[argyllcms] Re: targen options

Gerhard Fuernkranz wrote:
> in later version of Argyll, there are new targen options
> -r, -R, -q, -i, -I, -a for full-spead patch generation.
> Are these additional methods useful for particular cases,
> or are they rather for evaluation purposes only?

They were mainly for testing and evaluation (creating
data for my CIC2004 paper).

> Btw, what about the -A option?
> Is it generally recommended to use -A 1.0 for genreating
> targets? Could you please tell me a bit more detailed
> than the manual, what -A actually does?

It's still rather experimental. Without it the default algorithm
minimizes the maximum distance from any point in the device
space to the nearest test point. targen uses a first pass
algorithm to get within 2:1 of the optimum, and then uses an
iterative algorithm to get closer to optimum.

With adaptation, an existing response of the device is used
to try and minimize the expected error in the profile. The
-A option allows choosing a hybrid somewhere between these
two approaches. Currently if any degree of adaptation, the
iterative improvement is disabled, since I haven't figured
out how to do this yet.

Technically, the fully Adapted mode chooses test points incrementally
such that they are placed at the location with the greatest current
error between the expected response of the device and the linearly
interpolated (delaunay triangulation) Lab value computed from the
current set of test points.

Graeme Gill.

Other related posts: