[argyllcms] Re: Number of patches vs. profile quality

  • From: "Alan Goldhammer" <agoldhammer@xxxxxxxxxxx>
  • To: <argyllcms@xxxxxxxxxxxxx>
  • Date: Wed, 04 May 2011 08:11:32 -0400

I can only provide my own experience as a "new" user of ArgyllCMS.  I've
spent a couple of weeks getting to know the software and profiling Ilford
Gold Fiber Silk on an Epson 3880 with a ColorMunki.  I decided not to burn
through a lot of paper and restricted the patch set to 980 on 5 letter-size
pages.  I could not use the Ilford provided profile from their website as it
is absolutely terrible (not enough color density and some colors terribly
off).  In addition, my ColorMunki profile could not be used as it is a
version 4 profile and not compatible with ArgyllCMS.  The resulting profile
had a pretty low standard error (and Max error) and using a test print
(http://www.outbackprint.com/printinginsights/pi049/essay.html) achieved
very good results.  In particular, the sky in the arch picture was visually
very close to what is on the monitor.  It is hard using a ColorMunki to get
reliable spot readings to confirm color values (I have ordered an i1 spectro
which should come shortly and will use this instead of the CM for future


With respect to B&W printing, I only use the Epson ABW print driver.  It
gives a better Dmax and linearity than the normal color driver.  You can
obtain profiles for Epson printers using this driver
(http://people.csail.mit.edu/ericchan/dp/Epson3800/abwprofiles.html) or
build your own using QuadTone RIP.  It would be good if ArgyllCMS could
support this in a future release; it ought to be straight forward to
implement since only the B/W gradient is important and there are no colors
to be concerned with.  I do not have any experience with other printer
brands that might have dedicated B&W printing implementations.


One thing I would like to see in the next update is the print out of patch
locations in profcheck.  It took some time to figure out that the numbers of
the patches in the logfile that I can print out did not correspond to the
their actual location but rather the order of patches in a non-randomized
set.  I was able to go into the *.ti3 file and extract the patch locations
so that if I wanted to re-read a row I would know which patches had the
highest errors.  Even so, I suspect that the error rates are more a function
of the instrument rather than reading since doing multiple readings didn't
make a big difference.


Will I be able to get a much better profile from a larger patch set?  I'm
interest to see but I'm waiting until the i1 comes since the needed patch
size for the CM uses over twice as much paper as an i1.  In addition, the i1
can compensate for papers with OBAs (the Ilford paper that I do a lot of
printing on has some in it but far less than other papers that have an equal
white brightness.  My bottom line is how the profile performs on a real
print (though I am interested in the measurements as well but it's of
secondary importance).




From: argyllcms-bounce@xxxxxxxxxxxxx [mailto:argyllcms-bounce@xxxxxxxxxxxxx]
On Behalf Of Vittorio Villani
Sent: Wednesday, May 04, 2011 5:09 AM
To: ArgyllCMS List
Subject: [argyllcms] Re: Number of patches vs. profile quality


Hi Greame.


I am talking only about RGB printers here (my limited experience). I am
using Epson and I did some prints with Canons too.

I think that you can find something that may be interesting in this post:


I suggest you to read this post:

It would be nice to have your opinion on this.


I can give you my 2 cents.


1)      Some have reported the "blue shift issue", that you have already
commented and fixed in the 1.3.3 beta.


2)      Usually people know as "experts" (I am just a beginner) report that
it is very useful to have some low-saturation patches. I don't know how the
actual Argyll targen works; but I saw that it is often suggested to add some
patches to have a better profile, using procedures similar to the one
described in the message I have linked.


3)      I am not seeing noticeable improvements with my test prints using
more than 2500 patches. I found that 2000/2500 is a good number. I am using
targen with -G and -g256; I have seen that more patches around the grey axis
give me better results in BW printing (less colour casts). I did not try -s
and -p parameters, many suggest to use -s.


4)      I think that the possibility of making a pre-conditioning profile is
very interesting. I don't know how this actually works, but talking with
other people I found that some could get good results with less patches
(1000/1500) using a 800 patch pre-profile or a profile provided from the
paper/ink manufacturer. 

Hope this can be useful.

Vittorio Villani

> Date: Wed, 4 May 2011 13:11:45 +1000
> From: graeme@xxxxxxxxxxxxx
> To: argyllcms@xxxxxxxxxxxxx
> Subject: [argyllcms] Number of patches vs. profile quality
> I am interested in hearing about peoples experience with profile
> quality vs. number of test patches using Argyll. In particular, I'd
> like to get an impression of what aspect of a profiles quality are seen
> as improving as more patches are used. If this aspect is color accuracy,
> is there a particular area of the color gamut that is seen as the
> critical area ? If so, what area is it ?
> [I'm wondering if the profiling efficiency can be improved by some simple
> changes to the patch distribution.]
> thanks,
> Graeme Gill.

Other related posts: