Pages: [1]   Go Down

Author Topic: Lab value math question.  (Read 3384 times)

Some Guy

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 729
Lab value math question.
« on: June 15, 2014, 02:08:05 pm »

Warm tone paper has a max. white of L=85.  Best black is showing L=5.

If one puts prints a 21 gray step-wedge onto the paper, then am I correct to assume that the range is 80/21 (or an L=4) for each steps Lab value?  This is for B&W only.  Maybe Gamma Gray 2.20 saved as well.

Follows this pattern:

Step 1 L=85 (Paper white base)
Step 2 L=81
Step 3 L=77
Step 4 L=73
.
.
Step 19 L=13
Step 20 L=9
Step 21 L=5  (Best black)

Was trying to use a Stouffer R2110 21 step-wedge for comparison, but the two paper bases (Glossy vs. matte warm tone) are throwing me off.

SG
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Lab value math question.
« Reply #1 on: June 15, 2014, 02:32:18 pm »

Warm tone paper has a max. white of L=85.  Best black is showing L=5.

If one puts prints a 21 gray step-wedge onto the paper, then am I correct to assume that the range is 80/21 (or an L=4) for each steps Lab value?  This is for B&W only.  Maybe Gamma Gray 2.20 saved as well.

Hi,

You may want to give this calculator a try. You can convert from (calibrated) Stouffer Densities to Lab if you want to match (although your paper base will never be as bright), or you can just make an even step delta in 'Y' or in 'L*' between minimum and maximum, depending on how you want to use it.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

Some Guy

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 729
Re: Lab value math question.
« Reply #2 on: June 15, 2014, 05:56:33 pm »

...
You may want to give this calculator a try.....


Thanks Bart, but I cannot get the calulator to open on my end without Windows throwing up all sorts of Java warnings.  Seems the site's Java is too old and mine is the newest version so all I get are security issues and warnings from both Firefox and Internet Explorer both.

SG
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Lab value math question.
« Reply #3 on: June 15, 2014, 06:54:02 pm »

Thanks Bart, but I cannot get the calulator to open on my end without Windows throwing up all sorts of Java warnings.  Seems the site's Java is too old and mine is the newest version so all I get are security issues and warnings from both Firefox and Internet Explorer both.

Bruce Lindbloom's Java applets are safe. I also get those warnings, but have never experienced anything going wrong after allowing them to run. Alternatively you calculate the values yourself, using the formulas on his site.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

Some Guy

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 729
Re: Lab value math question.
« Reply #4 on: June 15, 2014, 08:01:00 pm »

Okay, I'll revisit later.

Discovered - for some odd reason - in scanning the Stouffer step-wedge that my i1 PhotoPro 2 head differs in readings in the whites.

If I "Scan" the step-wedge (sweep), the L values are 8-9 points lower in the whites than if I "Spot" read them into the iProfiler software?  I was using the M0 scale, but somehow they differ a lot between the "Scan" and "Spot" reading modes.  ???  The Colorport seems to agree with the Spot reading of the iProfiler software as well, just the Scan mode is way off base.

Oddly, the blacks are within 0.2 of the L values between the two methods, but the whites differ from 86 to 94 in the whites (Spot mode is the higher L=94 one.).

I'm guessing I should use the Spot mode and not the Scan mode for the step-wedge.

SG
Logged

MHMG

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1285
Re: Lab value math question.
« Reply #5 on: June 15, 2014, 08:07:26 pm »

Warm tone paper has a max. white of L=85.  Best black is showing L=5.


A measured L= 85 for media whitepoint is too low for just about any paper except perhaps the "metallic RC" media. Lmin = 5 is reasonable for pigment based inks on glossy or luster papers. Some dye-based systems get get down to 3 or even as low as 1 on glossy media(but L*= 1 is short-lived because rereading the Lmin patch after a few weeks or even just a little dirt/oil/grease buildup on the print will push that L-1 one back up to 3 or higher.  You may want to check your instrument's calibration given that typical media whitepoints are 93-98 for Lmax. As for where the other printed patches of a 21 patch gray scale step wedge will end up (assuming the scanned step wedge or synthetically generated step wedge is indeed linear on L* increments), it all depends on how linear  the printer is laying down ink from white to max black. The math of [(Lmax-Lmin)/number of patches] for calculating incremental L* differences between patches only holds up with perfectly linearized print output, and that situation is rarely the case.

cheers,
Mark
http://www.aardenburg-imaging.com
Logged

samueljohnchia

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 498
Re: Lab value math question.
« Reply #6 on: June 15, 2014, 10:23:28 pm »

Discovered - for some odd reason - in scanning the Stouffer step-wedge that my i1 PhotoPro 2 head differs in readings in the whites.

Yes, I discovered this bug and reported it to X-rite months back. They finally released V1.5.4 which solves measurement issues with the i1 Pro in scan mode after I passed on my experiment findings to them. Just upgrade to the latest software.
Logged

Some Guy

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 729
Re: Lab value math question.
« Reply #7 on: June 16, 2014, 12:08:05 am »

Yes, I discovered this bug and reported it to X-rite months back. They finally released V1.5.4 which solves measurement issues with the i1 Pro in scan mode after I passed on my experiment findings to them. Just upgrade to the latest software.

I have version 1.5.4 already.  Passed their newest test software too.  Does seem like a bug of some sort.  Blacks are fine whether scanned or spot (within 0.2 or so), but the whites are of by 8 points or so.  I thought the paper was 86 but turns out in Spot it is 94.  Big difference.

So "Spot" it is for a while.  Seems more consistent and repeatable too in readings.

I may report it to them as well.

SG
Logged

samueljohnchia

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 498
Re: Lab value math question
« Reply #8 on: June 16, 2014, 01:15:00 am »

Could be a problem with corrupted XRD files or two XRD services trying to run concurrently. I had that issue too. The older version must be disabled. If it doesn't work, good luck contacting them. It took me forever and I only learned of the update from this forum and via the software reminder. Very serious issue since you cannot expect to measure in patch mode for printer profiling targets.

I've recently reformatted my computer and running V1.3.2 with the older XRD for measurements. That one tested to be fine for me.
Logged

Some Guy

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 729
Re: Lab value math question.
« Reply #9 on: June 16, 2014, 10:58:16 am »

Thanks Samuel.

I put in a case with x-rite this AM and already heard back.  Sent in the measurement files, and also a JPG of the scanned target (Below).

I'll also install the newest XRD (Verison 2.3.4) off their site too.  Think I already did, but I'll do the uninstall and reinstall to be safe.

Could be the target file.  I see a lot of differing ones where some have black bars between the target patches, white bars, a black leader bar or not, etc.  Keith Cooper has variations of the targets too, just I can't get them to be accepted correctly with Colorport.

This is the one I was using that gives me the difference between the white between Scan and Spot (By 8 points too!).  It's 5 points of L between each patch in Gamma Gray 2.20 in the TIFF I have.  iProfiler just doesn't agree between "Spot" and "Scan" modes with it on the white end, but black end is okay.

SG
Logged

samueljohnchia

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 498
Re: Lab value math question.
« Reply #10 on: June 16, 2014, 11:52:42 pm »

You're welcome.

I just managed to get in touch with local support and they were very nice and are willing to help me escalate my issues. Great!

Ah ok, I've just tested it again with V1.5.4 and latest XRD. It indeed solves the issue of problematic light/white patch measurements when adjacent patches are of significantly varying contrast. This is how the smart patch generator lays out targets anyway.

When they are monotonic (steadily increasing or decreasing in tonal value from left to right) like in the jpeg you uploaded, the problem is still there. Light/White patches measure significantly darker than they should be. I'm 7 L* points lower when scanning a printer target on Harman Gloss paper.

Let's see where this goes with X-rite.
« Last Edit: June 17, 2014, 12:06:31 am by samueljohnchia »
Logged
Pages: [1]   Go Up