Topics

monitor profiling: i1Display Pro vs i1Photo Pro 2?

 

I have an i1Photo Pro 2 (a $1600 spectrophotometer) which I use mainly for making paper profiles, but which also is used to profile my wide-gamut monitor. I  have a fellow who claims to know a lot about this (I've only been doing it 15 years) who insists that the $240 i1Display Pro (a colorimeter) "produces a visibly superior monitor profile."

I've read that the latter is better in the dark tones, and worse in the bright than the former, that's pretty much the extent of my understanding.

I'm wondering if anyone here has any experience and/or comments on this. Is the i1Display Pro actually better for monitors? (Mac OS; basiCColor software).
--
Tracy Valleau, moderator

www.valleau.art

Jeff Grant
 

'A visibly superior monitor profile' is a brave call. On what basis is the assessment being made? I have an i1Pro2 that I use with Eizo ColorNavigator. For my B&W printing, it passes all the standard tests for being able to see differentiation in the bright brights and dark darks. How much better than that does it need to be? Unless you are using an Eizo CG or the NEC equivalent, the monitor remains the weak link IMHO.

 

Thanks, Jeff. After posting my original question, I had a further exchange with this person (who shall remain nameless.) In fact, because we each have the same model monitor, I asked him to send me his monitor profile, so I could toss them both into ColorThink Pro and see them together.

They are basically identical, and where they are not, it's in the 1% range, and I could easily chalk that up to calibration differences.

Thanks for responding.  Happy 4th!

--
Tracy Valleau, moderator

www.valleau.art

 

Your profile comparison gives you a practical answer. The theoretical answer (more in the direction your unnamed friend was leaning), is based on some basic differences between a spectro and a colorimeter. Having worked in that field for many years, here’s a rough summary of the theory:

Splitting a number of photons up into many buckets, then counting what’s in each bucket, and adding them together again (as a spectro does) is not as clean a process as just measuring the whole lot in one, or a few, batches (as a colorimeter does). This effect matters most at the dark end, where a small number of hits are being used to determine a given level, and where color doesn’t matter, just luminance level. But it also requires a good colorimeter to make low noise measurements in the darks. 

On the other hand, a spectro is much better at measuring highly saturated colors (such as the primaries of a wide gamut display) than a colorimeter, which is why tables for display types or models have been added to some calibration software since more exotic, and saturated, display types have appeared. These tables are there to assist colorimeters in getting measurements on saturated colors that are closer to what high-end spectros get. 

No way of knowing (since I am not privy to all the internal coding in the software involved) if your results are so similar because the devices are accurate and consistent, or if the results have been improved via a table or algorithm to make the primaries more similar. But it looks like the results are consistent, though we can’t really be sure that means accurate, without a reference instrument to define accurate. 

The other place to compare results is in the smoothness of the darks. This is best done by filling the entire screen with an RGB image consisting of full height bands of 0,0,0, 1,1,1, 2,2,2, and so on, up to about 20,20,20. Changing between the two profiles (and their LUT tables!) should show visible differences between the two profiles (both in density and possibly neutrality), though it won’t tell you which one is “right”, since you are not displaying both on the displays they were made for. 

This is the end where the colorimeter might prove superior, by creating smoother, more accurate dark gradients, but only on the display the profile was created for. 

Once you start comparing the available devices and software packages in a lab, using laboratory-grade reference spectros, you soon become dissatisfied with the results of all of them (and, to some degree, even the reference spectros!), on an absolute, theoretical level; at which point attempting to produce the results most useful for the type of user involved becomes the practical goal. This result is somewhat different for users of vector colors (graphic designers), raster colors (photographers), or scientific colors (technicians). Thus the perceptual, saturation, and colorimetric rendering intent choices... though how those play out varies amongst developers. 

C. David Tobie

On Jul 4, 2018, at 12:46 AM, Tracy Valleau <tracy@...> wrote:

Thanks, Jeff. After posting my original question, I had a further exchange with this person (who shall remain nameless.) In fact, because we each have the same model monitor, I asked him to send me his monitor profile, so I could toss them both into ColorThink Pro and see them together.

They are basically identical, and where they are not, it's in the 1% range, and I could easily chalk that up to calibration differences.

Thanks for responding.  Happy 4th!

<ScreenShot.jpg>
--
Tracy Valleau, moderator

www.valleau.art

 

Thank you Charles for an exceptionally clear explanation. I had no idea about the actual internals of either, and learning that a spectrophotometer is basically a CCD, like a digital cameral  (which I believe I do understand) the explanation about the darks makes all the more sense.

Further, you  are right in correcting me about "accurate" - perhaps "appropriate" or "internally consistent" would be better?

We both share the same monitor, and his profile is from the colorimeter. Looking at them together and in more detail, I do notice one thing that I'm going to try to track down today:

His profile is Version 2  and mine is Version 4. Same model of monitor and same software; different devices.
his white point is 100,0,0 and mine is 100, -3, -19.  Then my blue is -18 compared to his (-126 vs -108).  I likely have my monitor dimmer than his, setting mine at 120 cd/m.

If you know what this is likely to be, or what direction to point me in, please let me know

Once again, thank you for a perfect (OK: appropriate   ;-) response.

(My spectro in "Latest Front" and his colorimeter on the right hand.)


--
Tracy Valleau, moderator

www.valleau.art

 

To followup (and add a curiosity) :

The B* axis may just be my monitor. I ran the spectrophotometer thru X-Rite's diagnostic software, and it passed with flying colors. I have used both Xrite's and BasICColor's software to profile my monitor, and both show the same -19b. I have a i1Display Pro on order (should not have sold mine when I got the i1Pro 2, eh?) and will see if that shows the same thing. The color deltas are all less than 1, and I can barely see the difference between CCIELAB L 0 and 1 in the blacks. ("Barely" is how it should be, I believe.)

So, I have a quite workable profile, as I have had for a while now. What triggered all this was the purchase of BasICColor's software, which has some nice reporting in it.

Here's the curiosity (at least for me, as I've never experienced this before) and it's the kind of thing geeks like me love.

When I created the profile and used the software to save it, as with most such profiling software, that new one becomes the current profile in use. I kept a blackpoint clipping test open in Photoshop as I was running tests to set the black level in the profile, so it was easy to see the new profile's success (or not). ("Success" being defined  as me seeing the values as I wanted to see them. If I set the black level too low, I could not distinguish between low levels, and conversely, too high meant the lows values were too easy to see.)

Here's the curiosity: I'd get a profile I liked, and saved to the library; bring photoshop forward to verify it. Happy, I quit the profiling software... and about 1-2 seconds later, the Photoshop image changed to way too dark to see either of the lower values. In short, the system quit using the new profile, and reverted to the generic  profile (which is automatically created by the OS.) Once I figured out what was happening (but not why) I tried all kinds of  things to "fix" the symptoms: replacing the generic (didn't work despite  mucking about with  the internals using  ColorThink Pro; rebooting; removing all profiles and rebuilding.)

So why was the user profile being dropped by the system, and replaced with the generic one? I started working for Apple Computer in 1978, and in all the intervening years have never seen the OS "revert" anything, even so much as a profile.

The issue seems to have been resolved by me removing the display profile from the user/Library and placing it in the root Library. My mac seems stable now. Permissions? They look fine to me.

So, I'm sharing this just in case someone out there knows what was causing it, and so we can all learn from it. Otherwise, stick this away in the memory bank, and try my solution if it happens to you...
--
Tracy Valleau, moderator

www.valleau.art