You don't need much patches to create a good printer profile

[pharmacist]

Inspired by @Ink stained Fingers I generated a small 165 patch target that can be printed on a small 10x15 cm/4x6 inch photo card and with some modifying the grey and near neutral patches in the patch editor in i1Profiler to obtain an optimum patch distribution in the RGB color space I could generate a very good printer profile with only 165 patches:

Above a printer screen in gamut view (i1Profiler) between the 165 patch target (blue) and the 581 patch target (orange). They are practically the same in volume size and almost identical.

Actually the target is even better than the 50 + 50 method of the Colormunki: the ccStudion/i1Studio method uses a interpolation method between 2 points in the color space not taking care about the intermediate points if the behaviour is not linear (curved behaviour), which gives some problem like in the example below between the 165 patch target (i1profiler but I generated a v2 version so iccview.de could handle it) and the Colormunki profile:

You can see clearly the wired cube in the green/blue/yellow area of the color cube (165 patch target with i1profiler) can cover a significantl area that is not covered by the 50+50 Colormunki profile (the colored cube). I had this problem in the past on my P800 using a chinese pigment refill ink with a particular light cyan version that was way too light causing a strange gradient in the blue skies, that I could only deal with using the ArgyllCMS method (480 patches), before I could hold on the i1Pro2 spectro.. Note this profile is for the ET-8550 using original ink, so if you use the Colormunki method to create the profile: please take: probably you need an optimization scan with some green/blue colors to improve that area. The simple 165 patch i1Profiler profile does immediately a good job and is on pair with a larger 581 patch target.

You can download those 2 sets of files for the i1Profiler (i1pro2 and i1iO version included)

 

[Ink stained Fingers]

The simple 165 patch i1Profiler profile does immediately a good job and is on pair with a larger 581 patch target.

I only can confirm @pharmacist 's findings above with a test I did on his previous patch sheet version with 154 patches - this shows that a pretty good profile is possible with such a low patch count.

 

[pharmacist]

I was wondering if this can be done too with ArgyllCMS. I will make a target for ArgyllCMS also using 165 patches on 10x15 cm photo card.

 

[Ink stained Fingers]

I was wondering if this can be done too with ArgyllCMS

You may try to compare prints done either with an ArgyllCMS profile or an i1Profiler or ColorMunki or .... profile generator, but be aware that every company has some room for variations how rendering intents are implemented - it would be the perceptuakl or the relative colorimetric mode affected mainly. So differences in print output will most likely appear but you cannot judge which version is better or more correct , just look how many adjustment options there are for profile generation with the i1Profiler package.
But you may try to do such comparison with ArgyllCMS profiles gnerated from a low or high patch count and print the same patch sheet with these different profiles and scan those prints. I don't know which math tools you may have to compare that data e.g. in a spreadsheet.

Give it a try to fit the 165 patches onto a 10x15 cm card for ArgyllCMS

 

[Ink stained Fingers]

I did another test and compared a print done with a profile with 165 color patches vs. a print with a profile with 2880 patches, printed on a ET-8550 with 106 Epson inks and the Ultragrossy standard driver setting , with rel col. rendering intent and BPC off. The test print covers all colors - in- and out of gamut.
The averaged DeltaE deviations across 450 patches on the test sheet just comes to DeltaE= 1.18. I would say that prints with either one of such high or low count profiles deliver the same print output.

I did this test with a glossy cast coated Acction 210gr budget paper, just a plain average paper; I did the same print comparison test on another paper - the glossy 300gr Netbit paper which just delivers a deltaE average of 1.58.

All this testing shows that a high patch count during profile generation is not necessary - already 165 patches - fitting on a 10x15 cm card - can deliver a very accurate profile.

It may be possible to get more info out of the data tables generated by the profiling software i1Profiler, but I don't see a need for me at this time. The recently mentioned Babelcolor patch tool may be of help with the data tables, but I'm not going to delve into that program at tlhis time since I don't want to doctor around in the profiles.

 

[pharmacist]

Many thanks for the comparison: Actually the Spyderprint High Quality target with only 225 patches seems to produce already very good profiles. Like my 165 patch target both targets are based on RGB 5x5x5=125 patches in the RGB color cube space and extra patches up to 165 or 225 patches. The 225 patch target can be printed on a single 10x15 cm photo card, but rather cramped together, so scanning is a bit fidling, but it can be done. The 165 patch target can be scanned very easily without problems, also with the i1iO device and is much easier to handle and to be scanned. I had to manually change the grey step ramp and the near neutral ramps in order to get an even distribution and add additional grey/neutral steps around the neutral axis to get a better sampling around this visually most important part of the profile to obtain a good and visually neutral looking profile. The result is a target that has 2 extra grey steps and 3 triplets of near neutral RGB steps above the initial 154 target generated by i1profiler filling up the limited surface of a 10x15 cm photo card while keeping the patch size at 8x8 mm. With some PS manipulation the whole target was fitted on a 10x15 photo card at 360 dpi resolution.

@Ink stained Fingers : how is your findings compared to your older 96 patch target, does it a better job or is the difference only marginally better ?

 

[Ink stained Fingers]

the test with a 096 patch profile delivered a deltaE of 1.4 vs. the high patch count - 2880 profile which is slightly wider
than the DeltaE of 1.18 with 165 patches , the increase of patches from 96 to 165 patches for profiling delivers a measurable improvement reducing the deltaE

https://www.printerknowledge.com/th...e-needed-for-a-good-profile.16935/post-144751

So the profile based on 165 patches delivers good results and is easy to use since it fits onto a small 10x15 cm sheet

 

[The Hat]

So the profile based on 165 patches delivers good results and is easy to use since it fits onto a small 10x15 cm sheet

So when it comes to profiles, smaller is better, or as good as !