm just talking through my
hat.

Many Thanks - Damian (McGuckin)

Pacific Engineering Systems International, 22/8 Campbell St, Artarmon NSW 2064
Ph:+61-2-99063377 .. Fx:+61-2-99063468   | unsolicited email not wanted here !
Views and opinions here are mine and not those of any past or present employer

Re: [openvrml-develop] [Fwd: [www-vrml] Problems with VRML viewers under Linux]

[clayton c. <dr...@sm...>]

I dont know if you took a look but there was some coding errors on his
part for instance he had

Transform{
set_scale is setscale
children[
...
]

}

so i sent that back to him to check , probaly the windows viewers just
by passed or have some psuedo parsing for this or ???

[openvrml-develop] LOD processing

[Damian M. <da...@er...>]

During such processing, let's assume we have a single VRML for most of
the VRML description, but that the various levels of any LOD nodes are
stored as some Inline URLs.

Does the Parser expand/read all the URLs at the time the primary file is
being initially parses or when you are close enough to the object
described by the node that a particular level of that node is rendered
within VrmlNodeLOD::render.

Thanks - Damian (McGuckin)

Pacific Engineering Systems International, 22/8 Campbell St, Artarmon NSW 2064
Ph:+61-2-99063377 .. Fx:+61-2-99063468   | unsolicited email not wanted here !
Views and opinions here are mine and not those of any past or present employer

[openvrml-develop] orientationInterpolator

[Vieri Di P. <vie...@ya...>]

I have two identical scenes. The first contains an
object rotating clockwise. In the second, the object
rotates counterclockwise. I would like to load the
scene once and only modify the "key values" of the
OrientationInterpolator when necessary. Is this
possible? 
Regards,
V



__________________________________________________
Do You Yahoo!?
Send FREE video emails in Yahoo! Mail!
http://promo.yahoo.com/videomail/

[openvrml-develop] orientationInterpolator

[Vieri Di P. <vie...@ya...>]

I have two identical scenes. The first contains an
object rotating clockwise. In the second, the object
rotates counterclockwise. I would like to load the
scene once and only modify the "key values" of the
OrientationInterpolator when necessary. Is this
possible? 
Regards,
V



__________________________________________________
Do You Yahoo!?
Send FREE video emails in Yahoo! Mail!
http://promo.yahoo.com/videomail/

Re: [openvrml-develop] Re: Installation of Latest GTKLOOKAT CVS sources

[Thomas F. <tf...@ro...>]

Damian McGuckin wrote:

> I downloaded this.
> 
> I installed openvrml in the default location.
> 
> I have GTK-1.2 installed in /usr/lib such as /usr/lib/libgtk.a and the
> includes are in /usr/include/gtk-1.2
> 
> Typing
> 
> 	aclocal -I /usr/local/share/aclocal
> 
> and it says
> 
> aclocal: configure.in: 9: macro `AM_PATH_GTK' not found in library
> 
> It won't be found in /usr/local/share/aclocal because it came with the
> release of the O/S.
> 
> What's the problem?
> 


I don't get this problem on my system but instead get the following 
error when running configure

checking for gtk-config... /usr/bin/gtk-config
checking for GTK - version >= 1.2.0... yes
./configure: line 2599: syntax error near unexpected token `OV_PATH_GTKGL(,'
./configure: line 2599: `OV_PATH_GTKGL(, { { echo "$as_me:$LINENO: 
error: Missing GtkGLArea. See 'config.log' for details." >&5'

GtkGLArea is correctly installed on my system under /usr
[tflynn@cr471342-a gtklookat]$ rpm -q gtkglarea
gtkglarea-1.2.2-10

Does anyone know what's going on here?

Thanks,
Tom

[openvrml-develop] [Fwd: [www-vrml] Problems with VRML viewers under Linux]

[Thomas F. <tf...@ro...>]

The following message was posted to www-vrml last week from a guy trying 
to get his world to load in ANY browser under Linux. His results are 
posted below with OpenVRML 0.11.2

I've tried his world (attached to this mail) with the latest version of 
CVS and get the following assertion failure on load

../../../../src/openvrml/OpenVRML/nodeptr.h:94: OpenVRML::Node* 
OpenVRML::NodePtr::operator->() const: Assertion `this->countPtr' failed.

 with the following stack trace.

in OpenVRML::VrmlScene::load() at VrmlScene.cpp:286
in OpenVRML::VrmlScene::replaceWorld() at node.ptr.h:93
in OpenVRML::ProtoNode::addToScene() at proto.cpp:292
in OpenVRML::ProtoNode::instantiate() at proto.cpp:188
in OpenVRML::VrmlNamespace::cloneNodes() at VrmlNamespace.cpp:246
in () at VrmlNamespace.cpp:724
in OpenVRML::NodeTransform::accept() at vrml97node.cpp:9238
in () at nodeptr.h:99
in () at VrmlNamespace.cpp:724
in OpenVRML::NodeCollision::accept() at vrml97node.cpp:1519
in () at nodeptr.h:93
in OpenVRML::NodeGroup::setChildren() at vrml97node.cpp:3632



This world loads fine in all the big browsers on Windows. Anyone have 
any insight as to what's going wrong here?

Later,
Tom

-------- Original Message --------
Subject: [www-vrml] Problems with VRML viewers under Linux
Date: Thu, 03 Jan 2002 17:43:33 -0600
From: Aleksandar Donev <ad...@Pr...>
Reply-To: Aleksandar Donev <ad...@Pr...>
Organization: Princeton Materials Institute, Princeton University
To: www...@we...
CC: Aleksandar Donev <ad...@Pr...>, "John A. Stewart" 
<joh...@cr...>, Tom Flynn <tf...@us...>



Hello,

I just got back to my Linux box and tried to install a VRML viewer to 
look at my animation, and am having trouble getting a browser to display 
the attached file RSA.25-1.0.wrl (there are PROTOs defined in the other 
attached file). The files work just fine under all the Windows browsers 
I've tried. For example, here are the errors I get:

FreeWRL (installation was long but simple):
Invalid field 'set_scale' for node 'Transform'

OpenVRML (installation was a pain due to SpiderMonkey, which is 
installed on my machine under Mozilla, and is almost impossible to 
install standalone. Had to tinker with configure file to make it and use 
the right libmozjs and header files):
lookat: VrmlNamespace.cpp:1067: void 
{unnamed}::NodeRouteCopyVisitor::copyRoutesFromNode (VrmlNode &): 
Assertion `fromNode' failed.
Aborted (core dumped)

Both of these work with other (simpler) VRML models I used for testing!

Xj3D browser not yet installed due to painful installation 
procedure...Clearer and shorter instructions, rpm's, and other widgets 
would be most welcome :)

Could someone that has VRML under Linux running please try and see if 
they can get the attached file to work and let me know.

Thank you,
Aleksandar

-- 
__________________________________
Aleksandar Donev
Complex Materials Theory Group (http://cherrypit.princeton.edu/)
Princeton Materials Institute & Program in Applied and Computational Mathematics 
@ Princeton University
Address:
   419 Bowen Hall, 70 Prospect Avenue
   Princeton University
   Princeton, NJ 08540-5211
E-mail: ad...@pr...
WWW: http://atom.princeton.edu/donev
Phone: (609) 258-2775
Fax: (609) 258-6878     
__________________________________

 

Re: [openvrml-develop] Re: Vrml97Parser Namespaces

[<br...@en...>]

Quoting Damian McGuckin <da...@er...>:

> 
> I am trying to retrofit my changes to Vrml97Parser.cpp
which inline
> some of the heaviest used routines and same squillions
of subroutine
> operations.
> 
> Compiling with the file as pulled down off the CVS
tree, when compiling
> VrmlScene.cpp, you get the message
> 
> g++ -DHAVE_CONFIG_H -I.
-I../../../../src/openvrml/OpenVRML -I../../..
> -I../../../../lib -DOPENVRML_LIBDIR_=\"/usr/local/lib\"
> -DOPENVRML_PKGDATADIR_=\"/usr/local/share/openvrml\"
-g -O2 -c
> ../../../../src/openvrml/OpenVRML/VrmlScene.cpp -MT
VrmlScene.lo -MD -MP
> -MF .deps/VrmlScene.TPlo  -fPIC -DPIC -o VrmlScene.o
> In file included from Vrml97Parser.g:112,
>                  from
../../../../src/openvrml/OpenVRML/VrmlScene.cpp:50:
> Vrml97Parser.hpp:318: warning: This file contains more
`{'s than `}'s.
> 
> It compiles OK but I HATE messages like this.
> 
> What does this mean?

Heh, heh... It's harmless. I don't know how to make it
go away, because it's a perfectly valid complaint. I'm
just outsmarting it.

ANTLR's rather inflexible about the scope in which it
can generate a parser, so I'm forcing the issue here. I
wanted to have the parser inside an anonymous namespace
so it would not be a part of the public API. Because of
ANTLR's peculiarities, I had to accomplish this by
opening the namespace in the header file, and closing it
in the implementation file.

Braden

[openvrml-develop] Re: Installation of Latest GTKLOOKAT CVS sources

[Damian M. <da...@er...>]

I downloaded this.

I installed openvrml in the default location.

I have GTK-1.2 installed in /usr/lib such as /usr/lib/libgtk.a and the
includes are in /usr/include/gtk-1.2

Typing

	aclocal -I /usr/local/share/aclocal

and it says

aclocal: configure.in: 9: macro `AM_PATH_GTK' not found in library

It won't be found in /usr/local/share/aclocal because it came with the
release of the O/S.

What's the problem?

Thanks - Damian (McGuckin)

Pacific Engineering Systems International, 22/8 Campbell St, Artarmon NSW 2064
Ph:+61-2-99063377 .. Fx:+61-2-99063468   | unsolicited email not wanted here !
Views and opinions here are mine and not those of any past or present employer

[openvrml-develop] Re: Vrml97Parser Namespaces

[Damian M. <da...@er...>]

I am trying to retrofit my changes to Vrml97Parser.cpp which inline
some of the heaviest used routines and same squillions of subroutine
operations.

Compiling with the file as pulled down off the CVS tree, when compiling
VrmlScene.cpp, you get the message

g++ -DHAVE_CONFIG_H -I. -I../../../../src/openvrml/OpenVRML -I../../..
-I../../../../lib -DOPENVRML_LIBDIR_=\"/usr/local/lib\"
-DOPENVRML_PKGDATADIR_=\"/usr/local/share/openvrml\" -g -O2 -c
../../../../src/openvrml/OpenVRML/VrmlScene.cpp -MT VrmlScene.lo -MD -MP
-MF .deps/VrmlScene.TPlo  -fPIC -DPIC -o VrmlScene.o
In file included from Vrml97Parser.g:112,
                 from ../../../../src/openvrml/OpenVRML/VrmlScene.cpp:50:
Vrml97Parser.hpp:318: warning: This file contains more `{'s than `}'s.

It compiles OK but I HATE messages like this.

What does this mean?

Thanks - Damian (McGuckin)

Pacific Engineering Systems International, 22/8 Campbell St, Artarmon NSW 2064
Ph:+61-2-99063377 .. Fx:+61-2-99063468   | unsolicited email not wanted here !
Views and opinions here are mine and not those of any past or present employer

[openvrml-develop] Fw: qtlookat patch to enable win32 build

[nldudok1 <nld...@ol...>]

 Hi openvrml developers,
 I made a patch to the qtlookat code so qtlookat can be build on win32 with
Visual C++. I know there is already a native win32 binary but qt is
multiplatform and with qt it is very easy to add GUI features.
 I also have a binary but I don't know how to submit it to the openvrml
 project.
 Should I send it to one of the project
 admins or the mailing list?
 Also haven't I tested if qtlookat still builds correctly on unix (it
should)
 Greetings,
 Martin Dudok van Heel

[openvrml-develop] loadFromString

[Vieri Di P. <vie...@ya...>]

Hi. I'd like to know if there's a way to load a vrml
scene from memory, not from a file. What I do is put
the node information (just like a regular wrl file)
into an ostrstream. Then I call
myvrmlScene->loadFromString(myostrstream.str()).
Shouldn't this replace the
myvrmlScene->load(wrlfilename..) method correctly? My
problem is that when I call loadFromString I get a
segmentation fault. I was wondering if someone could
help me out. 
Thanks in advance,
Vieri
P.S.: my system is Linux Mandrake 8.1 and the openvrml
lib version is recent.


__________________________________________________
Do You Yahoo!?
Send your FREE holiday greetings online!
http://greetings.yahoo.com

RE: [openvrml-develop] Transparency

[bose <bo...@pa...>]

> -----Original Message-----
> From: ope...@li...
> [mailto:ope...@li...]On Behalf Of Damian
> McGuckin
> Sent: Friday, December 28, 2001 4:55 PM
> To: ope...@li...
> Subject: [openvrml-develop] Transparency
>
>
>
> This has been open for over a year.  I'd like to close it.
>
> There are two issues with transparency, one associated with materials and
> one with images that have transparent parts.  Apparently neither work.
>
> I have looked at the bug page and it doesn't give me much of a hint except
> to say that with z-buffers it is a problem.
>
> In 'ViewerOpenGL::insertTexture', it mentioned
>
> 	// Enable blending if needed
>   	if (d_blend && (nc == 2 || nc == 4))
>     		glEnable(GL_BLEND);
>
> Don't we need an extra
>
> 	glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
>
> or does the one in ViewerOpenGL::initialize cover us forever.  There is a
> glDisable(GL_BLEND) in ViewerOpenGL::redraw() and I thought that this
> clobbered the glBlendFunc().

If you define it once then again you don't need to redefine, unless
you want to change the source and destination parameters.
glDisable(GL_BLEND) has no effect on these two parameters.

> Heaven only knows what to do with Material transparency.
>
> Can anybody direct me to where there are some simple files to try this
> out?  And/or any theory that might be useful especially in the context of
> OpenVRML and its usage in the context of OpenGL.
>
Rendering code for OpenVRML is based on OpenGL. For implementing
transparency one can make use of OpenGL "Alpha Test" features. Render your
object list twice, the first time accepting only fragments with alpha values
of one (all the opaque objects), and the second time accepting fragments
with alpha values that aren't equal to one (all translucent objects). Turn
the depth buffer on during both passes, but disable depth buffer writing
during second pass (Read only). When the translucent objects are drawn,
their depth values are compared to the values established by the opaque
objects (Z-buffer algo.), so they aren't drawn if they are behind the opaque
ones. If they're closer to the viewpoint, however, they don't eliminate the
opaque objects, because the depth buffer values can't change (Read only).
Instead, they are blended with the opaque objects.

But mind that there will be drop in performance due to two passes for the
same scene.
For more details you can see the Chapter 10 of "The Official Guide to
Learning OpenGL" red book (Addison-Wesley Publishing Company).

Happy New Year

Thanks,

S. K. Bose

[openvrml-develop] Re: Numbers Close together

[Damian M. <da...@er...>]

While there is an inline routine, fpequal(a, b) which tests whether 2
numbers are close together to a floating point tolerance, there are
3 instances where this is not used but maybe should be.

field.cpp
		(1.0 - cosom) > FPTOLERANCE)		Anonymous

vrml97node.cpp
		(dotval+1.0) > FPTOLERANCE)		S.K.Bose
 		(1.0-dotval) > FPTOLERANCE)

which would appear to be saying

		cosom is damn close to +1
+
		dotval is damn close to -1
+
		dotval is damn close to +1

respectively.

If this is the case, then the way this is written will cause some
cases, where what I think they are saying, will not actually happen.

Is this a potential bug or am I being paranoid?

Also, FPTOLERANCE is 1.0e-6 in the main library, but 1.0e-7 in the
OpenGL stuff.  Should they be consistent?

And

	const float FPTOLERANCE(1.0e-6);

	inline fpzero(float a)
	{
		return fabs(b) <= FPTOLERANCE;
	}

will probably kill the optimizer because it will trigger a call to fabs.
If you really want it inline, it will often be faster to do something like

	inline fpzero(float a)
	{
		static float FPTOLERANCE = 1.0e-6;

		return (-FPTOLERANCE <= a && a <= FPTOLERANCE);
	}
or

	inline fpzero(float a)
	{
		static float FPTOLERANCE = 1.0e-6;

		return ((a < 0.0 ? -a : a) <= FPTOLERANCE);
	}

and I hope I got that right.

If you want really optimal and machine portable code of a number within 3
binary digits of the machine precison, you can try

	inline fpzero(float a)
	{
		return 1.0f + a * 0.125 == 1.0f;
	}

but most optimizers kill portability of this by almost always returning
true so we instead do

	inline fpadd1(float a)
	{
		return 1.0f + a * 0.125;
	}

	inline fpzero(float a)
	{
		return fpadd1(a) == 1.0f;
	}

to at least fool g++/gcc.  I've yet to try it on some other compilers.

I hope we all understood this while we are nursing our sore heads from
last night.

Many Thanks - Damian (McGuckin)

Pacific Engineering Systems International, 22/8 Campbell St, Artarmon NSW 2064
Ph:+61-2-99063377 .. Fx:+61-2-99063468   | unsolicited email not wanted here !
Views and opinions here are mine and not those of any past or present employer

[openvrml-develop] Re: Transparency

[Greg R. <ne...@po...>]

Tom Flynn <tf...@us...> wrote:

> Damian McGuckin wrote:

> >This has been open for over a year.  I'd like to close it.
> >
> >There are two issues with transparency, one associated with materials and
> >one with images that have transparent parts.  Apparently neither work.
> >
> ><snip>
> >
> >Heaven only knows what to do with Material transparency.
> >
> >Can anybody direct me to where there are some simple files to try this
> >out?  And/or any theory that might be useful especially in the context of
> >OpenVRML and its usage in the context of OpenGL.
> >

> There are a couple of simple transparency examples available with the 
> NIST Test Suite at 
> http://xsun.sdct.itl.nist.gov/~mkass/vts/html/vrml.html . Unfortunately 
> I don't have a pointer to which tests would be of interest though ... 
> I'll try and post some URLs after I get back home from holidays next week

I believe I've got a fairly complete set of possible transparency combos
here:

	http://www.libpng.org/pub/png/pngvrml.html

(specifically, http://www.libpng.org/pub/png/pngvrml/pngboxes.wrl, but
the link above has screenshots of how each major section _should_ look,
except that some of the transparent holes are too small to discern.)

Also, all of the textures are included on this page as standard HTML images:

	http://www.libpng.org/pub/png/png-textures.html

The holes are more visible here.

Greg

-- 
Greg Roelofs            ne...@po...             http://pobox.com/~newt/
Newtware, PNG Group, Info-ZIP, AlphaWorld Map, Philips Semiconductors, ...

[openvrml-develop] Re: Doc2 - adding big buffer

[Damian M. <da...@er...>]

> >On Sun, 23 Dec 2001, Damian McGuckin wrote:
> >
> >>I need to increase the speed of loading big, multi-megabyte VRML files.
> >>
> >The change was a lot simpler than what I had tried in the first place.
> >
> >I changed the size of the buffer declared in doc2 to 65535 and achieved
> >nearly a factor of two performance gain on big files.  The load time went
> >from 44 seconds down to 24 on a 1000Mhz Pentium over a 100Mbit/sec
> >network.
> >
> >Not as much as I had hoped but better than nothing.
> >
>
> How did this change affect the load time for smaller worlds? I don't
> imagine it would cause too big of a hit but I'd like to do some testing
> to be sure. Maybe the buffer size should be a command-line arg of
> lookat...or maybe the default should just be bumped up to a larger
> number...if you've already done some more investigation here let me know
> so I don't duplicate work.

The files with which I was playing were up to 6MB big.

For small worlds coming off a local disk, what I did made made absolutely
no difference.

For big worlds comming off a local disk, the change was marginal but
detectable if you measured it.

Both the above tests were IDE disks and 800Mhz Pentium III under XP.
Using a slower machine under NT, the change was almost independent of
processor speed.

However, the timings I quoted in my earlier email were based on loading a
file coming off an NFS network disk. It was a 100Mb/sec network and
1000Mhz Pentium under Linux with the display being an X terminal.  The
buffer size of 10 was woeful.  I'll try a 10Mb/sec network later.

For a 6MB world, a buffer size of 65535 made basically the same impact as
a 6MB buffer. Using 32767 was almost as good but detectablely slower.

While my short term interest is in making sure that my development & test
environment works well and that displays to others using it also work
well, my guess is that even small worlds over today's slowest ADSL links
will approach closely this benchmark scenario. If this is what people want
to use this stuff for, then such applications will need some reasonable
buffering.

As my world ends up taking 90MB when fully loaded, I figured grabbing an
extra 64KB from memory was not a big penalty.  However, it might be better
implemented as a run-time parameter.  Mind you, I also wanted something
where the change to the source code was as simple as possible and you
can't do much better than a 1 line change.  I figured even I can't
introduce a bug when .... "all I did was change one line"!!!!

Thanks - Damian (McGuckin)

Pacific Engineering Systems International, 22/8 Campbell St, Artarmon NSW 2064
Ph:+61-2-99063377 .. Fx:+61-2-99063468   | unsolicited email not wanted here !
Views and opinions here are mine and not those of any past or present employer

[openvrml-develop] Re: Matrix Operations

[Damian M. <da...@er...>]

Is there any reason why we can't have most of the Matrix Operations in
Single Precision ( i.e. float ).  It appears to me that this would save
a lot of work in many places.

Just my $0.02c worth.

Regards - Damian (McGuckin)

Pacific Engineering Systems International, 22/8 Campbell St, Artarmon NSW 2064
Ph:+61-2-99063377 .. Fx:+61-2-99063468   | unsolicited email not wanted here !
Views and opinions here are mine and not those of any past or present employer

Re: [openvrml-develop] Re: Transparency

[Christopher K. S. J. <ck...@di...>]

Damian McGuckin wrote:
> 
> > Sort them once at the begining.  Then resort on every single change in
> > viewpoint. Assume that as you move around the list will remain mostly
> > sorted, and choose a sorting algorithm to fit that assumption.
> 
> If you rotate yourself, say 90 degrees, doesn't the list get horribly out
> of sort.  Mind you, this is simply a Z-buffer problem.
> 

 The theory is that there should be many intermediate
steps in between, each of which results in only a small
numbers of postion swaps in the sorted list. Spatial
coherency is a fairly common assumption for real time
scene graphs, even though, as you point out, it's not
always true.

 Of course, it's only the transparent surfaces that
need to be depth sorted, so as long as you keep to
a reasonable number of them, the sorting shouldn't be
a big deal.

 A good intermediate step might be to keep all the
transparent surfaces in an unsorted list and render
them last. That's not quite right (transparent
surfaces layered over other transparent surfaces
would have subtle artifacts), but would be much better
than the current situation (transparent surfaces
layered in with any surface at all produce very ugly
artifacts)

 In the bad old days when processor time was expensive
and 3d harware was rare, you also had "screen door"
transparency. It's ugly, but relatively easy to 
implement, since you don't have to keep a separate
(sorted) list of the transparent surfaces. A google
search should turn up a bunch of references.



-- 
Christopher St. John ck...@di...
DistribuTopia http://www.distributopia.com

[openvrml-develop] Re: Transparency

[Damian M. <da...@er...>]

Christopher St John wrote:

> Sort them once at the begining.  Then resort on every single change in
> viewpoint. Assume that as you move around the list will remain mostly
> sorted, and choose a sorting algorithm to fit that assumption.

If you rotate yourself, say 90 degrees, doesn't the list get horribly out
of sort.  Mind you, this is simply a Z-buffer problem.

> If you hunt around on the net, there are some detailed explanations of
> how it all should work, I seem to remember the SGI Performer docs have
> a discussion.

I'll go fishing.

On the underlying graphics library issues, are there some notes floating
around anywhere on the overall design of the Scene Graph, and how this is
then mapped into the OpenGL library.

I also noticed that there is a significant processing load caused by the
virtual function ViewOpenGL::insertReference which adds a reference into
the call list.  In our test cases, during loading of the large VRML file,
this causes squillions of calls to glCallList().  While a virtual function
can create load, this appears to create nearly half the processing load of
tessShellVertex(), gluTessEndPolygon(), tessShellBegin() & insertShell()
combined so I don't believe it is just the 'virtual'ness of the function.
I'm wondering if there was a way to buffer up these calls, say N at a
time, and send them out in a single call to glCallLists().  Mind you, I
can't see the optimal spot where I would flush out those remaining in the
buffer when the scene was fully interpreted.

Thanks - Damian (McGuckin)

Pacific Engineering Systems International, 22/8 Campbell St, Artarmon NSW 2064
Ph:+61-2-99063377 .. Fx:+61-2-99063468   | unsolicited email not wanted here !
Views and opinions here are mine and not those of any past or present employer

Re: [openvrml-develop] Transparency

[Tom F. <tf...@us...>]

Damian McGuckin wrote:

>This has been open for over a year.  I'd like to close it.
>
>There are two issues with transparency, one associated with materials and
>one with images that have transparent parts.  Apparently neither work.
>
><snip>
>
>Heaven only knows what to do with Material transparency.
>
>Can anybody direct me to where there are some simple files to try this
>out?  And/or any theory that might be useful especially in the context of
>OpenVRML and its usage in the context of OpenGL.
>

There are a couple of simple transparency examples available with the 
NIST Test Suite at 
http://xsun.sdct.itl.nist.gov/~mkass/vts/html/vrml.html . Unfortunately 
I don't have a pointer to which tests would be of interest though ... 
I'll try and post some URLs after I get back home from holidays next week

Later,
Tom

Re: [openvrml-develop] Re: Doc2 - adding big buffer

[Tom F. <tf...@us...>]

Catching up on old email...

Damian McGuckin wrote:

>On Sun, 23 Dec 2001, Damian McGuckin wrote:
>
>>I need to increase the speed of loading big, multi-megabyte VRML files.
>>
>The change was a lot simpler than what I had tried in the first place.
>
>I changed the size of the buffer declared in doc2 to 65535 and achieved
>nearly a factor of two performance gain on big files.  The load time went
>from 44 seconds down to 24 on a 1000Mhz Pentium over a 100Mbit/sec
>network.
>
>Not as much as I had hoped but better than nothing.
>

How did this change affect the load time for smaller worlds? I don't 
imagine it would cause too big of a hit but I'd like to do some testing 
to be sure. Maybe the buffer size should be a command-line arg of 
lookat...or maybe the default should just be bumped up to a larger 
number...if you've already done some more investigation here let me know 
so I don't duplicate work.

Later,
Tom

Re: [openvrml-develop] Re: Transparency

[Christopher K. S. J. <ck...@di...>]

Damian McGuckin wrote:
> 
> And then, how do we change the depth-sorted order as we move around
> within the VRML world?
> 

 Sort them once at the begining.  Then resort on 
every single change in viewpoint. Assume that as
you move around the list will remain mostly sorted,
and choose a sorting algorithm to fit that assumption.
If you hunt around on the net, there are some detailed
explanations of how it all should work, I seem to
remember the SGI Performer docs have a discussion.


-- 
Christopher St. John ck...@di...
DistribuTopia http://www.distributopia.com

[openvrml-develop] Re: Transparency

[Damian M. <da...@er...>]

On Fri, 28 Dec 2001, Damian McGuckin wrote:

> In 'ViewerOpenGL::insertTexture', it mentioned
>
> 	// Enable blending if needed
>   	if (d_blend && (nc == 2 || nc == 4))
>     		glEnable(GL_BLEND);
>
> Don't we need an extra
>
> 	glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
>
> or does the one in ViewerOpenGL::initialize cover us forever.  There is a
> glDisable(GL_BLEND) in ViewerOpenGL::redraw() and I thought that this
> clobbered the glBlendFunc().  I am a long way away from being an OpenGL
> guru.

And glBlendFunc really wants primitives sorted in Depth order.  Aren't
then loaded them in the order we found them in the VRML file in the first
place. And then, how do we change the depth-sorted order as we move around
within the VRML world?

Questions, questions?

Thanks - Damian (McGuckin)

Pacific Engineering Systems International, 22/8 Campbell St, Artarmon NSW 2064
Ph:+61-2-99063377 .. Fx:+61-2-99063468   | unsolicited email not wanted here !
Views and opinions here are mine and not those of any past or present employer