Menu
▾
▴
[gl4java-usergroup] Capabilities of GL4Java state machine?
|
From: Mark M. <pat...@lm...> - 2003-03-26 23:16:50
|
Should have included the following in my last posting... (again, I thank =
you
for any help you can provide)
When Java3D queries the max-supported-texture-size on the Wildcat III 611=
0
machine (from within our application), it gets an answer of 2048. When
GL4Java queries this value with glGetIntegerv(GL_MAX_TEXTURE_SIZE, ...) f=
rom
within our same application on the same machine, it gets an answer of 102=
4.
What is up with that? Could the GLCapabilities be used to increase the ma=
x
supported texture size when the GLContext (state machine) is being create=
d?
Is GL4Java hardcoding this ceiling of 1024?
Pepijn Van Eeckhoudt wrote:
> OpenGL does not require square textures. If in practice it is necessary
> to use square textures, it's a driver/hardware limitation.
> The problem is probably that your hardware can't handle textures that
> are that big. A lot of consumer level cards can only handle textures up
> to 1024x1024. You can query this maximum value using glGetIntegerv(
> GL_MAX_TEXTURE_SIZE ).
>
> Pepijn Van Eeckhoudt
>
> On zondag, maa 23, 2003, at 11:38 Europe/Brussels, Florent Geffroy
> wrote:
>
> > I think the error occurs because your 2048*1024
> > texture is not a square. I think OpenGL can only
> > handle square textures !
> > You'll need to add black rectangles on both parts of
> > your image, in order to have a 2048*2048 one, and then
> > play with the texture coordinates of your polygon...
> >
> > Hope this helps...
> >
> > --- Mark Montana <pat...@lm...> a
> > =E9crit : > All,
> >>
> >> Follow up texture-mapping questions...
> >>
> >> First, I fixed my original problem that was causing
> >> textures in my application to
> >> be all white. I had a special initialization method
> >> to set all the texture mapping
> >> parameters. I forgot to surround those GL calls
> >> with:
> >> glc.getGLContext().gljMakeCurrent(); AND
> >> glc.getGLContext().gljFree();
> >> A rather newbie-sounding mistake, but it's been a
> >> while since I wrote rendering
> >> code in GL4Java. I hate that you have to perform
> >> those steps in the non-draw
> >> methods, but I realize it is a necessary evil
> >> (Magician had the same thing).
> >>
> >> On to my 2 new questions (somewhat more advanced):
> >>
> >> (Q1) I am loading fairly large textures of satellite
> >> imagery of the Earth. Two of
> >> them are 1024x1024 (powers of two) in RGB format.
> >> These two images load and render
> >> just fine. The third texture is 2048x1024 (also
> >> powers of two) in RGB format. This
> >> one gets a GL error during my call to
> >> glTexImage2d(), and the error is "invalid
> >> value." I determined that value that makes it
> >> unhappy is the image width (2048).
> >> If I lie to OpenGL and tell it that the width is
> >> only1024, the image loads and
> >> renders without an error (of course it looks screwed
> >> up, but I expect that since I
> >> lied about the true width). Any insight on why it
> >> cannot handle the 2048 width?
> >>
> >> I know that it is not related to the hardware,
> >> graphics card, graphics card
> >> driver, or machine configuration. I know this
> >> because our application is half
> >> GL4Java and half Java3D. The Java3D display loads
> >> the same exact textures and
> >> renders them with no problem. I also do not think it
> >> is related to the total
> >> amount of texture memory available; again because
> >> Java3D has no problem, and I can
> >> disable the Java3D display and disable to two
> >> 1024x1024 textures in the GL4Java
> >> display...and the GL4Java display still gets the
> >> same error on the 2048x1024
> >> texture.
> >>
> >> Could it have something to do with the OpenGL
> >> context or configuration that I am
> >> requesting???
> >> Could it be related to the fact that I am using
> >> display-lists???
> >>
> >> [[ I inserted my texture mapping method at the
> >> bottom of this message. ]]
> >>
> >> (Q2) My textures are rendered in a 2D display, no
> >> lighting, no normals, no
> >> materials. The polygons underneath the texture are
> >> white. The texture comes out
> >> way to dark with GL_DECAL or GL_REPLACE. When I try
> >> GL_MODULATE or GL_BLEND, it
> >> gets really washed-out. We had this same exact
> >> problem in the Java3D display with
> >> these textures. We fixed it there by using some
> >> texture-environment settings that
> >> are unique to Java3D. (In other words, the Java3D
> >> people used some of the more
> >> advanced features of blending functions and/or
> >> filters to create a new blending
> >> equation that works well.) I was considering a
> >> decompile of the Java3D code to see
> >> how they did it, but I thought I would as for some
> >> advice on this problem first.
> >>
> >>
> > ---------------------------------------------------------------------=
--
> > -----------------
> >>
> >> My texture map initialization code (sorry for the
> >> formatting if your mail tool
> >> doesn't support fixed-width fonts):
> >>
> >>
> >> private void initialize()
> >> {
> >> // Must make the GLComponent's OpenGL context
> >> current when outside a
> >> // GLEventListener callback method. Then don't
> >> forget to unlock() it.
> >>
> >> glc.getGLContext().gljMakeCurrent(); //
> >> Make current to call OpenGL
> >>
> >>
> >> GeoTiffImage image =3D
> >> geoTiffDef.getGeoTiffImage();
> >> Raster raster =3D image.getAsRaster();
> >> Object data =3D raster.getDataElements(
> >> 0, 0,
> >>
> >> raster.getWidth(),
> >>
> >> raster.getHeight(),
> >>
> >> null );
> >>
> >> if( ToolkitConstants.TK_DEBUG_PATCH2D =3D=3D true
> >> )
> >> {
> >> if( raster.getTransferType() =3D=3D
> >> DataBuffer.TYPE_BYTE )
> >> System.out.println( "Texture storage
> >> type =3D TYPE_BYTE");
> >> else
> >> {
> >> System.err.println(
> >> "TexturedPatch2d.initialize() - " +
> >> "What is the texture
> >> storage type?");
> >> System.err.flush();
> >> }
> >>
> >> System.out.println( "Num data elements =3D "
> >> +
> >>
> >> raster.getNumDataElements() );
> >> System.out.println( "Geotiff size =3D (" +
> >> raster.getWidth() + ", "
> >> +
> >> raster.getHeight() + ")" );
> >> }
> >>
> >>
> >> // Set unpacking to use 1 byte allignment
> >> gl_.glPixelStorei( GLEnum.GL_UNPACK_ALIGNMENT,
> >> 1 );
> >>
> >> // Ask for valid, unused texture name
> >> gl_.glGenTextures( 1, textureName );
> >>
> >> // Bind a texture object to the name
> >> gl_.glBindTexture( GLEnum.GL_TEXTURE_2D,
> >> textureName[0] );
> >>
> >> // Set clamping preferences
> >> gl_.glTexParameteri( GLEnum.GL_TEXTURE_2D,
> >> GLEnum.GL_TEXTURE_WRAP_S,
> >> GLEnum.GL_CLAMP );
> >> gl_.glTexParameteri( GLEnum.GL_TEXTURE_2D,
> >> GLEnum.GL_TEXTURE_WRAP_T,
> >> GLEnum.GL_CLAMP );
> >>
> >> // Set mag and min filter preferences
> >> gl_.glTexParameteri( GLEnum.GL_TEXTURE_2D,
> >> GLEnum.GL_TEXTURE_MAG_FILTER,
> >> GLEnum.GL_NEAREST );
> >> gl_.glTexParameteri( GLEnum.GL_TEXTURE_2D,
> >> GLEnum.GL_TEXTURE_MIN_FILTER,
> >> GLEnum.GL_NEAREST );
> >>
> >> gl_.glTexEnvi( GLEnum.GL_TEXTURE_ENV,
> >> GLEnum.GL_TEXTURE_ENV_MODE,
> >> GLEnum.GL_MODULATE ); //
> >> GL_REPLACE, GL_BLEND, GL_DECAL
> >>
> >> gl_.glHint(
> >> GLEnum.GL_PERSPECTIVE_CORRECTION_HINT,
> >> GLEnum.GL_FASTEST );
> >>
> >> // Define the 2D texture
> >> gl_.glTexImage2D( GLEnum.GL_TEXTURE_2D, 0,
> >> GLEnum.GL_RGB,
> >> raster.getWidth()/2,
> >> raster.getHeight(),
> >> 0, GLEnum.GL_RGB,
> >> GLEnum.GL_UNSIGNED_BYTE,
> >> (byte[]) data );
> >>
> >>
> >> if( ToolkitConstants.TK_DEBUG_PATCH2D =3D=3D true
> >> )
> >> {
> >> int errorCode =3D GLEnum.GL_NO_ERROR;
> >>
> >> if( (errorCode =3D gl_.glGetError()) !=3D
> >> GLEnum.GL_NO_ERROR )
> >> {
> >> System.err.println(
> >> "TexturedPatch2d.initialize() - " +
> >> "GL Error: " +
> >> glu_.gluErrorString(errorCode) );
> >> System.err.flush();
> >> }
> >> }
> >>
> >>
> > =3D=3D=3D message truncated =3D=3D=3D
> >
> > ___________________________________________________________
> > Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en fran=E7ais !
> > Yahoo! Mail : http://fr.mail.yahoo.com
> >
> >
> > -------------------------------------------------------
> > This SF.net email is sponsored by:Crypto Challenge is now open!
> > Get cracking and register here for some mind boggling fun and
> > the chance of winning an Apple iPod:
> > http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en
> > _______________________________________________
> > gl4java-usergroup mailing list
> > gl4...@li...
> > https://lists.sourceforge.net/lists/listinfo/gl4java-usergroup
> >
>
> ------------------------------------------------------------
> Name: PGP.sig
> PGP.sig Type: application/pgp-signature
> Encoding: 7bit
|
