🚀 go-pugleaf

In article <1992Oct05.203641.13429@bmerh85.bnr.ca> carrasco@bcars6e1.bnr.ca (Sylvano Carrasco) writes:
>cjwein@watcgl.uwaterloo.ca (Chris Wein) writes:
>>It also corrects an error from the 2nd edition which is the worst I have
>>ever seen.  I really couldn't believe it when I saw it the first time.
>
>Gee, Chris how nasty of you to put such a leading statement like this and not
>follow up. Not like you at all.
>I used that text for a course and don't remember anything that horrendous.
>Then again maybe I just wasn't paying attention.

I knew about 2 seconds after I posted that article about G and W that
I should have mentioned the error so here goes...

In the section about the discrete cosine transform they define
the 1D DCT as ...cos((2x+1)u*pi/N)...  When they combine two 1D
DCT's to form the 2D transform they pull out the 1/N as the argument
to the cos to form ... (1/N^2)cos((2x+1)u*pi)*cos((2y+1)v*pi) which
is essentially saying that cos(x/n)*cos(y/n) = (1/n^2)cos(x)cos(y).

It is corrected in the Gonzalez and Woods version.
-- 
==============================================================================
 Chris Wein                           | cjwein@watcgl.waterloo.edu 
 Computer Graphics Lab, CS Dept.      | cjwein@watcgl.uwaterloo.ca
 University of Waterloo               | (519) 888-4548