Better Dmax with Profilemaker 5 than i1Profiler

[jani80k]

Hi guys,

I know this thread is old. I just tried i1profiler on the weekend and I must say, I am surprised, how light the blacks are compared to PMP5.

I observed that setting the profile generation options to "Colorful" improves this situation somewhat but also boosts the contrast.

It would be great if someone could give me some input how to fix this problem.

I read this thread here but I cannot derive any actions to be taken that could help in solving the problem.

Is everybody just tolerating this lower DMax now? Boosting the contrast cannot be the final solution because it alters the image...

Cheers,
Jani

[samueljohnchia]

Adjusting the contrast does nothing for dmax. Also, it affects only the perceptual tables only. Either live with it or use PM5. A lot of things are wrong with i1Profiler.

[Czornyj]

i1Profiler stopped crashing blacks since v1.4 or v1.5 - what revision are you using?

[jani80k]

i1Profiler stopped crashing blacks since v1.4 or v1.5 - what revision are you using?

Really? Where did you get this information from, can you provide any links to Release Notes/Bugereports or any threads in this or other forums that confirm this. I am using 1.5.0.

[pixeldoppelganger]

Just adding a side comment that might help for those exploring all of this... 
I found the Saturation rendering intent w/ an increase of ink density on my 9900 produced some wonderful color and Dmax, however, it's smoothness
is quite harsh. This using i1Pro built profiles.

Speaking with DataColor at a Tradeshow, and via emails, here's a bit of information that might help:

"Datacolor does offer an exceptionally good Saturation Intent in our printer profiles, and a good alternative for where that is not ideal: our 'Perceptual' Intent is actually not Perceptual (meaning it does not blindly shrink the gamut), instead it is halfway between our Colorimetric Intent, and our Saturation Intent. This allows max color saturation with Sat, a step back from that with Percep, and then the most literal color with Colorimetric. Its a good system. "

I'm going to pull the trigger on their SpyderPrint, and see how this compares to i1Pro

[JRSmit]

Ja i i do not know your print environment, nor do i have pmp5. I use iproviler 1. 5 witb i1pro2 and have little issue with max black. Depends on papers but on a gloss pe photopaper around 2.5 and on a fine art mat around 1.6 to 1.7 . This with a smooth tone curve.

[digitaldog]

"Datacolor does offer an exceptionally good Saturation Intent in our printer profiles, and a good alternative for where that is not ideal: our 'Perceptual' Intent is actually not Perceptual (meaning it does not blindly shrink the gamut), instead it is halfway between our Colorimetric Intent, and our Saturation Intent. This allows max color saturation with Sat, a step back from that with Percep, and then the most literal color with Colorimetric. Its a good system. "

Sounds like a lot of marketing speak to me. There are no rules in how one creates a perceptual rendering intent. Just like there are no rules in how a film manufacturer could build a rendering into their film stocks. So the: does not blindly shrink the gamut seems a bit disingenuous, implying everyone does this incorrectly but they have some magic process. You should test this but not based on what they wrote above.

[Czornyj]

Really? Where did you get this information from, can you provide any links to Release Notes/Bugereports or any threads in this or other forums that confirm this. I am using 1.5.0.

I had horrible problems with crashed shadows on c-prints - profiles generated by earlier i1Profiler revisions had distinctly crashed blacks. With i1Profiler 1.5.0 the problem is virtually eliminated, RGB 0,0,0 is only slightly neutralised, so the resulting RGB values after conversion are usual 0 for one channel, and 1-2-3 for other two channels. It's still not perfect, but much better than before.

[Some Guy]

I found using i1 Profiler software with the i1 PhotoPro 2 head today (Windows 8.1-64), if I use Qimage as the pre-print tuning software, I can set a boost to the black part of the curve (In Qimage's "+Color" tab) to regain shadows lost or crunched up in a 21 step gray tablet.  If I don't, often it will bunch the blacks up around the 70% part of the 21 step scale and anything beyond 70% of the scale goes dark.  I saved it as a filter to apply to other normal images and it did help with regaining shadow detail.

Been monkeying with it today on a new glossy paper and it needed a good kick in the black and Profiler's Optimization didn't address it either.  Oddly, letting PhotoShop Manage Colors produced a smoother gray scale too that the one the i1 produced that bunched up the blacks.  Saw zero affect to applying Perceptual, Absoute Colormetric, Relative Colormetric, or Saturation in Qimage either from 4 prints made with each.

SG

[samueljohnchia]

however, it's smoothness is quite harsh. This using i1Pro built profiles.

Setting the saturation slider in i1Profiler to zero from the default of 40 does a wonderful thing for smoothness.

[samueljohnchia]

V1.5 renders the grayscale slightly darker towards the black point than earlier versions of i1Profiler for me too. This is by lowering the separation between the darkest tones, albeit slightly, and it does nothing to better the dmax for what I profile. My profiling experience is limited to relatively neutral native output, so I don't face any issues with dmax. Like Jan, I've been happy with the quality of the profiles after quite a bit of tweaking and investigation. V1.5 compared to earlier versions of i1Profiler has smoother gradations in dark blues - that the main tweak they have made in the profiling engine. It does sacrifice a bit of blue edge gamut separation.

But V1.5 is very buggy for me. Lost of connectivity with i1 Pro 2 at random (halfway during a large patch measurement process is not fun), increased errors in measurements and a strange bug when measuring light patches or paper white, very slow profile generation (sometimes very very slow at random) and more resource hungry than V1.3.2. Crashes without warning frequently.

[Scott Martin]

i1Profiler, like Monaco Profiler before it neutralizes the grey axis all the way tot the black point, whereas PMP leaves the black point alone. So PMP will leave the black point at RGB=0,0,0 whereas MP and i1P might set the black point to RGB=0,23,8 to try and neutralize a magenta, yellowish cast. I've suggested that i1P give us the ability to choose between the two behaviors but I think that's wishful thinking. Neutralizing the black point is probably the right approach and a colorful black point represents a failure in the calibration process prior to profiling. All said I'll take the i1P results and exceptional Perceptual rendering anyway. Perceptually it is fantastic. If you have a beef with it the correct approach is to improve the calibration prior to profiling.

[narikin]

So re-reading this old thread, what is the recommended path for getting best out of i1 profiler now - use Sat or Colorful intent? increase ink density +5/10/15% on the Epson panel? reduce or increase the smoothness slider?

It is very confusing - but I, like others here, found that i1 Profiler reduced the DMax compared to profiles made on older software. Generic profiles made by the paper manufacturer 3-5 years ago on PMP often give far better max density in all colors than i1 profiler, but do not have the smooth gradations Profiler does. You win something, you loose something - it is very smooth, with no uneven color steps, but at a real hit to max density of each 100% CMYK squares. not good. The explanations of i1 attempting to neutralise the black at cost to density make some sense, but do not explain all.

I was lucky enough to download the optimizer chart & a page grab when it was released (the blog post and link has subsequently been removed by x-rite) and that helps a bit with greyscale smoothness, but I can't notice any meaningful change to max density.   (I suspect that grey scale optimizer chart will be added back to Proilfer 2.0 as a 'feature' when it is released, and we'll all have to shell out a $500 upgrade fee to get it!)

[digitaldog]

It is very confusing -

It is and you can thank X-rite for putting in so many sliders that in some respects do little or nothing or turn around and produce visible differences. All without a lick of documentation! Best you can do is test and come up with a rough set of settings. Or move to another solution (expensive I know). Or use legacy software (the reason I spent all of $160 on eBay for an old MacBook that can even boot OS9). Or continue to put pressure on X-rite to get their act together.

[Scott Martin]

So re-reading this old thread, what is the recommended path for getting best out of i1 profiler now

Start with an excellent calibration (ink densities set optimally for chroma peaking, and perfect linearization) and you'll get perfect profiles from i1P. Calibration and profiling are two different pieces to the puzzle that must fit together to get optimal results.

Now, if you're using a RIP there are ways an analyzing the ink densities and developing optimal linearization curves externally and then using them in the RIP but we won't go go down that rabbit hole here. If you're using the driver you may need to do some experimentation to determine what the optimal settings are for a given media. Again, too much to go into here.

i1Profiler makes awesome profiles. If you're having problems with Dmax then that's a calibration problem, IMO, not a profiling problem. Don't blame the profile if it's the calibration (driver) settings that aren't right. Yes, PMP may have delivered better Dmax from a poor calibration but there are problems with PMP's approach too, namely colorful blacks. The bottom line for me is that if you're using optimal driver and or RIP settings you'll love the blacks you'll get from i1Profiler profiles, especially with Perceptual.

use Sat or Colorful intent?

Perceptual.

increase ink density +5/10/15% on the Epson panel?

Again, you may need to do some testing to determine the optimal settings for a given media. More is not necessarily better! If you have the tools, chart the chroma of the CMYK channels and look for the peaks. This is an advanced topic and a rabbit hole... If you don't have the tools, I wrote an article years ago about a beginner's approach to determining optimal media settings in a printer driver with the need for a spectro here: http://www.on-sight.com/how-to-determine-the-optimal-media-selection-for-any-paper/ It doesn't mention the ink density slider in the Epson driver because it didn't exist back then, but you get the idea and can include this slider in your testing.

reduce or increase the smoothness slider?

I'll save you the hassle of a lot of testing and encourage you to just set it to 70.

[narikin]

Start with an excellent calibration (ink densities set optimally for chroma peaking, and perfect linearization) and you'll get perfect profiles from i1P. Calibration and profiling are two different pieces to the puzzle that must fit together to get optimal results.

Thanks Scott -  I have to be honest here and declare that a profile I paid to have custom made by you ~4 years ago (before I bought my own iSiS, etc) on i1 profiler was not as good as the generic PMP one from the paper company in the DMax. Your one certainly had far better smoothness  but significantly reduced DMax. Same media settings. You remade it twice with boosted Color and Sat, but it changed nothing in the max densities reached. I used the old generic one for my prints.  I do not say this to embarrass you, far from it, I have learned a lot from your helpful posts, but simply to point out that Color Pros like yourself and Andrew here, could not get results from X-rite latest and greatest that equaled the old profiles from Monaco etc in a very very important area - maximum color density.  It is a remarkable deficiency in an expensive program.

[samueljohnchia]

Interestingly, I have managed to produce a profile with better dmax in i1Profiler than i1Match (which is basically the PMP profiling engine at default settings) with the same measurement file.

I'll save you the hassle of a lot of testing and encourage you to just set it to 70.

I have done a lot of testing too and I humbly disagree. Setting it to 70 will reduce overall gamut slightly and will also reduce the profile's black point slightly. Keeping it to 65 or less will not do either but is still effective enough to smooth any bumps in the CLUTs. If possible, just leave it at zero.

Also, I think narikin was referring to the two default choices for profile settings in i1Profiler, colorful or saturation, not the rendering intent. Neither will do anything to affect dmax or the overall gamut.

[Scott Martin]

Thanks Scott -  I have to be honest here and declare that a profile I paid to have custom made by you ~4 years ago (before I bought my own iSiS, etc) on i1 profiler was not as good as the generic PMP one from the paper company in the DMax.

What's your real world name? I'd like to look back at those profiles. If it was 4 years ago I was using Monaco Profiler back then... I'd like to know what process we're talking about...

Also, how are you determining the Dmax differences? Are you comparing color gamut renderings or taking measurements off of the prints? I would argue that the later is the only fair way to do such a comparison.

Your one certainly had far better smoothness  but significantly reduced DMax. Same media settings. You remade it twice with boosted Color and Sat, but it changed nothing in the max densities reached. I used the old generic one for my prints.  

I see quell the appropriate thing would be to have let me know at the time so we could discuss and address it. Bringing it up in public years later is poor form. Perhaps we did discuss it before - let me know your real world name and I'll take a look.

...simply to point out that Color Pros like yourself and Andrew here, could not get results from X-rite latest and greatest that equaled the old profiles from Monaco etc in a very very important area - maximum color density.

I've yet to see i1Profiler produce a profile that's inferior to Monaco Profiler. When I compare PMP, MP and i1P profiles on a variety of printing processes I rarely see what you're talking about. Would love to take a look at what you're seeing. Tell us more could you!

[digitaldog]

I have learned a lot from your helpful posts, but simply to point out that Color Pros like yourself and Andrew here, could not get results from X-rite latest and greatest that equaled the old profiles from Monaco etc in a very very important area - maximum color density.  It is a remarkable deficiency in an expensive program.

I'd agree! I recently rebuilt some profiles for a client where we had the time and energies to send the same averaged data (a huge amount of it) through various profile products including i1Profiler, PMP, MonacoProfile and Copra. We had various output technologies (CMYK) on differing substrates. We sent a lot of actual images and synthetic's though each and what we found were instances where the Copra profile was better (in some areas), i1P was better in others, PROFILER better yet in other areas. Each has certain strengths and weaknesses. You have observed one area where you are not as happy with i1P. And that's no surprise. The kinds of images you send are telling to! Some synthetic image (like a Granger Rainbow but better) tells you one thing about the color engine. The Roman 16's tell you another. In the end, you have to send a lot of such tests through the profile and look at the areas that are most important to you to select which product will be used to build the final profile.

[digitaldog]

I've yet to see i1Profiler produce a profile that's inferior to Monaco Profiler.

Based on my work, that indicates you need to send a lot more differing images, real and synthetic through both because depending on a lot of factors, MonacoPROFILER can indeed show better results.