Media Configuration Tool (MCT) for Canon

[henrikolsen]

Starting a new thread on this topic, spawned from this one: http://forum.luminous-landscape.com/index.php?topic=123444.msg1030968#msg1030968.

--

I might chime in with some info on MCT usage (on a Pro-1000), as I've recently spent a good deal of time testing out things in it that I cannot find properly documented.

A couple of starting points, some might be mentioned earlier, some might be new to some, some might be qualified speculation (missing hard facts from Canon).

Create calibration issues a feed adjustment, whether done or not in step 1. The reported values do _not_ agree, and the bug is reported to Canon (both the extra feed adjustment initiation, and the - much worse - discrepancy between the two runs). It _might_ be that the value reported in create calibration target is a delta to the one in step 1 (often low value for me), but not sure. Rerunning step 1 adjustments is fairly consistent. If different value from create calibration run is overwriting the one from step 1 is unknown. Canon is provided with am1x exported media files for examination but haven't returned with results. Please do your own bug reports to Canon if you replicate this issue to force some more pressure on the trouble solving.

Create calibration target does just that . It just creates a target, but only the ability to later calibrate against it is registered, not an actual calibration. This might be (another) bug in MCT. As I recall from earlier tests some linearisation does change from this create calibration target, though calibration isn't reported until executed separately after target creation. I can dig out more data if needed. I've stopped using the calibration now, as changes seemed minor to end result of custom calibration, and since the extra feed adjustment it triggers seem buggy, and Canon has yet to come up with an explanation of the discrepancies.

For _some_ base media types the named quality scale if offset compared to built in media types, e.g. highest mean quality 2 for some media, but maps to quality 3 for others (in that case quality 2 can be set custom, but not as named setting). This is quite confusing. Quality range is 2-5, 2 being highest (see accounting manager).

Photo/glossy media cannot have a custom media type with the auto CO application as Pro Platinum which puts in from highlights to darks. The base media types for custom media seem to only apply to darker tones (as Glossy II). This has stopped me from using custom media for glossy papers, unfortunately - missing out on feed adjustments, named medias in selection and accounting etc. Overall/full setting still works, but often I just want auto, but with Pro Platinum CO curve, not the one excluding highlights where I think they matter a lot (perhaps more than shadows).

The three fine art base media type seem to render colors the same, and only differ in their default settings of ink limits and head height (and possible exclusion of rear feed for thicker papers). So the multiple options are just presets. Set back to same ink limit and head heights, they print the same. Also the Extra Heavy Art Paper base reported in MCT to use only MBK, is using a mix of PBK and MBK like the two others.

Only some base media types allow for use with "Contrast reproduction" in Print Studio Pro.

I'm sure there's more, but that's it for this time. Tons of time and testing for reaching some of this insight. All backed up by a lot of measurements and notes. Unfortunate, as I'd wished for all this to be clearly documented and resolved by Canon.

Please join with more knowledge and data.

[loganross]

I don't have any issues with custom media types, but then again, I have custom profiles made and do not need the color calibration portion of MCT. Typically, the MCT assistant selects heavyphotopaper or lightphotopaper as the suggested base. For some really thick papers, it has recommended canvas. 

My workflow is to enter in the paper thickness into MCT, accept whatever it recommends as a base profile, perform a feed adjustment, perform the ink test and select an ink level, and then save the newly created media type.  I do not print the color targets as I select my preferred color icc profile in my printing application.

[Panagiotis]

About the importance of color calibration on the PRO series printers.

http://www.learn.usa.canon.com/resources/articles/2016/ipf-pro1000/color-calibration-tool.shtml

"Calibration is a process that ensures that your paper is receiving the exact amounts of ink specified by the driver to create your desired colors. Ink output can change because of small variations in ink and paper production, as well as shifts in humidity and ambient temperature in your work area. It is usually a good practice to calibrate your media whenever you open a new package of paper, replace an ink cartridge, or at least once a month."

There is also the Device management console app which monitors the calibration status for every paper custom or not (that is papers with common or unique calibration) and suggest the appropriate time to renew the color calibration. More here:
http://www.learn.usa.canon.com/resources/articles/2017/printer-device-management.shtml

From the above link:

"The Calibration Advisability bar below the status window indicates whether a new calibration is recommended at this time. If the bar is green, calibration is not needed. If it is yellow, try to schedule a calibration when you can. If the bar is red, you should calibrate before printing to ensure accurate color in your prints."

EDIT: I would add that it seems logical and good practice to execute color calibration (unique if a calibration target was already created for this paper, common if a calibration target was not created) for a paper just before printing the targets to make an ICC color profile. By doing so the printer can go back to this state (just before printing the targets) anytime in the future by executing again the internal color calibration procedure.

[loganross]

I should clarify that Canon tech support informed me that the internal color calibration should rarely need to be performed.  The fact that there is the ability to perform at any time is fine, however, I will only perform when I think it is necessary.  I am quite happy with the results obtained by (1) having a proper editing/print prep workflow, (2) using a custom color icc profile, and (3) defining a custom Media type. 

Note that with respect to # 3, as I mentioned in my other thread, I am not sure that I can "see" a material difference yet.  Moreover, there were two key drivers for me in setting up custom Media Types:

(a) exact any additional quality benefit
(b) since I use a lot of third party papers, to make my life easier when it comes to remembering the correct corresponding media type to use

Note that (b) is as important to me as (a).

 

[henrikolsen]

I just stick to the common calibration, in my case on a Pro Luster.

This then can be used to maintaine the behaviour of the printer, and is the reference for custom icc calibrations, which can then also be shared among other similar printers.

A little data point for differences before and after a calibration, looking at a set of 96 sample patches covering the gamut well.

  Total errors:     peak = 9.612874, avg = 3.256066
  Worst 10% errors: peak = 9.612874, avg = 7.788206
  Best  90% errors: peak = 6.402975, avg = 2.729073

A significant change.

Also checked against Canons own icc for Pro Luster and that matches better after calibration (as expected), coming in at (CIEDE2000 this time)

  Total errors (CIEDE2000):     peak = 2.177756, avg = 0.831824
  Worst 10% errors (CIEDE2000): peak = 2.177756, avg = 1.779103
  Best  90% errors (CIEDE2000): peak = 1.449286, avg = 0.721675

Before it was

  Total errors (CIEDE2000):     peak = 4.667316, avg = 1.825594
  Worst 10% errors (CIEDE2000): peak = 4.667316, avg = 3.015105
  Best  90% errors (CIEDE2000): peak = 2.493174, avg = 1.687279

[Panagiotis]

Thanks! Very useful because it shows with actual measurements the importance of color calibration. Can you provide more info of the time passed and the printer usage between the two measurements?

I just stick to the common calibration, in my case on a Pro Luster.

This then can be used to maintaine the behaviour of the printer, and is the reference for custom icc calibrations, which can then also be shared among other similar printers.

A little data point for differences before and after a calibration, looking at a set of 96 sample patches covering the gamut well.

  Total errors:     peak = 9.612874, avg = 3.256066
  Worst 10% errors: peak = 9.612874, avg = 7.788206
  Best  90% errors: peak = 6.402975, avg = 2.729073

A significant change.

Also checked against Canons own icc for Pro Luster and that matches better after calibration (as expected), coming in at (CIEDE2000 this time)

  Total errors (CIEDE2000):     peak = 2.177756, avg = 0.831824
  Worst 10% errors (CIEDE2000): peak = 2.177756, avg = 1.779103
  Best  90% errors (CIEDE2000): peak = 1.449286, avg = 0.721675

Before it was

  Total errors (CIEDE2000):     peak = 4.667316, avg = 1.825594
  Worst 10% errors (CIEDE2000): peak = 4.667316, avg = 3.015105
  Best  90% errors (CIEDE2000): peak = 2.493174, avg = 1.687279

Custom calibration

[loganross]

I agree. I should have mentioned that yesterday. I rely on the common one which was performed when I got the printer.  You have made me curious enough that perhaps I will try unique calibrations from my matte papers.

I just stick to the common calibration, in my case on a Pro Luster.

This then can be used to maintaine the behaviour of the printer, and is the reference for custom icc calibrations, which can then also be shared among other similar printers.

A little data point for differences before and after a calibration, looking at a set of 96 sample patches covering the gamut well.

  Total errors:     peak = 9.612874, avg = 3.256066
  Worst 10% errors: peak = 9.612874, avg = 7.788206
  Best  90% errors: peak = 6.402975, avg = 2.729073

A significant change.

Also checked against Canons own icc for Pro Luster and that matches better after calibration (as expected), coming in at (CIEDE2000 this time)

  Total errors (CIEDE2000):     peak = 2.177756, avg = 0.831824
  Worst 10% errors (CIEDE2000): peak = 2.177756, avg = 1.779103
  Best  90% errors (CIEDE2000): peak = 1.449286, avg = 0.721675

Before it was

  Total errors (CIEDE2000):     peak = 4.667316, avg = 1.825594
  Worst 10% errors (CIEDE2000): peak = 4.667316, avg = 3.015105
  Best  90% errors (CIEDE2000): peak = 2.493174, avg = 1.687279

Custom calibration

[henrikolsen]

Thanks! Very useful because it shows with actual measurements the importance of color calibration. Can you provide more info of the time passed and the printer usage between the two measurements?

The example shows differences between an uncalibrated and calibrated state (measured the same day). Not between two calibrated states with a long time in-between. I might do a new measurement before/after doing a future recalibration to get some numbers on any possible drift. I expect it to be low, but data is king, so might as well know it instead of guessing or assuming.

[loganross]

To your earlier point, is there a way to compare the calibrated state from one paper to another?

The example shows differences between an uncalibrated and calibrated state (measured the same day). Not between two calibrated states with a long time in-between. I might do a new measurement before/after doing a future recalibration to get some numbers on any possible drift. I expect it to be low, but data is king, so might as well know it instead of guessing or assuming.

[Panagiotis]

The example shows differences between an uncalibrated and calibrated state (measured the same day). Not between two calibrated states with a long time in-between. I might do a new measurement before/after doing a future recalibration to get some numbers on any possible drift. I expect it to be low, but data is king, so might as well know it instead of guessing or assuming.

I understand now. Probably that was the reason there was such a significance difference. Unfortunately I cannot contribute measurements since I don't have an instrument. Any additional info would be greatly appreciated. Thank you.

[loganross]

I was wondering with regards to ink setting, does it relate to the maximum amount of ink that the paper can handle (i.e., more ink can be used when necessary to increase gamut), or does it refer to the general amount of ink used for a paper? I ask, because if it is a case of the former, I would prefer to use "Largest" so long as the paper can handle it. However, if it is the latter, I would want to avoid "Largest" so not to waste ink unnecessarily.

[loganross]

Answer found in documentation. It refers to max Ink the paper can handle.  Most can handle largest.

I was wondering with regards to ink setting, does it relate to the maximum amount of ink that the paper can handle (i.e., more ink can be used when necessary to increase gamut), or does it refer to the general amount of ink used for a paper? I ask, because if it is a case of the former, I would prefer to use "Largest" so long as the paper can handle it. However, if it is the latter, I would want to avoid "Largest" so not to waste ink unnecessarily.

[loganross]

Jut got my new icc profiles back after performing unique color calibrations for my preferred papers (this required me to reprint targets).  In colorthink, it shows increased gamut for a few of the papers.  It also shows more linear neutral curves for some papers (which perhaps means more neutral b&w mode prints when not using icc profiles).  I am not sure whether the differences will be appreciable in actual prints, but I will post back.