Hehe, Oh I am sorry, but this gives me the giggles. I have built a generic
shape-like object that I can perform lots of high level optimizations
against. It stores its internal numbers in indexed format for a lot of
reasons, one is memory, another is animation, another is cast and compact
disk storage and retrievals. I can ask this object for whole slew of
different GeometryArray and IndexedGeometryArray, and it builds the
appropriate geometry array and returns it. Course I have to unwind the my
own arrays to return an unindexed shape. I was was really suprised when I
saw no frame difference whatsoever when I retreived an Index versus an
Unindexed GeometryArray, and when memory usage went up for the indexed
version I was pretty convinced it was "unindexing" it.
It sounds like we would be foolish to ever use an IndexedGeometryArray of
any kind since it just unwinds it at a cost of both cpu cycles and memory.
When you use reference geometry does the referenced data make it all the way
to the retained structures? Or does the shape node just use the referenced
geometry, but still translate and copy it to the retained versions?
Dave Yazel
----- Original Message -----
From: Doug Twilleager <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Monday, December 04, 2000 1:21 PM
Subject: Re: [JAVA3D] Indexed Triangle Arrays and Multitexture
Yes, Java 3D currently unwinds or unIndexifies indexed geometry before it
renders it. It only does it when the geometry data actually changes. It
does
this because the index format does not map well to low level API formats.
We
could look at fixing this in a future release. As has been discussed here,
the
indexed format in Java 3D needs updating.
For your problem, I am guessing that using geometry by reference will
probably
work best for you. It all depends on how often you plan on updating the
geometry.
If you plan on doing the update every frame, geometry by reference will be
the
way to go. Unfortunately, you will probably need to try both ways to find
which
one works better for your application.
Doug Twilleager
Sun Microsystems
>MIME-Version: 1.0
>Content-Transfer-Encoding: 7bit
>X-Priority: 3
>X-MSMail-Priority: Normal
>X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2615.200
>Subject: Re: [JAVA3D] Indexed Triangle Arrays and Multitexture
>To: [EMAIL PROTECTED]
>
>Ok I found the bug parade and located the bug you referenced. The note
from
>the developer is:
>
>When the indexed triangle strip array geometry is unIndexified, the
>texCoordSetMap is not passed to the createGeometryArrayData().
>
>What I am confused about is the fact that I am using an
>IndexedTriangleArray, not a strip array. The other thing I don't
understand
>is what unIndexified means... you don't suppose they are unwinding the
>indexed geometry and passing a non-indexed version? I was actually
suprised
>to see my memory usage increase when I started using indexed geometry,
>although by my calculations I has reduced the information by more than 50
>percent.
>
>Dave Yazel
>
>----- Original Message -----
>From: David <[EMAIL PROTECTED]>
>To: <[EMAIL PROTECTED]>
>Sent: Monday, December 04, 2000 10:16 AM
>Subject: Re: [JAVA3D] Indexed Triangle Arrays and Multitexture
>
>
>Hmm I am not seeing this bug on the list. Is there more than one list? I
>am looking at:
>
>http://developer.java.sun.com/developer/earlyAccess/java3D/java3d-bugs.html
>
>And I have verified this also happens for quads :( Did you find a work
>around? It looks like it IS multi-texturing, just that its using the tex
>coords from unit state 1.
>
>Dave Yazel
>
>----- Original Message -----
>From: Pedro Estrada <[EMAIL PROTECTED]>
>To: <[EMAIL PROTECTED]>
>Sent: Monday, December 04, 2000 9:41 AM
>Subject: Re: [JAVA3D] Indexed Triangle Arrays and Multitexture
>
>
>David,
>
>I encountered the same bug when doing terrain mesh multitexturing using
>indexed triangle arrays. It is Bug 4390217. It has been identified and
>corrected for 1.2.1 beta2. Check out the bug for more info.
>
>-Pedro
>
>----- Original Message -----
>From: "David" <[EMAIL PROTECTED]>
>To: <[EMAIL PROTECTED]>
>Sent: Monday, December 04, 2000 9:09 AM
>Subject: [JAVA3D] Indexed Triangle Arrays and Multitexture
>
>
>> I have re-implemented my terrain mesh as indexed triangle arrays and
>> everything is rendering fine with the exception of multi-texturing. I am
>> not ruling out a bug of mine, but I have checked all the obvious places.
>> The texture coords for unit state 2 are correct, the indexes for unit
>state
>> 2 are correct, the rendering attributes are correct. But the mesh is
>> rendering without multitexture. Before I spend too much time on this I
>> wanted to ask if there are any known bugs in 1.2.1 beta 1 OpenGL that is
>> know regarding multi-texturing with indexed triangle arrays.
>>
>> Dave Yazel
>>
>>
>===========================================================================
>> To unsubscribe, send email to [EMAIL PROTECTED] and include in the
>body
>> of the message "signoff JAVA3D-INTEREST". For general help, send email
to
>> [EMAIL PROTECTED] and include in the body of the message "help".
>>
>
>===========================================================================
>To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
>of the message "signoff JAVA3D-INTEREST". For general help, send email to
>[EMAIL PROTECTED] and include in the body of the message "help".
>
>===========================================================================
>To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
>of the message "signoff JAVA3D-INTEREST". For general help, send email to
>[EMAIL PROTECTED] and include in the body of the message "help".
>
>===========================================================================
>To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
>of the message "signoff JAVA3D-INTEREST". For general help, send email to
>[EMAIL PROTECTED] and include in the body of the message "help".
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".