Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: VitOne on April 14, 2011, 07:23:23 pm
-
I am an ArgyllCMS user. I am working only with RGB printers (Epson 4880 and 7900).
I am happy with the results that I am getting with ArgyllCMS and I would like to have some more informations about i1Profiler (the new software) because I'd like to buy it but I am not sure. I am using an i1Pro and a ColorMunki.
This is my ArgyllCMS workflow (an example):
targen -v -d2 -G -f2232 -g128 name
printtarg -v -ii1 -t360 -p310x450 name
chartread -v -H -B -T0.4 name
colprof -v -qu -i D50 -o 1931_2 -S AdobeRGB1998.icc -cmt -dpp -f -D A-E7900-Bamboo-2880dpi-i1Pro name
One of the thing that I like of this software is that I can repeat the readings that have a large DeltaE2000 using the -r paramter in chartread (I found the worst patches using profcheck) so that I can get a max Delta2000 lower than 1. Is it possible, with the Profiler software, to re-read only the worst patches (using spot reading)?
I read that OBA compensation will only work with the i1isis device. Is this true?
Is there any ArgyllCMS user that already made a comparison between this 2 softwares?
-
This is my ArgyllCMS workflow (an example):
targen -v -d2 -G -f2232 -g128 name
printtarg -v -ii1 -t360 -p310x450 name
chartread -v -H -B -T0.4 name
colprof -v -qu -i D50 -o 1931_2 -S AdobeRGB1998.icc -cmt -dpp -f -D A-E7900-Bamboo-2880dpi-i1Pro name
Greek to me. Speak English <g>.
I read that OBA compensation will only work with the i1isis device. Is this true?
Sort of, yes. There is a module called OBC which only works with the iSis. It consists of targets you visually examine. See: http://www.colorwiki.com/wiki/Optical_Brighteners_2_-_X-Rite_OBC
There is no software OBA compensation like we had in ProfileMaker Pro.
-
I read that OBA compensation will only work with the i1isis device. Is this true?
Is there any ArgyllCMS user that already made a comparison between this 2 softwares?
If you are already familiar with ArgyllCMS and use its reversed FBA compensation starting from a new UV-enabled Eye1 Basic measurements, the total purchase sum will be lower, no learning curve and more features in total. Covers calibration/profiling of wide gamut monitors too. I actually want to go that route more. So I am also interested in the answer tot that last question.
met vriendelijke groeten, Ernst
New: Spectral plots of +250 inkjet papers:
http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
-
Thanks for the answer Andrew.
I will try to explain the commands that I am using with ArgyllCMS, you can for sure find a better and probably more correct description in the very good ArgyllCMS online documentation here: http://www.argyllcms.com/doc/Scenarios.html#PP1
targen -v -d2 -G -f2232 -g128 name
I am making target test values optimized for an RGB printer (-d2) with 2232 patches and asking to have 128 patches with R=G=B (“nominally grey, of course they are not, otherwise the printer would probably need no calibration). I saw that a good number of patches with same RGB values allows to get a neutral BW later.
printtarg -v -ii1 -t360 -p310x450 name
I am using the previously generated patches to make an i1Pro-compatibile(-ii1) 360dpi 8bit TIFF image file (-t360) whit all the patches. This file will measure 310x450 mm. This is the file that I will print with PhotoshopCS5 (making myself sure to disable colour management, I personally use the “double-assign” method using 2 times the AdobeRGB profile to have as result a null transform).
chartread -v -H -B -T0.4 name
This is to read the previously printed test chard made with printtarg and printed with no colour management applied. I am using high resolution (-H) spectral mode to read it and I am asking for a very low tolerance (-T less than 1). So if the instruments gets big DeltaE 2000 errors while reading I will be asked to read again. I don’t exactly know (I have asked but I received no answer) what tolerance is –T 0.4, but I think that it should be something around max 5 DeltaE2000.
colprof -v -qu -i D50 -o 1931_2 -S AdobeRGB1998.icc -cmt -dpp -f -D A-E7900-Bamboo-2880dpi-i1Pro name
Here I am making the ICC profile. Quality is set to “Ultra”, illumination spectrum is D50, observer is the second , I am asking for OBA compensation (-f) and it will work for the selected illumination spectrum (D50 in this case, but I can use the actual illuminant if I want, there is the specific illumread utilty: http://www.argyllcms.com/doc/illumread.html) . Here: http://www.argyllcms.com/doc/FWA.html you can find more about how ArgyllCMS works with OBAs.
-cmt is telling Argyll to use as “input profile” a “standard monitor” and –dpp is telling to Argyll to use as “output” a “print observed under about 500lux”.
-S AdobeRGB indicates that I want the AdobeRGB profile to be used to manage the perceptual and saturation rendering intents.
Now the profile is ready to use. A very interesting function in ArgyllCMS is that I can check how the profile behaves using a validation process (comparing the Lab values to generate the profile to the Lab values that the profile is using). This is the command:
profcheck -v2 -k -i D50 -o 1931_2 -w profile.ti3 profile.icm > logfile.log
As output I will have a 3D graph with vectors (the more the length, the more the DeltaE is) so that I can have an immediate “idea” of how the profile performs and I will also have a log file called “logfile.log” that I can export to Excel (or other similar program) to read all the worst DeltaE patches. In this way I can decide to re-read the patches with the high DeltaE errors. You can locate the patch and read it again. You will have to run charread with –r parameter (re-read) and you can read again in spot mode just the patchs that you want to.
After this step you can get results like this:
Profile check complete, errors(CIEDE2000): max. = 0.925663, avg. = 0.200191, RMS = 0.228265
This is for an Epson 7900 with Hahnemuhle FineArt Baryta paper, patch used were about 2500.
Notice that there are many other interesting tools and possibilities in ArgyllCMS.
For example:
1) The mailing list is very active, you get fast answer to your questions.
2) All the files that you generate are text files that you can open and edit. For example you can add new patches to a previously generated test chart.
3) You can use the average tool to average measurements files.
4) Graeme Gill (software author) is always listening, adds new functions and corrects previous bugs or weakness according to users’ needs and reports. Look this discussion (and all the mailing list :)): http://www.freelists.org/post/argyllcms/PurpleBlue-shift-problem-with-Epson-printer
-
If you are already familiar with ArgyllCMS and use its reversed FBA compensation starting from a new UV-enabled Eye1 Basic measurements, the total purchase sum will be lower, no learning curve and more features in total. Covers calibration/profiling of wide gamut monitors too. I actually want to go that route more. So I am also interested in the answer tot that last question.
met vriendelijke groeten, Ernst
New: Spectral plots of +250 inkjet papers:
http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
Thanks Ernst. I am trying to learn to use ArgyllCMS but I am also interested in the whole “market” of colour management programs, so I think I will give i1Profiler a chance, also because I read many positive comments about the quality of final profiles. The software seems to have its weakness (I really don’t understand why there is no OBA compensation if you don’t have an i1isis and the EULA seems to have some “strange” stuff in it) but I read that profile generated with this software are the best ever seen…
-
but I read that profile generated with this software are the best ever seen…
By reviewers not familiar with ArgyllCMS. Packed as subjective judgements so far. Given the lack of features it can not have covered trials on a wide range of print conditions either.
met vriendelijke groeten, Ernst
Dinkla Gallery Canvas Wrap Actions for Photoshop
http://www.pigment-print.com/dinklacanvaswraps/index.html
-
I have made more than a few profiles with Argyll, and for the past weeks, we have been mirroring many of our in-house software profile builds with i1Profiler. That effort has pushed a few hundred profiles through i1Profiler (an incredibly painful process, I might add, given the poor usability of the i1Profiler software). The targets have been measured on various instruments: iSis, Spectroscan, and iCColor. We are only using the i1Profiler-generated profiles for internal evaluations, not for public use.
For papers containing significant levels of optical brighteners, i1Profiler-generated profiles produce a yellow cast unless the OBC module is used or UV-filtered measurements are made. The OBC feature, as with much of the software, is designed with a single user, profiling his own readily-accessible printer in mind. To give X-Rite their due, the resulting profiles do indeed compensate very well for OBA in standard viewing conditions. The problem is the workflow. The printers we deal with are located remotely or, in the case of a press, not something that one wants to stop production on at regular intervals. The i1Profiler OBC module assumes a linear workflow: measure the target, build a profile, generate and print an evaluation target, check it, enter the parameters back into the software, and, finally, produce a corrected profile. Needless to say this does not fly when the printer you are profiling is thousands of miles away. There is no easy way of storing and managing a large number of intermediate measurements and OBC module outputs, so thanks, but no thanks.
The software-based OBA compensation algorithms in ProfileMaker did a reasonable job on the majority of inkjet and RA-4 paper stocks. The resulting profiles were more accurate, visually smoother, and offered improved neutral rendition over using UV-filtered measurements. Some commercial press papers threw ProfileMaker for a loop. Now with i1Profiler one needs to either go through the prolonged and painful OBC module (which only works on an iSis) or go back to the cover-your-eyes and pretend UV does not exist process of UV-filtered measurements.
ArgyllCMS has powerful OBA compensation routines. Our own software borrows from this approach. It is not perfect, and for highly brightened papers or strange illuminants the X-Rite OBC gives visually superior results in neutrals. (This is largely based on previous profiles we made with Profiler or ProfileMaker and the OBC module -- i1Profiler has problems loading custom illuminant spectra). For the majority of real-world printing applications, however, I submit that the software-compensation approach makes for happy photographers and customers. Were X-Rite to improve the usability of their OBC module, it could be a valid contender. Of course, it only works with an iSis. This presents yet another example of i1Profiler's muddled marketing: it lacks high-end functionality of the previous generation of X-Rite products (PMP and Profiler), offers a workflow designed only for the please-lead-me-by-the-hand beginner, yet omits crucial OBA compensation support for the very mass-market measurement device that the software aims at.
Comparing the overall Argyll profile quality to i1Profiler gets tricky. As you are no doubt aware, minor changes to the profiling parameters in Argyll can have significant effects on the profile quality. If you are willing to burn through a large amount of time and paper, Argyll profiles can be stunningly good. For our own usage, we decided that Argyll was too brittle; i.e. the amount of fiddling with customized targets, remeasuring, rebuilding profiles with a range of input parameters, and the amount of paper needed to evaluate the results simply was not worth the performance. If you are a one-person shop, print on a limited set of papers, and use stable printers (e.g. Epson inkjets), the Argyll may make sense. Devote a few days to optimizing the process for your environment and artistic taste, and you are set.
-
Thanks for the answer.
I am mainly making profiles for me, and I am interested in profile quality, of course I don't want to loose tons of paper.
I didn't understand why you said that you need a lot of paper with ArgyllCMS. Once you have printed the target and read it you can use the same readings, add other readings, and make many profiles with different settings (you can also make a profile with OBA compensation and one without, or many profiles for different illuminants).
Did you mean that with Argyll you need a bigger number of patches? I saw that with 2000 patches you can get very good results.
Are you talking about RGB printers? What kind of rendering intents have you experimented?
-
To cut a long story short, Argyll is typical Unix command-line software, hugely powerful and infinitely flexible but a pain to setup.
To use Argyll you need to read the documentation, then you really need to read the documentation again, and by the way did I mention that the documentation is sometimes wrong, so you need to read the forums.
On the other hand Argyll is *free*, mostly works, gives give decent results, and you can use it to supply clients with profiles.
MP, PMP, i1Pro and cousins are basically expert systems in a box, and you are paying not only for the profiling engine but also for the work that went into pre-configuring your workflow into the common mouse-selectable use-cases.
Edmund
Thanks for the answer.
I am mainly making profiles for me, and I am interested in profile quality, of course I don't want to loose tons of paper.
I didn't understand why you said that you need a lot of paper with ArgyllCMS. Once you have printed the target and read it you can use the same readings, add other readings, and make many profiles with different settings (you can also make a profile with OBA compensation and one without, or many profiles for different illuminants).
Did you mean that with Argyll you need a bigger number of patches? I saw that with 2000 patches you can get very good results.
Are you talking about RGB printers? What kind of rendering intents have you experimented?
-
> Is there any ArgyllCMS user that already made a comparison between this 2 softwares?
I prefer ArgyllCMS by far. However I add manually a lot of patches that are close to "neutral" to those created by targen, as well as the whole set of Munsell original colours and CMY gradients. For some printers I found using Argyll "Calibrating Printers" workflow to be very beneficial.
-
This presents yet another example of i1Profiler's muddled marketing: it lacks high-end functionality of the previous generation of X-Rite products (PMP and Profiler), offers a workflow designed only for the please-lead-me-by-the-hand beginner, yet omits crucial OBA compensation support for the very mass-market measurement device that the software aims at.
Ethan.
Thank you for the short review. The day Argyll gets nice GUIs for some platforms ............
One wonders if its monopoly in CM hard and software causes this muddled marketing. The products will sell anyway. Even when alternative software can be used the spectrometers are X-Rite models, Qualux supports two X-Rite models and only recently the Spyder3. The majority of the spectrometers supported by ArgyllCMS are X-Rite models. There are Barbieris and Spyders but they fall in other categories. There was a time when 3 or 4 spectrometers of an equal quality competed in the Eye 1 Basic market. In hindsight one may wonder if the EU should have allowed this merger. On the other hand a HongKong based software company could resurrect the SpectroCam with better electronics and produced cheaper in China.
met vriendelijke groeten, Ernst
New: Spectral plots of +250 inkjet papers:
http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
-
It is difficult to create a GUI for ArgyllCMS without dumbing it down to less flexibility and functionality.
-
It is difficult to create a GUI for ArgyllCMS without dumbing it down to less flexibility and functionality.
Why? I've done a fair amount of C and C++ programming and Microsoft has made the GUI stuff very easy so that one only needs to worry about what you want the program to do rather than the interface. From what I've seen of the Argyll documentation, application of a GUI interface should not be that difficult (and don't ask me to do it; I'm retired from programming). Most of the commands are basic switches that can be dealt with in a straight forward manner.
-
The difficulty has nothing to do with programming difficulty. To me, ArgyllCMS strength is in allowing many different workflows. It is impossible to wrap a GUI around all those, especially when it comes to printers. However it is easy enough to make a GUI for a "regular" user who for some reason can't use command line. Such GUI exists for monitor calibration for example - http://hoech.net/dispcalGUI/
-
Well, maybe it is time to start getting all the CM experts to design what they want, and start tailoring something to make Argyll (or adapt Argyll) to suit. If there are issues with i1Profiler then a decent Argyll front end that does 90% of what people want (with the option to CLI it for special options) would certainly force a bit of healthy competition. If people are interested let me know and I can see about devoting some time to this.
-
Here is one GUI:
http://www.digifab.com/ArgyllCMSGUI/
FYI, Steve is the one that writes the RIP that Digifab sells.
Some more info on GUIs for Argyll:
http://colorhacks.blogspot.com/search/label/ArgyllCMS
-
> Some more info on GUIs for Argyll: http://colorhacks.blogspot.com/search/label/ArgyllCMS
The link to JavaScript wrapper is changed to http://x3.ntf.uni-lj.si/~gojc/ArgyllCMS_GUI/
-
To cut a long story short, Argyll is typical Unix command-line software, hugely powerful and infinitely flexible but a pain to setup.
To use Argyll you need to read the documentation, then you really need to read the documentation again, and by the way did I mention that the documentation is sometimes wrong, so you need to read the forums.
On the other hand Argyll is *free*, mostly works, gives give decent results, and you can use it to supply clients with profiles.
MP, PMP, i1Pro and cousins are basically expert systems in a box, and you are paying not only for the profiling engine but also for the work that went into pre-configuring your workflow into the common mouse-selectable use-cases.
Edmund
Thanks for the answer to all of you. I think that ArgyllCMS results are much more than "decent". I am interested in the “long story”, if you have time to write it I will be happy to read it.
I am more interested in profile quality than in usability, although I personally think that ArgyllCMS is very flexible also because of his workflow based on command lines. After some painful time spent trying to understand how the whole thing was working I am getting very nice and fast procedures that allow me to get excellent results in a short amount of time. I have different workflows already set in a simple txt file, and I already reported one (that I found good for many fine-art papers).
At the moment I am happy with ArgyllCMS the way it is and the author, Graeme Gill, told many times on the mailing list that a GUI for this software is far from being released by him. I know just some basics about programming, but I think that it is much more difficult to make a software like Argyll than to make a good and complete GUI for it. I found dispcalGUI a clear example of a nice and complete GUI for Argyll. Anyway I am not really interested in this topic, my original question was more about profile quality than usability of the software. Anyway of course usability is an essential requirement for a successful application.
I should receive soon the new i1 Publish software, and I will for sure compare the 2 software.
I prefer ArgyllCMS by far. However I add manually a lot of patches that are close to "neutral" to those created by targen, as well as the whole set of Munsell original colours and CMY gradients. For some printers I found using Argyll "Calibrating Printers" workflow to be very beneficial.
I would be very happy to know something more about the patches you added and how you did it. Can you please explain? Are you using txt2ti3? Can you describe what patch are you using and the whole procedure? Thanks.
-
Edmund is absolutely correct in his assessment. Argyll's strengths and shortcomings both stem from its flexibility. The reason why I mentioned having a large roll of paper on hand before starting your first Argyll profiling effort is that the multitude of adjustable parameters strongly influence profile quality. Determining the optimal configuration requires several iterations of profiling and test printing.
i1Profiler and other commercial packages include more intelligence about workable construction parameters. You need to supply the smarts with Argyll or, as Edmund says, read the forums and mailing lists.
My experience with Argyll is that the printer module is better optimized for CMYK than RGB. You can indeed get good RGB results, but doing so requires tweaking the target. If you are working on your own printer where patch and page count are not critical, adding a series of neutral and near-neutral patches helps. I have not used Argyll's target creation utilities, so I am not the one to ask for advice on that count.
-
Thanks.
I don't understand why you say that you need to tweak the target to get good RGB results. I know many ArgyllCMS users and just a few of them does more than “targen -d2 -fXXXX name” to build their target. The –d2 parameter already makes optimized patches for RGB printer, and, also if I have experimented a lot, I saw just small improvements in using more and different parameters.
The –G parameter might be good (but it does not change a lot the final profile if you are using a number of patches above 2000 for what I saw). I personally got no increase in profile quality using more than a preliminary profile in targen and, for what I saw, it is also difficult to see improvements using a single pre-conditioning profile (-c parameter) if you are using 2000 patches or more in the first profile. Anyway the “nice” thing about Argyll is that you can reuse the targets you have printed expanding them with new ones without having to do everything from the beginning, because all the files that the software make are easy-editable txt.
Please consider that I am talking about people that need calibration for printers like Canon ipf series, Epson Stylus Pro, HP Z, so ink jet printers for photography; and I have direct experiences only with a few printers. What I said is related only to my limited direct experience, and maybe we are talking about different “printing environments”.
I would like to have a clarification about the procedure of adding neutral and near neutral patches: do you mean R=G=B patches or what?
I have no personal experience in CMYK so I will not say anything about it.
-
Maybe you should explain why you think Argyll will be better than the software that came with your instrument?
Color consultants with good eyes like Ethan or Iliah have very special needs, and appreciate freedom and flexibility, but why should Joe Average take the trouble to use Argyll if he is not already on Linux?
The only reason I ever used Argyll was that I did some profile linking. Later I constructed a turnkey system as a request from a spectro manufacturer - Barbieri - who wanted a Linux solution.
Now I have another reason, linked to license issues with i1Pro.
Edmund
Thanks.
I don't understand why you say that you need to tweak the target to get good RGB results. I know many ArgyllCMS users and just a few of them does more than “targen -d2 -fXXXX name” to build their target. The –d2 parameter already makes optimized patches for RGB printer, and, also if I have experimented a lot, I saw just small improvements in using more and different parameters.
The –G parameter might be good (but it does not change a lot the final profile if you are using a number of patches above 2000 for what I saw). I personally got no increase in profile quality using more than a preliminary profile in targen and, for what I saw, it is also difficult to see improvements using a single pre-conditioning profile (-c parameter) if you are using 2000 patches or more in the first profile. Anyway the “nice” thing about Argyll is that you can reuse the targets you have printed expanding them with new ones without having to do everything from the beginning, because all the files that the software make are easy-editable txt.
Please consider that I am talking about people that need calibration for printers like Canon ipf series, Epson Stylus Pro, HP Z, so ink jet printers for photography; and I have direct experiences only with a few printers. What I said is related only to my limited direct experience, and maybe we are talking about different “printing environments”.
I would like to have a clarification about the procedure of adding neutral and near neutral patches: do you mean R=G=B patches or what?
I have no personal experience in CMYK so I will not say anything about it.
-
Thanks Edmund.
I don’t think that I have ever said that Argyll is better than the software that came with my instrument, and I don’t think so. I am using different software and I started to “play” with Argyll for personal interests more than for real need, I had a ColorMunki and the software by X-Rite is not really “complete” also if it works good. I started with a ColorMunki and I also used PM 5 and i1Match and I am happy with both of them. I don’t think that using Argyll is a trouble, it just needs some time to learn it. Joe Average could be a ColorMunki user, or he would like to use is OEM i1Pro D with an update software, or it could be just a person, like me, searching for the best possible result for fine-art printing.
I was just curious to have some input about profile quality from other people. You are using Argyll just because it is the only possible solutions? I know may user like me, with just a “simple” printer like the ones that I mentioned before, that are more happy with the results they get from Argyll than with the results from i1Match. The person capable of making the best fine-art prints that I have ever seen is using a Barbieri SpectroSwing and ArgyllCMS…
-
Here is a workflow I found to be pretty robust with ArgyllCMS for RGB printers:
1. printer calibration as described http://argyllcms.com/doc/Scenarios.html#PC1 but I use -s256 for targen
2. do a simple preliminary profile programming targen to -d2 -G -s256 -g256 -f1860
3. take Munsell values from http://www.cis.rit.edu/research/mcsl2/online/munsell_data/1929.dat and convert them to RGB values using the profile above, taking into account those are defined for illuminant C (I have my own tools to do this, Excel spreadsheets do exist on the web to calculate necessary transforms, but BabelColor PatchTool may be easier, and a useful program anyway)
4. use targen to create something like 3000 patch target, using preliminary profile for optimized coverage
5. use Excel or PatchTool or your own simple programme to convert the target to something ColorPort can import
6. produce a desaturated (near-grey) target using MeasureTool or pre-made target and stripping all unnecessary patches (I do it differently, but this works well too)
7. take the target from 5. to ColorPort, add Munsell collection, add desaturated patches, print and measure with ColorPort
8. use txt2ti3 and proceed with Argyll
-
The difficulty has nothing to do with programming difficulty. To me, ArgyllCMS strength is in allowing many different workflows. It is impossible to wrap a GUI around all those, especially when it comes to printers. However it is easy enough to make a GUI for a "regular" user who for some reason can't use command line. Such GUI exists for monitor calibration for example - http://hoech.net/dispcalGUI/
I am using discpalGUI on an Ubuntu system. That is a GUI that works nicely and as far as I can judge doesn't compromise the underlying features of ArgyllCMS on monitor control. The same positive experience is described on the page that discusses the Argyll GUIs. The other GUIs seem to have their flaws. I was on the linked pages before and the 2008 date as the most recent isn't encouraging and some links no longer lead to the right pages. That and your last message explaining how you create RGB printer profiles suggests to me that better GUIs are needed + better manuals. Measure Tool and ColorPort do not have an Argyll equivalent I guess, so something is lacking in the drivers/features too. A GUI that does allow the creation of profiles without too much hassle but with an advanced button on each page for the special tricks + enough backfeed on what the underlying modules do is needed to make ArgyllCMS a usable alternative in the market.
I am also surprised about the number of patches Argyll users prefer. Reading the Argyll forum I see similar high numbers for printing systems/media where I doubt the systems have the consistency in time etc to be the right base for profiles created that way. For example 3000-6000 patches on an Epson 4800 with offset paper + third party inks. I see more problems created there with small sized patches to make those numbers practical where a smaller number of patches, more readings and correct averaging + iteration could prove to be better. Just a gut feeling.
met vriendelijke groeten, Ernst
New: Spectral plots of +250 inkjet papers:
http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
-
Dear Ernst,
You are right and I was unclear. dispcalGUI is an excellent and highly usable system, and that is why it is the only GUI I mentioned directly.
For printer calibration a GUI around target creation and printing is something that needs to be addressed first of all. It needs to combine the features of targen, ColorPort, PatchTool and allow individual/group patch editing. Alternatively, a GUI can be a combo of targen, PatchTool, and individual/group patch editing while ColorPort/MeasureTool/Barbieri target reader/etc can be used as is to support those instruments that are not supported by Argyll directly.
As far as my experience re number of patches, even with all "native" consumables I can't make a profile that I like using less than 2500 patches, normally it takes 3000 and more. That is with any profiler, not just Argyll.
-
The question of how many color patches are necessary for a good printer profile depends on the color mode, printer's characteristics, the profiling software, and how intelligently the patches were chosen. For the sake of argument, let's define the minimum target patch count as the number of patches where adding an additional 200 (or 10%, whichever is larger) makes no visible difference in profile accuracy or, more often, output smoothness and resistance to visual artifacts. Relative, eyeball-gauged profile accuracy can most easily be assessed by neutrality of grayscales. Visual artifacts include abrupt transitions or banding in shadows and smoothness in the transitions to maximally saturated colors. If a profile shows hue shifts - blues turning purple or reds going orange - this indicates far too few patches or, more likely, poor profiling software.
The single largest factor in determining how many color patches are necessary is the output color mode. RGB has three degrees of freedom, CMYK four, and n-color more still. A basic color sampling strategy is to pick color points spaced evenly from the lowest to highest possible values. Building a sampling space containing 9 points on each axis requires 729 points for RGB (3 axes, so 9^3 = 729). The same grid interval for CMYK requires 6561 color patches (9^4 = 6561). Yes, ink limiting may reduce the total number of CMYK points, but the fact stands that you need significantly more color samples to describe a CMYK printer than for one that uses RGB input. Intelligent patch distribution is a must for CMYK or n-channel printers. I'll only talk about RGB-mode profiling from now on, primarily because that is what I am well-versed on.
A well-behaved printer, one that does not show abrupt transitions in color, density, or tone, can be profiled using a smaller array of patches than a less predictable device. Non-linear output can be modeled accurately only by having a sufficiently fine mesh of evaluation points. A predictable printer can be adequately profiled (using our above definition of no visual improvement after adding more patches) with as few as 729 patches with Monaco Profiler or PMP. Prime examples of such a printer were the Fuji Pictrography series. Looking through our measurement DB, the fewest patches see us using for a Pictro profile using Argyll was 1500. This work dates back a few years, but not that much has changed in terms of Argyll's fundamental behavior.
Typical inkjets benefit from at the Profiler maximum target of 1728 patches or as much as 4096 for PMP. This assumes a regular RGB grid spacing of 12 points for the 1728 patch count target or 16 for the 4096. Using the default RGB target that Argyll spits out, I agree with VitOne: 2000 patches is the minimum you need to use.
Building a target that concentrates on problem areas can reduce the necessary patch count. Profiler uses fixed layouts, but PMP (and i1Profiler) allow customized targets. A good starting point is the neutral axis. All our targets include many neutral patches and near-neutrals - e.g. R=0. G= 0, B=15, (7, 7, 15), ... , (239, 239, 255). Repeat for red, green, C, M, and Y. Additional points can be added in areas where the printer behavior gets non-linear. Argyll benefits from this technique.
A slightly different approach is to first characterize a small target to determine the printer's basic behavior then use that information to optimize the color patch distribution for the main target. ProfileCity originated this method; it has since been used by DataColor, ColorMunki, and Argyll. At the cost of making multiple measurements and generating unique targets for each printer, you can create a superior accuracy profile from fewer overall patches. This is what Iliah describes above as the first step in his profiling flow.
Argyll is the only package I am aware of that splits the correction curves calculated from the initial lin target apart from the profile. They need to be manually inserted into the profile later. This makes the math somewhat easier but seems like an unnecessary step to me. Since we build profiles on a remote basis, the attractions of an iterative approach are outweighed by the practicalities of only requiring customers to print a single target. This means we either needed to print a huge target, compromise on accuracy with a smaller patch set, or get clever with how we weight and process the data. We went the latter route, the details of which would only interest serious geeks.
-
Nathan,
A nice exposé and with the contours of the print environment + workflow sketched I can understand the choices.
Iliah sketched profiling an RGB-device printer so using 3000 patches is halfway the 1728 to 4096 you quote.
Predictable and not-predictable is something else than the consistency of printers in time, etc. Building a precise profile on a non -linear, more complex structure is alright if the print conditions are stable or can be kept (more or less) stable by calibrating. The practice of reducing patch sizes to make high number patch targets possible may increase bad readings and by that the repeatability of the measurements becomes critical. Not that homogene non-inkjet media and disputable consistency in inks used is what I also referred to. This thread on the ArgyllCMS forum in particular:
http://www.freelists.org/post/argyllcms/Black-turning-down-problem-help
met vriendelijke groeten, Ernst
New: Spectral plots of +250 inkjet papers:
http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
-
Ernst,
The forum post you linked to in many ways shows both the strength and complexities of Argyll. Brief summary: An Argyll user found a paper and ink combination where her inkjet profiles created blue bands in what should have been dark neutral ramps. The cause was the interaction between third party ink and printing on plain paper, the combination of which could not hold high ink density. A patch with C=M=Y=100, K=0 printed lighter than when K was included. Argyll's behavior is to maximize mathematical accuracy and color gamut at the expense of output smoothness. Also, many of Argyll's most powerful adjustments can only be applied printing with independent control over the black channel, i.e. in CMYK mode.
Looking at the output the forum posted included, I suspect that i1Profiler could tame the behavior with use of the "smoothness" slider. An enterprising type with more time than I could experiment as all the reference and measurement data are included in the profiles. The online discussion went through a series of possible Argyll parameter adjustments, some that Graeme marked as "experimental". With enough test prints and time, Argyll can do the job in most cases.
Cheers,
Ethan
-
I've downloaded the Argyll package but have not had the opportunity to do any experimenting. I suspect what Ethan, Edmund and others are commenting on is the time and effort that needs to be put into Argyll to get a good profile versus getting a pre-packaged approach as one would get with i1Profiler. At what point this time becomes real money is not clear to me at present but may be after fooling around with it for a while. I like the idea of using my ColorMunki as the chart reader since this will not require a further investment in hardware.
-
When I bought my ColorMunki my hope was that I could progress from the CM software to ArgyllCMS and see my profiles improve. In this way I could leverage a relatively inexpensive instrument to make good profiles, since purchasing an Eye-One was not an option for me. That has turned out to be the case.
However, I too am curious about what the i1Profiler would buy me. I'm glad to see the topic discussed.
Terry.
-
Ernst,
The forum post you linked to in many ways shows both the strength and complexities of Argyll. Brief summary: An Argyll user found a paper and ink combination where her inkjet profiles created blue bands in what should have been dark neutral ramps. The cause was the interaction between third party ink and printing on plain paper, the combination of which could not hold high ink density. A patch with C=M=Y=100, K=0 printed lighter than when K was included. Argyll's behavior is to maximize mathematical accuracy and color gamut at the expense of output smoothness. Also, many of Argyll's most powerful adjustments can only be applied printing with independent control over the black channel, i.e. in CMYK mode.
Looking at the output the forum posted included, I suspect that i1Profiler could tame the behavior with use of the "smoothness" slider. An enterprising type with more time than I could experiment as all the reference and measurement data are included in the profiles. The online discussion went through a series of possible Argyll parameter adjustments, some that Graeme marked as "experimental". With enough test prints and time, Argyll can do the job in most cases.
Cheers,
Ethan
Nathan,
I have followed that thread at that time and there were more messages that questioned the approach. First of all the user was trying out ArgyllCMS and printer software that was newly developed + a third party ink that I have used in the past and its consistency wasn't that good. Not between deliveries and quite fast pigment settling in the printer. The paper a plain paper quality printed with 127 LPI screening. Patches 6 mm square. One of the observations by the user was that about twice the number of patches actually made the profile worse. Others advised the user to start with a better quality paper. In general the impression I got was that there was no rigid base for creating good profiles and none of the Argyll tweaking or more patches delivered a satisfying result. This sure was an unpredictable printer/media combination and ArgyllCMS could have delivered something in the end but given the inconsistencies the effort wasn't worth it, next week this profile doesn't fit. A gut feeling again.
If there is consistency in the printer+media and the measurements are accurate and repeatable I can see the advantage of more patches to tame/describe unpredictable printer output. Less patches for a well behaving printer, one that obeys to certain expectations about color mixing, I can understand that too. But the situation as sketched above had all the elements to use averaging, fewer patches, smoothing and deliver a usable profile for the time being.
It doesn't matter much in this discussion but I did read something else:
>>After careful investigation, I discovered that on this paper and with theseinks (MISPro K4, only Photo K and Light K used to simulate littler dots) the CMY combination (100 100 100 0) is actually DARKER than CMYK (100 100 100 100)!<
Using PK on a plain paper is already a bad choice and the user mentioned that.
This is an extreme case but I was reminded of it when reading the patch numbers in this thread. I have learned something about the use of more patches in this thread but I still think it can be wasted on lost causes too.
met vriendelijke groeten, Ernst
Try: http://groups.yahoo.com/group/Wide_Inkjet_Printers/
-
Here is a workflow I found to be pretty robust with ArgyllCMS for RGB printers:
1. printer calibration as described http://argyllcms.com/doc/Scenarios.html#PC1 but I use -s256 for targen
2. do a simple preliminary profile programming targen to -d2 -G -s256 -g256 -f1860
3. take Munsell values from http://www.cis.rit.edu/research/mcsl2/online/munsell_data/1929.dat and convert them to RGB values using the profile above, taking into account those are defined for illuminant C (I have my own tools to do this, Excel spreadsheets do exist on the web to calculate necessary transforms, but BabelColor PatchTool may be easier, and a useful program anyway)
4. use targen to create something like 3000 patch target, using preliminary profile for optimized coverage
5. use Excel or PatchTool or your own simple programme to convert the target to something ColorPort can import
6. produce a desaturated (near-grey) target using MeasureTool or pre-made target and stripping all unnecessary patches (I do it differently, but this works well too)
7. take the target from 5. to ColorPort, add Munsell collection, add desaturated patches, print and measure with ColorPort
8. use txt2ti3 and proceed with Argyll
Thanks for the complete answer. I am not expert in Colour Management and so I need some help to understand.
I hope that I don’t good too far from this topic asking you:
a) Can you please explain to me why you are taking the Munsell values? Is there a way to use ColorThink to convert the values you have linked to the RGB values of the profile? Why you suggest to do this?
b) How do you decide what patch you want to remove in point 6? How many desatured patches are needed?
c) Can you explain why you are converting Musell values in the RGB of the preliminary profile? How can I do this? How can I be sure that all the values are inside the gamut of the preliminary profile that I made? Do I need to use a specific rendering intent to do this?
d) Why you are using desatured target?
e) How many patches does this method require? 1800 + 3000 + Munsell (957 patches) + desatured (I don’t know how many). Are you using this huge amount of patches also with other profiling programs or you find that this is needed only with ArgyllCMS?
f) Can you give me your PayPal address so that I can offer you a beer?
-
Thanks for the complete answer. I am not expert in Colour Management and so I need some help to understand.
I hope that I don’t good too far from this topic asking you:
a) Can you please explain to me why you are taking the Munsell values? Is there a way to use ColorThink to convert the values you have linked to the RGB values of the profile? Why you suggest to do this?
I wondered the same thing myself. The linked data set contains the values from the 1929 Munsell Book of Color. This set had limitations, which were addressed in a 1943 revision. RIT has data sets from the 1943 renotation, which is probably more useful. The "Real" data (http://www.cis.rit.edu/research/mcsl2/online/munsell_data/real.dat). These values represent colors that fall within the Macadam limit - the theoretical maximum color gamut for reflected colors. Note that many of these colors will not be found in the real world, as the Macadam limit assumes perfect 100% reflectance at all wavelengths and that the object is also a black hole capable of reflecting 0% of incident light. All the Munsell color values were calculated for CIE Illuminant C, and therefore need pre-conversion to D50.
In all, I can't see much real-world use in including the Munsell colors (Real, Total, or circa 1929) in a printer test target. The only time we used a small subset of these was for a brewery advertising campaign. Their beer colors were evaluated using Munsell colors, and we wanted to get those spot-on.
b) How do you decide what patch you want to remove in point 6? How many desatured patches are needed?
[SNIP]
d) Why you are using desatured target?
I can't answer fir Iliah, but we see improved neutral accuracy from including ramps of neutral and near-neutrals at a closer spacing than the overall patch array. A series of patches spaced every 8 RGB points apart should suffice for the near-neutral, or desaturated, colors. For most printers, having the dominant color 16 RGB units more saturated than the rest is a good default. As I mentioned in a previous post (admittedly with typos in the key example), an example series for red-dominant would be (15, 0, 0), (23, 7, 7), ... , (255, 239, 239). Repeated for each primary (RGBCMY) this gives 6 * 32 = 192 additional color patches.
-
Using Munsell colours improve profiles, both to my eyes and to the eyes of my friends-photographers. If you look into history, that is how Munsell collected the data the reason becomes obvious - the system is a grid in its own, with "perceived equidistance" between colours. I tend not to use renotated colours because a lot of them are results of interpolation; and the total number of the colours is too large ;) The benefit of using several different grids, created on different principles, for profiling (and that includes neutral ramps, semi-neutral ramps, and colour ramps) is that it damps spontaneous oscillations which occurs when calculating nodes of 3-D LUTs. Another benefit is creating certain anchor points (this is achieved mainly using gray, coloured, and semi-neutral ramps).
Conversion to a preliminary profile means that printed colours will already be close in their look to correct reproduction, and this inhibits oscillations and provides more anchor points to make a well-behaved profile. With such a technique one needs more patches than usual (about 2x more) to make the profile smooth and avoid posterization.
To convert Munsell colours to RGB you can use an Excel spreadsheet (Google lists several), a self-written programme (you can borrow code snippets even from a FORTRAN public library or use LCMS, see Colorspace Conversions), free ColorLab, Bruce Lindbloom site (useful also to check how good is your own self-written converting routines), PatchTool etc. There are numerous ways to do this with as little or as much effort as you wish.
-
I've downloaded the Argyll package but have not had the opportunity to do any experimenting. I suspect what Ethan, Edmund and others are commenting on is the time and effort that needs to be put into Argyll to get a good profile versus getting a pre-packaged approach as one would get with i1Profiler. At what point this time becomes real money is not clear to me at present but may be after fooling around with it for a while. I like the idea of using my ColorMunki as the chart reader since this will not require a further investment in hardware.
Alan,
Go for it! You have perfectly understood the issues and have the right attitude. Provided you have time to hone your usage of the software, Argyll will reward you with very good profiles.
Edmund
-
Alan,
Go for it! You have perfectly understood the issues and have the right attitude. Provided you have time to hone your usage of the software, Argyll will reward you with very good profiles.
Edmund
Set up was pretty easy once I figured out how to change the PATH command (it's been a lot of years since I had to modify such a file (back in the DOS days!!). Everything is pretty intuitive as long as you are careful in the command line typing. I manged to generate a set of targets but for ColorMunki uses it does consume a considerable amount of paper. A 1860 patch set is 10 Letter size pages (of course you need the larger size patches to insure not making a mistake in reading). I'll give this a run on some Ilford Gold Fiber Silk and see how the profile turns out. Thanks to all who posted to this thread.
-
Set up was pretty easy once I figured out how to change the PATH command (it's been a lot of years since I had to modify such a file (back in the DOS days!!). Everything is pretty intuitive as long as you are careful in the command line typing. I manged to generate a set of targets but for ColorMunki uses it does consume a considerable amount of paper. A 1860 patch set is 10 Letter size pages (of course you need the larger size patches to insure not making a mistake in reading). I'll give this a run on some Ilford Gold Fiber Silk and see how the profile turns out. Thanks to all who posted to this thread.
If you want you can try this to save some paper with your ColorMunki: http://www.youtube.com/watch?hl=en&v=9PebBQOpP3I&gl=US
I made many profiles with CM and i1Pro patches; I can say that max DeltaE2000 is below 1.0 using profcheck and reading again the patches that had the wrost DeltaE: http://www.freelists.org/post/argyllcms/How-to-improve-an-RGB-printer-profile-using-ti2-and-ti3-files This was one of my first experiments with ColorMunki and the i1Pro patches, one good thing of Argyll is that you can, as I said, find and read again the patch that were not read correctly. I was looking for a similar function on i1Profiler or other software but I could not find anything, can anybody help me?
Other questions:
1) In Argyll I can set the resolution and the bit depth of the output TIFF file. Is i1Profiler I saw that is only possibile to have 100dpi output files. On Epsons is usually better to send to the printer files at 360dpi. Is it possibile ot set 360 dpi output in i1Profiler?
2) I think I read that the last version is not the 1.0 that I can find on the X-Rite web site. Can you tell me what is the last version of the software and where I can get it?
3) I read that i1Profiler is not working well with custom targets. Does anybody have experience in this? Do you suggest to use only his own patches? What number of patches will you use for a profile of an Epson 7900 with luster/baryta papers?
-
If you want you can try this to save some paper with your ColorMunki: http://www.youtube.com/watch?hl=en&v=9PebBQOpP3I&gl=US
I think the tolerances are too close to do it that way. I did get everything installed and printed out a too ambitious target of 2000 patches (this uses 10 letter size pages with rows 14x14. Even then it took some time getting the hang of things. It worked pretty easily once I got it going with a straight edge to guide the Munki. I also didn't have any problem installing the CM driver as some others have. Anyway I created a profile and it was a disaster. It had an overall light blue cast with no paper white showing at all. I don't know if there was a misread or a calibration problem (I did get an OK when asked to calibrate it). I'm too tired to spend any more time on this tonight. I did print out a patch set on two pages with fewer patches just so I can experiment since reading 10 pages is a bit of a drag.
I'm getting used to all the command line arguments but I guess have been spoiled over the years with Windows programs. It's been a while since I've had to write any code or do interesting things.
-
Just updating the experience. I don't think that anything was wrong with the profile that I created using the 2000 patches. I was too bleary eyed last night to see things straight. I did print out a smaller patch set (500) since that is easier to read multiple times if necessary. The interesting thing is that I got a better profile with the 500 patches than with the larger set (there were probably a lot of bad reads with the larger set since the standard errors were larger). I need to figure out how to view and compare profiles now. I'm not sure that I want to spend the big $$$s on ColorThink Pro and have downloaded and am looking at Gamutvision which is 1/4 the price. ColorThink 2 now runs on Win7 so that one is in the running as well. I would welcome recommendations for this and what is really the best way to evaluate profiles using the Argyll tools. One thing I have found out both using Gamutvision and my standard test print is that the Ilford profile for Gold Fiber Silk is really not very good. Colors are too light and the gamut is smaller than the profiles I have prepared from ColorMunki software and Argyll.
Thoughts on a path forward are welcome.
-
Hi Alan,
I for one appreciate your update.
I very much need to try the ArgyllCMS software w/ my ColorMunki.
Do you find the 500 patch profiles better than the CM software ones?
How much better are they? By a large degree, or slightly so?
As I am just a hobbyist and not making monies from my work, I choice the Gamutvision route.
I like it very much!
It does require a steep learning curve, as the site offers limited documentation.
Though, the documentation offered is very good, it leaves a lot of areas untouched.
I searched for other information, and found a couple of older nice articles here on LULA.
And, I found some useful postings on Naturescapes.net
That was pretty much it.
I was saddened to see that the Gamutvision site forums are almost void of posts!
I was going to ask detailed questions here, though, thought better of it, as I would probably bore the others who use the ColorThink software.
Anywho, FYI, there is a discussion of options and their meanings of Gamutvision on going there:
http://www.naturescapes.net/phpBB3/viewtopic.php?f=44&t=195316
Ed June
Just updating the experience. I don't think that anything was wrong with the profile that I created using the 2000 patches. I was too bleary eyed last night to see things straight. I did print out a smaller patch set (500) since that is easier to read multiple times if necessary. The interesting thing is that I got a better profile with the 500 patches than with the larger set (there were probably a lot of bad reads with the larger set since the standard errors were larger). I need to figure out how to view and compare profiles now. I'm not sure that I want to spend the big $$$s on ColorThink Pro and have downloaded and am looking at Gamutvision which is 1/4 the price. ColorThink 2 now runs on Win7 so that one is in the running as well. I would welcome recommendations for this and what is really the best way to evaluate profiles using the Argyll tools. One thing I have found out both using Gamutvision and my standard test print is that the Ilford profile for Gold Fiber Silk is really not very good. Colors are too light and the gamut is smaller than the profiles I have prepared from ColorMunki software and Argyll.
Thoughts on a path forward are welcome.
-
Hi Alan,
I for one appreciate your update.
I very much need to try the ArgyllCMS software w/ my ColorMunki.
Do you find the 500 patch profiles better than the CM software ones?
How much better are they? By a large degree, or slightly so?
As I am just a hobbyist and not making monies from my work, I choice the Gamutvision route.
I like it very much!
It does require a steep learning curve, as the site offers limited documentation.
Though, the documentation offered is very good, it leaves a lot of areas untouched.
I searched for other information, and found a couple of older nice articles here on LULA.
And, I found some useful postings on Naturescapes.net
That was pretty much it.
I was saddened to see that the Gamutvision site forums are almost void of posts!
I was going to ask detailed questions here, though, thought better of it, as I would probably bore the others who use the ColorThink software.
Anywho, FYI, there is a discussion of options and their meanings of Gamutvision on going there:
http://www.naturescapes.net/phpBB3/viewtopic.php?f=44&t=195316
Ed June
Ed,
I had to carefully read the Argyll documentation as it is quite sparse and just when you think you have done something right, it turns out to be wrong. I finally decided to build serious profiles for both Ilford Gold Fiber Silk and Museo Silver Rag using Argyll and the ColorMunki (I have ordered a i1 Pro but it's on back order and won't be here for about 10 days or so). With the CM I settled on a 980 patch set as that prints out on five pages; I didn't want to use more than that because this was just a pilot experiment and reading too many pages is tedious. The 980 patch set gave quite good readings as judged by the standard errors (in the case of the Museo paper it was a little higher but that paper has a warmer cast than does the Ilford paper). The gamut plots as viewed in a simple viewer looked good. For me the real test was the test print that I use which is the Outback test print that is a composite of a number of the Atkinson images. I found that these profiles did the best job yet on the sky in the arches image. I'm not going to do much more until the i1 comes in since I do want to read specific patches to see how close the profile comes to matching (reading patches with a CM is not as easy).
Both of my Argyll profiles were better than those supplied through the manufacturers' websites. The IGFS profile supplied by the manufacturer is particularly poor.
-
I very much need to try the ArgyllCMS software w/ my ColorMunki.
Do you find the 500 patch profiles better than the CM software ones?
How much better are they? By a large degree, or slightly so?
Hi Ed.
I bought my ColorMunki expecting good, but not necessarily great, profiles. That's what I've found. I also bought it hoping that I could leverage ArgyllCMS to take things up a notch.
I've profiled a canvas, a coated canvas, a matte paper and a gloss paper with ArgyllCMS. The ArgyllCMS profiles have been better. My comparison was to ColorMunki profiles, other than the canvas profile which was provided by the manufacturer.
My first profile was done using ~1300 patches and the rest use more than 2000 patches. I can't speak to 500 patch profiles, although that may change shortly. I have a paper where I don't want to commit to 12 letter sized sheets, and I plan to profile it with either 392 or 784 patches.
ArgyllCMS does a nice job of matching to the paper white point, and I find that skins tones come out very nice. I find them noticeably better than the CM profiles. I also see a better print to screen match with the ArgyllCMS profiles, which means I have confidence in these profiles whereas I'm now reluctant to use the CM profiles. When I compare gamut volumes my ArgyllCMS profiles come out somewhat larger.
Terry.
-
Thanks to everyone in this thread for the good info. Today I had some time to try ArgyllCMS with my Color Munki, and the results have been excellent! Especially on matt paper (Photo Rag 308) the difference between a basic Color Munki profile and a custom 2264 patch (five full 11x17 sheet in high density mode) Argyll profile are quite dramatic! With Argyll the Munki is now all grown up!
I will be slowly reprofiling by papers with this new approach, and I feel that Argyll profiles are as good as any I have seen! For anyone willing to tacle the command line, this is an outstanding combination of hardware and software!
Stefano
-
Here is a workflow I found to be pretty robust with ArgyllCMS for RGB printers:
1. printer calibration as described http://argyllcms.com/doc/Scenarios.html#PC1 but I use -s256 for targen
2. do a simple preliminary profile programming targen to -d2 -G -s256 -g256 -f1860
3. take Munsell values from http://www.cis.rit.edu/research/mcsl2/online/munsell_data/1929.dat and convert them to RGB values using the profile above,
Refering to point 3 of your workflow do you think is it worth converting the Munsell Lab values to Bruce Lindbloom's CIELAB UP profile (a uniform perceptual profile) to avoid the classical "blue turns purple" problem and only after that converting them to RGB values using the Argyll profile created in point 2?
For those interested in the "blue turns purple" problem read http://www.brucelindbloom.com/UPLab.html (http://www.brucelindbloom.com/UPLab.html)
-
Refering to point 3 of your workflow do you think is it worth converting the Munsell Lab values to Bruce Lindbloom's CIELAB UP profile (a uniform perceptual profile) to avoid the classical "blue turns purple" problem and only after that converting them to RGB values using the Argyll profile created in point 2?
IMHO, the blue to purple shft in Lab is related to luminance differences in that color space. The a/b channels are not fully chrometrically decoupled from luminance in that colorspace.
Cheers,
Bart
-
Refering to point 3 of your workflow do you think is it worth converting the Munsell Lab values to Bruce Lindbloom's CIELAB UP profile (a uniform perceptual profile) to avoid the classical "blue turns purple" problem and only after that converting them to RGB values using the Argyll profile created in point 2?
ArgyllCMS uses CIECAM02 instead of CIELAB for profile calculation, I suppose.
-
ArgyllCMS uses CIECAM02 instead of CIELAB for profile calculation, I suppose.
Thank's for ponting it out.
In CIECAM02 space there is no "blue turns purple" problem.
Since Argyll is using CIECAM02 as default space I guess one can skip the conversion to CIELAB UP profile and apply the previously created Argyll profile to get RGB values.
-
Here is a workflow I found to be pretty robust with ArgyllCMS for RGB printers:
1. printer calibration as described http://argyllcms.com/doc/Scenarios.html#PC1 but I use -s256 for targen
2. do a simple preliminary profile programming targen to -d2 -G -s256 -g256 -f1860
........
7. take the target from 5. to ColorPort, add Munsell collection, add desaturated patches, print and measure with ColorPort
8. use txt2ti3 and proceed with Argyll
Hi Iliah. I was wondering about the step function (-s256) you used here. I read the Argyll description and saw this
The steps are evenly spaced in device space by default, and the total number of test patches will be the number of colorants times the value specified with the -s flag.
I don't understand how that translates into why you selected 1860 patches. It implies there is a relationship between -s and -f but I don't see it.