Angus Dorbie (dorbie++at++bitch.reading.sgi.com)
Tue, 25 Feb 1997 19:40:39 +0000
> Yeah, I think the limitations won't be a big deal in this case. Do you
> know if the IR texture color index table really is 10 bit, like it is
> leading me to believe, though the man pages say 8? It would be nice to
> support 1k materials. Can I expect normal texel rates in this index mode
> or should I expect a performance hit?
I think it's free. You should be able to get better than 8 bit table
although you'd need better than an 8 bit image to make use of this.
Be carefull of what question you ask though.
What you are actually talking about is the colour table width so for
say a 12 bit table you'd want a 12 bit intensity image an internal format
of GL_INTENSITY12_EXT for the table, a format of GL_RGBA and a type of
GL_UNSIGNED_INT_8_8_8_8_EXT.
You then sent 2^12 map entries of 32 bit RGBA data for the table, but this
would be too large so you could reduce the width parameter to reduce the
table size and give say 10 bits of look up at this quality or you could
have table entries of format GL_UNSIGNED_BYTE_3_3_2_EXT and retain the full
12 bit information at reduced back end image quality.
If you mean that you want 10 bit map entries then this is a different
situation. You could have 2^8 entries, an internal format of
GL_INTENSITY8_EXT and a type of GL_UNSIGNED_INT_10_10_10_2_EXT to
give more precise controll of the table colours but this is of less use
for an cmap image than a more precise internal format.
There are a plethora of in-betweens which you can play around with, just
realise what precision means when using these tables and watch you don't
accidentally loose it somewhere in the formats you choose.
Cheers,
Angus.
=======================================================================
List Archives, FAQ, FTP: http://www.sgi.com/Technology/Performer/
Submissions: info-performer++at++sgi.com
Admin. requests: info-performer-request++at++sgi.com
This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:54:44 PDT