Caricature challenge
Wilhelm Sanke
sanke at hrz.uni-kassel.de
Thu Dec 20 14:15:16 EST 2007
Most things have already been said, and I will try to keep my comment short.
Bush was not a specific target for me, but more of a random choice to
exemplify the new "caricature" features of my Toolkit. It could have
chosen Hillary or Barack, or Putin for that matter.
I remember, someone - about half a year ago? - posted the address of a
flash clip on this list which showed "Bush falling through the clouds -
forever". This drew a lot of responses, as far a s I remember not one
of them negative, and no one complained that this list was not the
proper forum for such occasional stuff. So I felt on the safe side with
my choice.
As has already been pointed out: No specific political or national
intentions on my side and least of all to antagonize Chipp as a citizen
of Texas (By the way, I very much liked the description of Texan
peculiarities in John Gunther's book "Inside USA").
And concerning Chipp's last post:
Chipp Walters chipp at chipp.com wrote on Wed Dec 19, 2007
>
>
> Terry,
> (snip)
> I think you'll find it difficult to distort imageData the way you want
> to in
> Rev without an external.
>
> Wilhelm's toolkit, while very complete and in it's right a wonderful
> set of
> scripts, isn't setup to do this sort of thing. In fact, part of the
> basis of
> his original toolkit were derived from some convolution matrix code I had
> generated years ago.
>
> best,
> Chipp
I think my attempt of a caricature indeed shows what can be done by
manipulation of imagedata with Revolution - even without externals,
although in different ways as demanded by Terry and proposed by Ken. I
had described this in more detail in my post, and these new routines
(which will be part of the next version of my ImagedataToolkit) had
not been specifically designed to produce caricatures, but simply to
copy and paste parts of an image elsewhere, to enlarge, shrink, and
flip them, and to integrate the selected ovals, rects, or polygons
with transparent fringes into the same or an another picture.
Manipulating the Bush photo with the six alterations was a matter of
about five minutes.
Chipp Walter - with his old stack he mentions - and Ken Ray - with his
imagedata webpages - were the ones that opened the world of imagedata
for me.
I give Chipp credit for his external in the introduction to the
Toolkit and have given him credit elsewhere. Chipp's external later was
replaced by Derek Bump's one, which did not require to script around
the Endian issue.
I had an exchange with Chipp about the changes I applied to his
non-external script to use matrix filters, i.e. without the external.
Here the intention of Chipp had been to demonstrate that Revolution was
capable to use matrix filters even without an external. Speed was no
consideration for Chipp at that time and speed was not needed in that
sample stack of Chipp as the used image was very small.
As I worked with larger images in my Toolkit I tried to rewrite the
non-external script and succeeded in reducing the execution time for
640x480 images from 90 to 7 seconds on a 2 GHz computer, which you can
see using the present Toolkit.
Mark Waddingham, in a post last week concerning bug # 5113, still found
additional potential in the script to reduce execution time even more,
which is now down to 4.7 seconds on my Windows computer and 3.4 seconds
on my MacBook Pro with Leopard.
Neverless I asked Mark, if also externals for 5x5 and 7x7 matrix filters
could be added for Revolution - and both for Mac and Windows.
But, of course, matrix filters are only a limited part of the features
of my present and upcoming Imagedata Toolkits.
Best regards,
Wilhelm Sanke
<http://www.sanke.org/MetaMedia>
More information about the use-livecode
mailing list