use-revolution digest, Vol 1 #1521 - 13 msgs

Cubist at aol.com Cubist at aol.com
Mon Jun 23 11:16:01 EDT 2003


In a message dated 6/23/2003 8:06:08 AM, 
use-revolution-request at lists.runrev.com writes:

>
>Message: 9
>From: "Mathewson" <richmond at mail.maclaunch.com>
>Subject: Rotating Images: the rot sets in...
>To: use-revolution at lists.runrev.com
>Date: Mon, 23 Jun 2003 08:21:13 -0400
>Reply-To: use-revolution at lists.runrev.com
>
>I'm playing around with rotating images and am wondering
>about the phenomenon of "fuzzy edges".
>
>Let us just suppose I have a GIF image which I want to
>rotate in increments of 45 degs:
>all very simple: described in Help files & so forth.
>
>What is not explained is that the edges of the image (if it
>is plain colour) or the whole image (if it is a picture)
>will deteriorate markedly as it rotates.  This effectively
>means that the
>'rotate' term is not much use.  
>
>If one has a family of images (in the case above I would
>need 8) that each represent the image rotated at certain
>increments one can keep replacing images (like a slide
>show) to give an impression of rotation.  One could use an
>animated GIF.....or a Quicktime Movie....or....blah, blah,
>blah.   Makes your stack much larger!
>
>The problem with this is if one is using the rotating image
>as the template for the windowShape......
>
>The 'rotate' term is great, in theory; but how can one keep
>the image quality intact?
  Any time you rotate an image, you're going to get *some* "fuzziness". If 
you rotate a rotated image, the "fuzziness" accumulates. Therefore, never rotate 
an already-rotated image. Rather than do "rotate AlreadyRotatedImageX 1 
degree", keep a running total of whatever your angle *is*, and instead do "rotate 
OriginalZeroDegreeImage N degrees". That way, you've got exactly *one* 
rotation's-worth of "fuzziness" at any given time.



More information about the use-livecode mailing list