Re: Q1 ?

Glenn Linderman on  wrote...
| I'm surprised that Q1 would require a rewrite of so many
algorithms.
| 
The problem is that while Q8 requires one byte per value,
and high Q
levels also reference values at byte level,  Q1 would
require bit level
processing, whcih is a very different way of handling data.

Otherwise you may as well use Q8, restricting yourself to
just black and
white!

A rewrite would be needed to support Q1 effeciently for
memory and
processing.


  Anthony Thyssen ( System Programmer )    <A.Thyssengriffith.edu.au>
 -----------------------------------------------------------
------------------
  "Ye castle looks awfully like a home page to thy
eyes! Castle bricks doth
   not fit down ye old ethernet cable, though stupidity,
apparently, does."
                        --- Dave Connors <dconnorsgil.ipswichcity.qld.gov.au>
 -----------------------------------------------------------
------------------
     Anthony's Home is his Castle     http://www.cit.gu.
edu.au/~anthony/
_______________________________________________
Magick-users mailing list
Magick-usersimagemagick.org
http://studio.imagemagick.org/mailman/listinfo/magick
-users

On approximately 3/29/2007 10:44 PM, came the following
characters from 
the keyboard of Anthony Thyssen:
> Glenn Linderman on  wrote...
> | I'm surprised that Q1 would require a rewrite of so
many algorithms.
> | 
> The problem is that while Q8 requires one byte per
value, and high Q
> levels also reference values at byte level,  Q1 would
require bit level
> processing, whcih is a very different way of handling
data.
>
> Otherwise you may as well use Q8, restricting yourself
to just black and
> white!
>
> A rewrite would be needed to support Q1 effeciently for
memory and
> processing.
>   
Ah yes, bit level vs byte level... that would be a
difference for sure, 
but couldn't that be abstracted to the pixel cache accesses?
 Writing a 
bit-level pixel cache would _certainly_ be more work that
flipping a Q8 
vs Q1 switch, I can see... but it would conceivably be much
less work 
than re-writing almost every algorithm.  And it is probably
a bigger 
time investment than I could make at present.

Now I can see that changing _only_ the pixel cache wouldn't
be as 
efficient as rewriting the algorithms to be aware of the
bit-level 
access, and the ability to handle/process multiple pixels at
a time in a 
single byte/word/dword, but it seems like it could save a
lot of pixel 
cache memory, which in turn would save lots of disk I/O on
RAM limited 
machines.  It would probably burn the same amount of CPU, or
perhaps 
slightly more or less, I would guess (hard to say for sure
without 
trying it).

-- 
Glenn -- http://nevcal.com/
===========================
A protocol is complete when there is nothing left to
remove.
-- Stuart Cheshire, Apple Computer, regarding Zero
Configuration Networking

_______________________________________________
Magick-users mailing list
Magick-usersimagemagick.org
http://studio.imagemagick.org/mailman/listinfo/magick
-users