Antialiasing and IR

New Message Reply Date view Thread view Subject view Author view

Rob Jenkins (robj++at++quid)
Fri, 8 Nov 1996 13:46:54 -0800

On Nov 8, 2:33pm, login IRIX wrote:
> Subject:
> Hello ,
>
> I have a little problem with antialiasing.
> Well, I use a image mapping to simulate shadow on a terrain. To do
> that,
> I use two faceted terrains ( Yes, I know, the number of polygons is twice: if
> somebody has another idea?) on each other. On the first one, I map a repeated
> texture, on the second I map a image with an alpha mask. This allows me to
> pertub my repeated texture.
> This method works well when I don't use antialiasing. In this case,
> antialiasing seems to modify the number of bits of alpha channel (I 'm sure
> that there are 8 bits at the beginning ).
>
> If somebody could help me...
>
> Thanks a lot. I am not a genius.
>
> PS: where could I find some sample frames of what you are doing with
Performer,
> guys? If you want, I will show you some snapshots of my application.
> =======================================================================
> List Archives, FAQ, FTP: http://www.sgi.com/Technology/Performer/
> Submissions: info-performer++at++sgi.com
> Admin. requests: info-performer-request++at++sgi.com
>-- End of excerpt from login IRIX

There's no guarantee you'll get all the framebuffer resources you ask for. I'll
guess you're using Reality Engine, if you ask for too many multisamples, one of
the first thing the framebuffer config does to satisfy your request is switch
from RGBA 12 bits per component pixels to RGB 8 bit per comp. This lack of
alpha isn't generally a problem as blended transparency can be done without
destination alpha planes but it sounds like you're relying on having
destination alpha. *Always* find out what framebuffer config you actually got (
man pfWindow talks about various ways ). You may be able to do the alpha
operations you're doing still without destination alpha by adjusting you're
blendfunction but most likely you need to ask for less samples ( or get more
RM boards ). If this is RE2 then here's some guidelines:

One RM is enough for:

640x512 resolution 16 samples 8 bits/component 24-bit Z
640x512 resolution 8 samples 12 bits/component 32-bit Z
960x680 resolution 8/4 samples 12 bits/component 32-bit Z
1280x1024 resolution 0 samples 12 bits/component 32-bit Z

Two RMs are enough for:

1280x1024 resolution 8 samples 8 bits/component 24-bit Z
1280x1024 resolution 4 samples 12 bits/component 32-bit Z
1600x1200 resolution 0 samples 12 bits/component 32-bit Z

Four RMs are enough for:

1280x1024 resolution 16 samples 8 bits/component 24-bit Z
1280x1024 resolution 8 samples 12 bits/component 32-bit Z
1600x1200 resolution 8 samples 8 bits/component 24-bit Z
1600x1200 resolution 4 samples 12 bits/component 32-bit Z

Cheers
Rob

-- 
________________________________________________________________
Rob Jenkins robj++at++csd.sgi.com
Silicon Graphics, Mtn View, California, USA
=======================================================================
List Archives, FAQ, FTP:  http://www.sgi.com/Technology/Performer/
            Submissions:  info-performer++at++sgi.com
        Admin. requests:  info-performer-request++at++sgi.com

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:53:55 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.