[argyllcms] Re: Profiling flexo presses

Roger,

> If indeed I choose a grid resolution of 33 for Lab to CMYK, which is the
> maximum with most profiling software, don't I get 100/(33-1) = 3.125% steps,
> at best?

Roger, that would be if all the PCS space were useful for device
gamut, which is not the case.

> You wonder why couldn't there be some scaling factor that allows making
> intelligent use of the encoding space.

the only way to do that is by means of input and output curves (which
a profile can have in addition to the tables). This is somewhat crude,
but still can result in an improvement of PCS space usage.

> Most CMYK profiles I seen only have 17x17x17 CLUT in the Device -> PCS
> direction. I was once told that increasing the grid resolution would create
> monster profiles?
Large, indeed ; but how do you define a monster? :)
 
> ECI2002 target has patches with 2% values. I'm not following?

Oh, I was not aware of that. Well, for building device->PCS these 2%
patches will very likely not be taken into account, given that at 17
grid points your starting value would be 6.25%, and for that you would
interpolate between the 5% and the 10%. Unless there were some routine
to extract linearization (curve) and tables from one single ECI2002
dataset.
And for PCS->device those 2% patches would only matter in the gray axis, if so.

> Hmmh? Building flexo profiles with ProfileMakerPro actually yields accurate
> Device -> PCS mappping, matching the original target measurements. Could it
> be that the "influence of readings from 2% patches" you refer to is not
> quite as you think?

I guess then you're dealing with a linearized flexo press. If your
grid points are situated at 0, 6.25%, 12.5% .. how could you
accurately represent a change of slope happening below 6%, or even
worse, dot dissapearance below 4%?

> > Now, if the profile were to be used for proofing (device->PCS) or with
> > a CMM that does on the fly table reversal (instead of using
> > pcs->device it just reverses device->PCS) such as argyll, then it's
> > better.
> 
> Could you please explain that point further?

Building a PCS->device table from target readings (device), basically
loses the accuracy of the original data and replaces it by
interpolated (thus less accurate) data situated at regular intervals
of PCS space (which the readings were not). As you don't know in
advance which points in PCS space you'll need to render, you just
select a "grid" and render it.
Afterwards when you want to render some PCS colors to device space,
you interpolate between those gridpoints that were already
interpolated from device space readings. Double loss of precision.
If instead of that, you keep the original readings (device->pcs table)
and when you want to find a device value for an original PCS point,
you do a reverse interpolation (I know what I want to say, but doubt
this is proper wording!) in the device space, you have much more
precision.
That's what the Imation CFM did (although I never saw evidence of
better quality!), and that's one of the options of Argyll CMM.

> Please excuse my ignorance, but you mean I could create such profiles with
> argyll? 41 grid points both ways?

I remember a post from Graeme in Colorsync-users boasting about a
gigantic profile with a large number of grid points, I don't remember
exactly how many... 65?

> > Furthermore, you just can't print 2% on a flexo press that has not
> > been linearized, at least in the highlights by means of a bump curve.
> 
> What do you mean roberto? All the flexo printers I know of routinely print
> 2% dots. Without the need of bump curves.

Roger, why do you think Barco charged (time ago) so much for a
software called Intellicurve, which was only a 16-bit multi-pass
linearization tool? The big problem in flexo is lack of linearity in
the highlights, big changes of slope and dot loss at some point which
I believe is 3% or 5%, don't recall exactly (of course depending on
linescreen, but usually specified at 133lpi).
If your friends are routinely printing 2% dots in flexo at 133lpi or
higher, there is a bump curve. The digital screened file does not have
2% there, for sure. Maybe the curve is implicit in the RIP (there
are/were dedicated rips for flexo such as the AlanRip), or somewhere
else you haven't noticed, or maybe it has been applied to the file in
Photoshop or whatever. There has to be a curve.

-- Roberto Michelena
   Infinitek
   Lima, Peru

Other related posts: