Profiling and such - some observations

[Ink stained Fingers]

I overprinted profile target sheets with a Gloss optimizer/Epson - experimentally - to check for an assumed impact onto the fade performance of dye inks - yes - it does. But there is another question remaining - how does a GO overprint impacts the overall gamut of such practice.
I'm using I1Profiler with an i1iO scan table for profile generation and tpyically use a sheet with 1100 patches on one sheet, that scans within a few minutes.
When looking to the gamut with or w/o a GO overprint I get this at L=25%

The outer red line shows the gamut of a typical dye ink on a glossy paper, in this case a HP premium glossy photo paper, this sheet got overprinted with GO on a R800,
and the green line shows the effect onto the gamut volume, it is slightly reduced, about equally across all luminance values, the total voume is reduced by about 8%. The black point is this case raises from L=5% to L=6%. The look of such print is slightly different to a print without GO, but only visible in direct comparison, this slightly reduced gamut volume is not visible in actual printouts, it only affects the saturated colors differing slightly. The results are very similar with other dye inks, and on a range of other papers, the GO might improve the fading performance of the inks but slightly reduces the available gamut, and the overall look is a matter of personal taste.
I'll soon post a similar look to pigment inks with/w/o GO.

 

[Ink stained Fingers]

I had a similar look to profiles with pigment inks - on a R800 first without GO, scanning for the profile patch data, overprinting with GO and scanning the patches again

The left image shows part of the gamuts, the bottom section, the print without GO results in the red gamut volume, the green shaded gamut body comes from the GO overprint, the GO overprint increases the gamut by about 15% everywhere from light to dark colors. more at the darker colors . The gamut crosssection - right image - is done at L=25 and shows the improvement. The black point is at L=13 without GO, and drops to L=7 with GO which is visible when you compare printouts.
I'm getting about very similar results with different pigment inks and a variety of photo glossy papers including silk/semimatt. And I went a step further, I was printing with pigment inks on a R265 and did the GO overprint on a R800 - a similar gamut increase is visbible, and this as well with genuine Epson R3000 pigment inks.
It is a kind of irritating that Epson is not offering a Gloss optimizer option for the mainstream printers
as Canon does with the Chroma optimizer. It's just the P400 with this option. This test is not related to the other fading tests, I wouldn't have the patience to wait for the results.....

 

[Ink stained Fingers]

The photo print problems with Epson's 4 color Durabrite printers have already been discussed and a kind of work around to use a photo black pigment ink instead of a matte ink and using the matte paper settings instead of a glossy paper setting.
I have profiled such a combination with this result - as print output from the recently installed WF7110W A3 printer:

That's the bottom part of the gamut volume at L=8, the pink line shows the gamut with the inks mixing to 'black' - actually a dark and strong violet, that's with the glossy paper setting.
The blue line sits at a neutral black point, that's the black point of the photo black with the matte paper setting. This makes a difference directly visible in printouts.

 

[Ink stained Fingers]

when you want to profile a particular paper with your printer you mostly have several choices for the driver settings, you turn the ICM option to off , but you have several choices for the paper selection -e.g. photo glossy, premium glossy, ultra glossy which refer to the names of the OEM papers for Epson (or similar for Canon). You don't know which one to select, the manufacturer may give you some advise - 'use our photo paper with the premium glossy settings' - or not. You have to try , some settings don't appear to make a difference in printouts or you may get some ink overload on the paper. And within each paper choice you typically have a few quality options - photo, optimal photo , optimal photo RPM (by Epson). These options slow down printing, some pretty much, and you may get again some ink overload or not. The paper selection seems to control the overall ink limit the printer can place onto the paper, the quality options seem to use some of most of that range. The quality option controls the number of overprints per line, with different droplet sizes - more print passes with smaller droplets. A real quality improvement - more fine details - is barely visible or not at all.
But what is the impact onto the gamut you can acheive with these options

That's a gamut cross-section at L=18, in the darker region, it is a castcoated budget paper, the profile patch sheet printed on a R800 with pigment inks, with the Premium Glossy Photo Paper setting, and with the Photo - yellow line - and Opt. Photo RPM settings - light blue line.
The gamut for the higher quality setting is smaller, and this not just at this level, but about everywhere below L=40 down to the black point.
There is another effect playing here, the paper is taking the higher ink level with the higher quality setting, but the color saturation drops. there is a point of saturation reversal that more ink does not give more color saturation but less. But this effect is not predictable to say it always would happen with the RPM option (with Epson printers). Typically darker colors like the darker violet need more ink to be mixed than for lighter colors. If you want to maximize the gamut size of your profiles you need to test several driver options and compare the profiles.

 

[mikling]

Good overall explanation. Getting a good profile is a lot of work as you can see. What is not shown either is that getting a good set of targets also has an effect on the overall profile. A good set of targets will vary by printer model in addition to what was observed above.

Then, the printer engine kicks in....how the inks are combined and whether or not the printer is allowed to output said color in quantities needed even before reaching the ink limits. In other words, the printer also has a kind of ink limit to its output that is not paper related but determined by firmware.

Essentially profiling also has limits as to what it can achieve.

 

[Ink stained Fingers]

Yes,

Then, the printer engine kicks in...

the driver/printer behaviour is a kind of black box we have to make the best use off - experimentally - since no details are available, only companies writing RIP software would get more information on a pretty confidential level, and since there is no RIP software for typical desktops there is no other way than by actually testing the options.
Getting a working profile is straight forward - step 1, 2, 3, but getting a profile with the most gamut, the best black point etc requires quite some of this testing and working with the parameters - and there are lots of them - paper, inks, driver settings - so there is no way to answer the question which printer offers the widest gamut

 

[The Hat]

Try a 3D printer, I have not printed a single sheet since I got mine, the only thing I use is a Pritt stick...

 

[Ink stained Fingers]

let me give you another example how driver options can make a difference. I'm using a WF2010W, a simple 4 color Durabrite printer, ink swapping with different cartridge sets is easily possible with this model - e.g. changing from pigment to dye inks and or swapping the matte black pigment ink to a photo black. I'm using a 4 color CMYK dye inkset for this test.
When using the typcial ICM off settings together with Ultra Glossy paper selection, it prints very dark, highly saturated. When printing the target sheet with the matte paper setting, ICM off it gets visibly lighter overall

One can start questioning whether such a print would give a decent profile at all, this lightness difference is a parameter the driver is applying to the printout via the paper selection.
This is the view of the profiles generated from both these scanned patch seets

The pink outline belongs to the Ultraglossy setting, the overall gamut is about 20% smaller than
the gamut with the matte paper setting - the green line and range at the bottom shows the extended gamut.
This WF2010W gamut is slightly smaller than the gamut of this same paper used on a R265. The R265 delivers a wider gamut in the lighter green range, most likely as an effect of the available light inks.
There is no significant difference in the lower lightness range so the WF2010W can be used as quite a good photo printer.

 

[Ink stained Fingers]

There is a question lingering around and being discussed heavily since a long time - how many color spots do you need on your target to get a decent profile - 50 or 480 or 3000 or ? Or how much would you give up using just a small number. I'm typically using 1107 spots - they just fit onto one page . the maxium the i1Profiler software let me do for a i1Pro Spectro. I actually don't really bother whether it's 800 or 1000 or ... since I'm using an i1iO robot desk for scanning I'm not feeling the nuisance doing manual scans. But what about the low end - I tested a target with 50 color spots - the minimum i1Profiler let me do, these colors are placed around the sRGB color space - with some regularity - all corners are populated, there are one or two color spots on every egde , there are one or two on every outer surface and the rest somewhere within the volume.

(I remember looking to the 50 color spots ColorMunki is using, that's about the same distribution)
With 1100 spots you have many more available, there are about 8 on every edge, and many more in a matrix on the boundary surfaces and within the gamut volume.
Running profiles with both targets yields a very good overlap of both gamut volumes, pretty much identical everywhere except at the bottom, the more critical area. The 50'er set just has one single control point at the black point and one more in direction of the -b axis, the next one is already at the corner.

The red body is the gamut created with 50 patch fields, the green body with 1100 patch fields.
The bottom tip at the black point has the same value, the green volume is slightly smaller , based on more control points and measured data, the math creating the profile tables just has to assume some standard behaviour between the black point and the only actual control point, the point where both volume shapes intersect.
I'm not showing the upper parts of the gamut volumes , they are very close together by a few %, they are due to printing and reading variations creating probably most of the small differences.
There is a last question whether these profile variations are visible in printouts , I tried a few prints with some test images with darker content, I don't see any. It might be possible to specifically create some special pattern in this affected color range - between the red and green gamut outlines - that would confirm the measurements but not help otherwise very much.
These tests are not done with a photo printer with whatever implied better performance but with the Epson WF2010W running with dye inks for this test and the matte paper setting to get the printer to use the black ink, and on a budget castcoated glossy paper.

 

[Ink stained Fingers]

We hear the question - which printer is the best - which printer delivers the most gamut - and similar including references to Adobe RGB etc.
'Which printer delivers the most gamut ' - we all know there is an intricate interrelation between the inks, the paper, the printer model and its driver settings.
Let me compare the L800, an Epson photo printer with ink tanks, and the base line WF2010W office printer normally running on Durabrite inks. I'm running the WF2010W for this test with the same Fujifilm DL inks as the L800 by use of refill cartridges, this gets the ink as a variable out of the way. And I'm using a cast coated glossy photo paper.

These are the gamuts at a luminance of L=22%, in the darker range.
Both printers perform pretty much equal at L=50, with a few % variation, but there are 3 lines in the graph:

Green boundary - that's the WF2010W with the matte standard driver setting
light blue - that's the L800 with the matte Photo driver setting
pink - that's the L800 with the Ultraglossy Photo driver setting

This directly shows the dilemma when creating a profile - you need to test various settings to find the optimal one, the Ultraglossy photo setting of the L800 does not deliver you the best gamut you can acheive with this printer. The matte paper setting yields a slightly wider gamut which is measurable but not directly visible in printouts.
The surprise is the performance of the WF2010W - delivering a significantly wider gamut, and this is visible in printouts in darker areas, with more details in the shadows - when you specifically look for it.
So which printer is the 'best' in this case ?