Skip to main content

What is in Java 3D 1.3.2 build8 ?

3 replies [Last post]
jada
Offline
Joined: 2004-03-17

We are happy to share what we have achieved in our last pre-beta build. We would like to thank everyone on this forum for your contribution in making this possible.
Please refer to Kevin's 1.3.2-build8 (last pre-beta build)
post for link to build8.

Java 1.3.2 enhancemets
=======================
1) Improved the speed and quality of scene antialiasing (from 2 samples to possibly 8 samples) if pbuffer is supported.

2) Eliminated memory bloat for IndexedGeometryArrays with BY_REFERENCE and USE_COORD_INDEX_ONLY flags set.

Java 1.3.2 new properties
==========================
j3d.viewFrustumCulling - If this flag is set to false, the renderer view frustum culling is turned off. The default value is true.

Note: Java 3D uses a 2 pass view culling. The first pass is a loose view culling of the spatial tree, and the second pass is a tight view frustum culling in the renderer before sending the geometry down to the low level graphics API. The j3d.viewFrustumCulling property is to control the renderer view frustum culling, and it will not affect the first pass view culling.

j3d.transparentOffScreen - If this flag is set to true the background of the off screen canvas is set to transparent. The default value is false.

j3d.usePbuffer - If this flag is set to false pbuffer will not be use for off screen rendering. The default value is true.

Bug fixed
==========

Bugs fixed in build8 :

Issue Summary
------------------------------------------------------------------------------------------
4 Speed of OffScreen Canvas3D
11 Seemingly inconsistent delays in updating attributes
18 Geometry missing in OnScreenCanvas when OffScreenCanvas is resized
20 Off-screen rendering doesn't work on Linux
23 Enhancement to Reduce Memory Bloat in IndexedGeometryArray
26 Texture3D not displaying when size over 512X512
28 Combine Mode & Missing Texture in OffScreen snapshot on Win/XP
33 Texture updates fail w/ Multitextured BY_REF Geoms
36 Vector3d hashCode() method distinguishes between -0.0 and 0.0
38 ConfiguredUniverse throws an exception when using Java Web Start
66 NullPointerException in renderOffScreenBuffer
71 OGL: OffScreen canvas ignore GraphicsConfigTemplate under windows
72 Exception in setOffScreenBuffer when Canvas3D not in View
73 Quat4f#interpolate: incorrect when q1*q2<0
74 Texture/color math difference (D3D vs OGL)
76 OffScreen rendering should use Pbuffer on Windows / OpenGL
77 Scene antialiasing is disabled on Windows / OpenGL
80 libj3dcore-ogl.so fails to load on Linux systems without certain GL extensions

Bugs fixed before build8 :

Issue Summary
------------------------------------------------------------------------------------------
3 lg3d: gets unsatisified link error on various ATI cards
12 Frame counter isn't always incremented correctly
13 J3DTimer fails on Windows XP with P4 3GHz
14 Random NullPException in uninitialized SoundSheduler
16 Issue with Runtime instantiation of Canvas3D
19 Resizing off-screen Canvas3D hangs on Solaris
21 Exception in WakeupOnElapsedFrames behavior stops renderer from running when it should

thanks,
Java 3D Team.

Reply viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
nvaidya
Offline
Joined: 2004-08-03

> Java 1.3.2 enhancemets
> =======================
> ..
> 2) Eliminated memory bloat for IndexedGeometryArrays
> with BY_REFERENCE and USE_COORD_INDEX_ONLY flags
> set.
>

Chien,

That enhancement will reduce memory bloat whether or not the BY_REFERENCE flag is set. Please check. And if that's indeed the case, it should also please those who may not use that flag.

Thanks

--Vaidya

jada
Offline
Joined: 2004-03-17

Thanks! Look like you are right, it is my oversight. I must have read too much into issue 23, when compiling this list.

- Chien.

nvaidya
Offline
Joined: 2004-08-03

> Thanks! Look like you are right, it is my oversight.
> I must have read too much into issue 23, when
> compiling this list.
>
> - Chien.

:). Yes, you are right - my testcase and its description were a red herring with respect to the BY_REFERENCE flag. It was while doing the fix that I fully understood the irrelevance of that flag to the whole issue.

Thanks

--Vaidya