Re: tilt-shift... but faster?

Oh, found the biggest culprit: was using "-set option:compose:args 15".
 Lowering that to 8 cut the time down significantly.  I assume that's the
blur radius?  Thanks!

-Aidan

On Tue, May 17, 2011 at 12:39 AM, Anthony Thyssen <[email protected]
> wrote:

> On Mon, 16 May 2011 23:33:29 -0400
> [email protected] wrote:
> | Hi all-
> | So first I wanted to announce a Ruby gem I created called magickly, which
> | essentially offers Imagemagick as a web service.  Please take a look -
> would
> | love any feedback:
> |
> | https://github.com/afeld/magickly
> |
>
> I suggest you make the announcement in the IM Forum Announce area
>
> A link to some basic documentation (not the project page) would be helpful!
>
> | convert source.png \\( +clone -sparse-color Barycentric '0,0 black
> | 0,%[fx:h-1] white' -function polynomial 2,-2,0.5 \\) \
> |
> | -compose Blur -set option:compose:args 15 -composite output.png
> |
> Hmmm that sparse-color line is a little wrong but not to any visible
> degree in the results.  Updating usage example
>
> For general speedups see...
>  http://www.imagemagick.org/Usage/api/#speed
>
> However the blur is very slow.. and for a short period, a patch to fix
> blurs of less that 1 sigma caused it to be 3 to 5 times slower still.
>
> That patch was removed until I can re-think the problem.
> If you IM is between versions  6.6.7-7 and 6.6.8-4  and Upgrade your IM.
>
> Perhaps a run-time switch to remove the use of a LUT for EWA filtering,
> which was the cause of the the slow down in that patch.
>
> Sorry thinking out loud!
>
>  Anthony Thyssen ( System Programmer )    <[email protected]>
>  --------------------------------------------------------------------------
>    Philosophy (n):  unintelligible answers to insoluble problems.
>  --------------------------------------------------------------------------
>   Anthony's Castle     http://www.ict.griffith.edu.au/anthony/
> _______________________________________________
> Magick-users mailing list
> [email protected]
> http://studio.imagemagick.org/mailman/listinfo/magick-users
>