[argyllcms] Re: Computing A2B tables
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Sat, 23 Jul 2005 10:42:14 +1000
Roger Breton wrote:
What is the prevailing wisdom in computing the A2B table? Do you prefer to
fit every measurements accurately or do you prefer to sacrifice accuracy for
smoothness?
It's really up to the profile vendor. Argyll uses a thin plate regular
spline
approach, and does sacrifice some level of accuracy for smoothness. This
can be controlled (programmatically), but generally I've tried to
maximize accuracy, rather than maximize smoothness. Even if
the "smoothness" factor is turned down, there is a limit to how
much it can smooth out "bumpy" sample data or device response.
Graeme Gill.
- References:
- [argyllcms] Re: Computing A2B tables
- From: Roger Breton
Other related posts:
- » [argyllcms] Re: Computing A2B tables
- » [argyllcms] Re: Computing A2B tables
What is the prevailing wisdom in computing the A2B table? Do you prefer to fit every measurements accurately or do you prefer to sacrifice accuracy for smoothness?
It's really up to the profile vendor. Argyll uses a thin plate regular spline approach, and does sacrifice some level of accuracy for smoothness. This can be controlled (programmatically), but generally I've tried to maximize accuracy, rather than maximize smoothness. Even if the "smoothness" factor is turned down, there is a limit to how much it can smooth out "bumpy" sample data or device response.
- [argyllcms] Re: Computing A2B tables
- From: Roger Breton