Thanks for the replies.

The two images are:

http://togaware.redirectme.net/access/p1.png
and
http://togaware.redirectme.net/access/p2.png

I'll give the suggestions a try.

Regards,
Graham

On Thu, Jun 4, 2009 at 12:25 PM, Anthony Thyssen
<[email protected]> wrote:
> On Thu, 4 Jun 2009 06:48:50 +1000
> Graham Williams <[email protected]> wrote:
>
> | I've been trying to find a solution to generating an animation, a
> | little like "-morph", that will "move" points in  a simple plot from
> | one location in an image to another location in a second image.
> | Essentially the two images have objects that are the same unique
> | colour in each image, and I want a series of images that move each
> | object to its new location, so that I end up with an animated gif that
> | looks like it is moving the objects "smoothly."
> |
> | I attach two images (generated using R) that illustrate the starting
> | and end points.
> |
> | Any ideas? The morph option is quite nice when the objects are not too
> | far from each other, but not quite what I am after:
> |
> | $ convert -delay 50 -morph 10 p?.png movie.gif
> | $ animate movie.gif
> |
> | Thanks for any help.
> |
> -morph is a very old 'quick hack' operation.  It only does a color morphing
> from one image to another.
>
> For true morphing you need to distort each of the two images to
> the corresponding position for EVERY STEP between the two images.
>
> That is you not only need to distort the first image to the second (and
> all the intervening steps between), but also distort the second image
> to the first image.
>
> When you have the two sequences, you then color blend the pairs of
> images (each distorted to the same points) so that you not only have a
> the images being spatially distorted, but also color blending from on
> to the other.
>
> At this time the IM -distort operator only has one method that can
> distort ANY NUMBER OF POINTS.  And that is 'shepherds' method.  I have
> yet to implement gridding and triangulation mesh distortion methods,
> though I do want to add them.
>
> WARNING; Shepards will not string rotating distortions.
>
> Fred's script pre-dates (and helped develop) that distortion, as is
> basically a shepherds distortion, with only ONE moving control point,
> and four fixed control points in the corners (which simplifies the
> maths enormously).
>
> Also he makes the morph processing even simpler by generating the
> distortions as a displacement map (Shepherds method is a displacement
> distortion).  (I have notes on this but again the example is not online
> yet)
>
> This means he only calculates the distortion ONCE for each direction,
> for each X and Y mappings (4 maps total), and varies the displacement
> vector to do linear spatial distortion for each of the intervening
> steps.  This is an enormous speed improvement, but only works for
> images that distort in straight lines (linearly) between the two images.
>
>
>  Anthony Thyssen ( System Programmer )    <[email protected]>
>  -----------------------------------------------------------------------------
>  "I would have been long since dead and gone, if I hadn't died."
>                        - Jolie the ghost -- "And Eternity" - Piers Anthony
>  -----------------------------------------------------------------------------
>     Anthony's Home is his Castle     http://www.cit.gu.edu.au/~anthony/
>



-- 
Regards,
Graham

_______________________________________________
Magick-users mailing list
[email protected]
http://studio.imagemagick.org/mailman/listinfo/magick-users

Reply via email to