[argyllcms] Re: targen samples filter out bug (v1.3.5)

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Thu, 15 Mar 2012 08:37:29 +1100

Nikolay Pokhilchenko wrote:
> I've noticed that "-F" parameter didn't work together with "-q". For example:
> 
> C:\Argyll_V1.3.5\bin>targen -v -d2 -p1.4 -w -f2880 -Q -        Test
> iRGB test chart
> Full spread patches = 2880
> Filtering out samples outside sphere at 0.000000 0.000000 0.000000 radius 
> 25.000000
> 
> Added 2880/2880
> Total number of patches = 64
> Execution time = 1.593000 seconds
> 
> C:\Argyll_V1.3.5\bin>targen -v -d2 -p1.4 -w -f2880 -q -F 0,0,0,25 Test
> iRGB test chart
> Full spread patches = 2880
> Filtering out samples outside sphere at 0.000000 0.000000 0.000000 radius 
> 25.000000
> 
> Added 2880/2880
> Total number of patches = 2880
> Execution time = 0.429000 seconds

Hi, why don't you think it worked ? In the second example, are there any
test patches outside an L*a*b* 25 range of 0 ?

[ Note that some classes of the full spread patches don't support generating
  the total number of patches while being filtered - instead the points are
  generated then filtered. Some of the algorithms can't reasonably support
  it (OFPS), in some it would be tricky and or slow (-t, -i, -I), and for some
  I haven't implemented it simply because the points are generated outside 
targen.c
  (-Q, -R) where the filter code is. The filter is a quick and dirty - ideally 
it
  would be more general, allowing an arbitrary region in device and/or 
perceptual
  space. I suspect implementing such a thing for all the patch types would be 
some
  months of work for not a whole lot of benefit.]

Graeme Gill.

Other related posts: