mesa3d-dev

[Mesa3d-dev] Bug in z-depth

[Klaus N. <kl...@ma...>]

Hi,

There seems to be a bug in the depth-buffer code. That looks very strong like an rounding-error or something. On
http://www.math.leidenuniv.nl/~klaus/bug.jpg
you can see a screenshot I made. With Mesa-3.2 it looks o.k.

The error was produced by using Crystal-space. I believe it works like this (but I'm not sure): The walls are drawn with activated z-Buffer (but without z-Buffer testing) and afterwards objects are drawn with z-Buffer-test activated.

Greetings

	Klaus

Re: [Mesa3d-dev] Bug in z-depth

[Brian P. <br...@va...>]

Klaus Niederkrueger wrote:
> 
> Hi,
> 
> There seems to be a bug in the depth-buffer code. That looks very strong like an rounding-error or something. On
> http://www.math.leidenuniv.nl/~klaus/bug.jpg
> you can see a screenshot I made. With Mesa-3.2 it looks o.k.
> 
> The error was produced by using Crystal-space. I believe it works like this (but I'm not sure): The walls are drawn with activated z-Buffer (but without z-Buffer testing) and afterwards objects are drawn with z-Buffer-test activated.

I'm assuming that you're using software rendering (Xlib driver).

When I've seen this sort of thing before it usually caused by using "bad"
near/far clip plane positions.  What are your arguments to glFrustum?
If the ratio of far:near is greater than 1000 you should adjust the
values.

Also, you could check the depth of your Z buffer with
glGetIntegerv(GL_DEPTH_BITS, &bits).  I'd expect it to be 16 bits, but
maybe it's not.

-Brian

Re: [Mesa3d-dev] Bug in z-depth

[Stephen J B. <sj...@li...>]

On Wed, 4 Oct 2000, Klaus Niederkrueger wrote:

> Hi,
> 
> There seems to be a bug in the depth-buffer code. That looks very strong like an rounding-error or something. On
> http://www.math.leidenuniv.nl/~klaus/bug.jpg
> you can see a screenshot I made. With Mesa-3.2 it looks o.k.
> 
> The error was produced by using Crystal-space. I believe it works like this (but I'm not sure): The walls are drawn with activated z-Buffer (but without z-Buffer testing) and afterwards objects are drawn with z-Buffer-test activated.
 
This looks like pretty typical Z buffer precision
issues.

Read this:

  http://web2.airmail.net/sjbaker1/love_your_z_buffer.html

...then tell us if you still think it's a Mesa bug.

----
Steve Baker                      (817)619-2657 (Vox/Vox-Mail)
L3Com/Link Simulation & Training (817)619-2466 (Fax)
Work: sj...@li...           http://www.link.com
Home: sjb...@ai...       http://web2.airmail.net/sjbaker1

Re: [Mesa3d-dev] Bug in z-depth

[Klaus N. <kl...@ma...>]

>  
> This looks like pretty typical Z buffer precision
> issues.
> 
> Read this:
> 
>   http://web2.airmail.net/sjbaker1/love_your_z_buffer.html
> 
> ...then tell us if you still think it's a Mesa bug.
> 

Well, I send in another email saying that the Depthbuffer has only 8 bits, but this email does not seem to appear on the list (stupid elm).

The thing I think is strange is that the same binaries of CS use 8 bits when they are running with Mesa-3.3 or 3.5 and 16 bits with Mesa 3.2, but maybe it is a bug of Crystal-space...

No idea.

	Klaus

Re: [Mesa3d-dev] Bug in z-depth

[Brian P. <br...@va...>]

Klaus Niederkrueger wrote:
> 
> >
> > This looks like pretty typical Z buffer precision
> > issues.
> >
> > Read this:
> >
> >   http://web2.airmail.net/sjbaker1/love_your_z_buffer.html
> >
> > ...then tell us if you still think it's a Mesa bug.
> >
> 
> Well, I send in another email saying that the Depthbuffer has only 8 bits, but this email does not seem to appear on the list (stupid elm).
> 
> The thing I think is strange is that the same binaries of CS use 8 bits when they are running with Mesa-3.3 or 3.5 and 16 bits with Mesa 3.2, but maybe it is a bug of Crystal-space...
> 
> No idea.

I forgot to mention in my previous message that the support for N-bit
depth buffers was new in Mesa 3.3.  In 3.2 and earlier you always got
a 16-bit depth buffer, unless you specially compiled Mesa otherwise.

-Brian