Re: Multi-color ICC profiles
Re: Multi-color ICC profiles
- Subject: Re: Multi-color ICC profiles
- From: Mark Stegman <email@hidden>
- Date: Wed, 23 Dec 2015 20:25:57 +1100
Max,
As Roger's OP was in reference to a printing press I assume he was
referring to so-called 'hi-fi' printing on an offset litho printing press.
I would have thought that G Mike Adams was right in observing that it was a
dying art as the demand for this type of printing was never mainstream and
the benefits were only apparent when the copy contained a lot of highly
saturated colours. In other words, it was a niche market with marginal
differentiation to conventional 4-col offset printing with significantly
higher setup costs. In times that are tough for connectional 4-col offset
printing I would be surprised that there is much interest outside of
'academic' circles.
Add to these challenges that of getting designers to embrace the concept
and the 'bugs' inherent in applications like Photoshop that limited the
ability to preview the end result and it ends up as a technology looking
for a purpose. In the end, the cost did not justify the means. As offset
printing in general continues to decline I would have thought that hi-fi
writing was bordering on the 'endangered species' list.
I am wondering if the distinction between multi-colour 'hi-fi' offset litho
printing and multi-colour inkjet printing where devices commonly (?) use
several 'shades' of C, M, Y and K as well as Red, Orange and Green is being
addressed. I thought sending RGB files to these devices and letting the RIP
handle the conversion was standard practice and the use of profiles for
achieving predictable results had settled on a range of standards-based
colour gamuts that reflected commonly used sets of 4-col offset printing
conditions as their target(s). What use are multi-colour profiles outside
of the RIP in this scenario?
Mark Stegman
Graphic Technologist
On 23 December 2015 at 07:31, Max Derhak <email@hidden> wrote:
> I wouldn't say that multi-color profiling is something of a dying art.
>
> I'd classify it as a fairly tricky thing to do with some serious caveats
> WRT V4 ICC profiles. Some of the challenges include:
> 1. Dimensional complexity makes characterization challenging. You need
> lots more patches to get good characterization. What works for 4 colors
> becomes overwhelming for 7 colors.
> 2. Ink Load issues (lots more ink possibilities
> 3. Dimensional complexity makes device to PCS A2Bx tags in profiles really
> really big or rather inaccurate. The size of the n-dimensiona tables is
> determined by the number of sampling points raised to the power of the
> number of device channels. Its an exponential problem often resulting in
> much sparser sampling of the device address space to keep the profile size
> reasonable.
> 4. Ink separation is much more involved to get good smooth transitions
> while achieving full gamut available with extra inks. In the B2Ax PCS to
> device tables there is only 1 way to specify how to achieve each Lab value.
> 5. Large differences in ink values for adjoining grid points in the B2Ax
> output table separation can result in greater interpolation errors between
> grid points.
>
> Thus when you use a multi-color icc profile in Photoshop it is very likely
> the preview will be less than desirable, and most likely not a faithful
> representation of the actual output.
>
> With up and coming iccMAX based multi-color profiles there are greater
> possibilities to either directly model mathematically the characterization
> of the inkset or use selection of lower dimensional sub tables for greater
> accuracy profiles of smaller size.
>
> Max Derhak (PhD)
> Principal Scientist Onyx Graphics
> Co-Chair International Color Consortion
>
> -----Original Message-----
> From: colorsync-users-bounces+max.derhak=email@hidden
> [mailto:colorsync-users-bounces+max.derhak=email@hidden]
> On Behalf Of G Mike Adams
> Sent: Tuesday, December 22, 2015 11:10 AM
> To: Roger Breton
> Cc: 'colorsync-users?lists. apple. com' List
> Subject: Re: Multi-color ICC profiles
>
> Roger,
>
> > Next step is the experimentation -- you'd imagine.
> >
> > So I installed the profile in my Windows 10 PC and turned to
> > Photoshop, naturally ;-) There must zillions of "advanced" proprietary
> packages that offer all kinds of "advanced" tools for those who thrive on
> multi-color profiles?
>
> Not really. Fact is that multi-color was always a pretty small niche of
> the industry, and nowadays many multi-color machine manufacturers as well
> as RIP manufacturers have “contoned” the process, meaning that it’s done
> internally, and what you generate is simply a CMYK profile, so, sadly, you
> could call multi-color profiling something of a dying art.
> >
> > So, I took a grayscale image and converted it to the muti-color profile.
> >
> > After conversion, I end up with three extra channels, Red, Green and
> Blue. Fair.
> >
> > Two things.
> >
> > First, read the CIE Lab values returned good neutrals (a* and b* = 0).
> > Second, the appearance of the image didn't evoke anything neutral but
> more like "red" cast?
> >
> > This may be more a question for the Photoshop Forum on Adobe's web site
> but, in case..., is the user expect to "Edit" the RGB Channels after
> conversion? To obtain a correct color appearance?
> >
> > It would seem that way…
>
> No.
>
> Fact is Photoshop has never rendered an accurate screen image of any
> multi-color profile once you go ahead and make the conversion. You may also
> have noted that if you install a multi-color profile and go look for it to
> soft-proof with, it won’t be there.
>
> However, you can get a pretty good idea of how a multi-color profile will
> look printed by going to convert to profile, selecting the mc profile you
> want, and clicking “preview.” You’ll probably have to drag the box out of
> the way, but what’s on the screen then is a pretty good representation.
> Then of course if you go ahead and click OK, it whacks out.
>
> A couple things though:
>
> First is that any multi-color profile — is made for a very specific
> printing condition. I can’t image really why you’d ever want to do a
> conversion in Photoshop to begin with. If you don’t have the entire
> printing condition at hand — machine/media/resolution/ink/density/RIP —
> then your converted result is going to be pretty useless to you.
>
> And if you do, well, then use the RIP to do the conversion. That’s what
> it’s for.
>
> Second is that properly made and properly used, the value of N-colors is
> to extend gamut beyond CMYK. Differing profiling packages have differing
> controls, and I guess it’s at least possible for someone making a
> multi-color profile to extend the N-channels into non-chroma areas. But
> there’s no reason to. So if you’re doing any tests starting with a
> greyscale image, once you get the other channels — whatever they are — they
> should all be empty.
>
>
> Mike Adams
> Correct Color
> _______________________________________________
> Do not post admin requests to the list. They will be ignored.
> Colorsync-users mailing list (email@hidden)
> Help/Unsubscribe/Update your Subscription:
>
>
> This email sent to email@hidden
>
> _______________________________________________
> Do not post admin requests to the list. They will be ignored.
> Colorsync-users mailing list (email@hidden)
> Help/Unsubscribe/Update your Subscription:
>
>
> This email sent to email@hidden
>
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden