Pages: 1 2 [3] 4   Go Down

Author Topic: Comparing spectros  (Read 18055 times)

GWGill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 608
  • Author of ArgyllCMS & ArgyllPRO ColorMeter
    • ArgyllCMS
Re: Comparing spectros
« Reply #40 on: June 19, 2017, 07:54:50 pm »

The key word was "affordable". Individually measured targets are VERY expensive.
Whilst not perfect, the charts mentioned are probably getting as close to spec. as is needed for the OP's purpose.
Hmm. Care to point to an available chart at any price that is measured to a traceable reference grade ?
BCRA/CERAM tiles etc. yes. But a chart produced using graphic arts type technology ?
- the charts mentioned so far in this thread seem to be of the type that typically use i1Pro2 grade instruments as their references, so unsuitable for checking the accuracy such instruments.
Logged

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Re: Comparing spectros
« Reply #41 on: June 19, 2017, 10:20:22 pm »

Here's the target that I was looking at initially: https://chromachecker.com/info/en/page/A2

Certainly not cheap, but more trustworthy?
« Last Edit: June 20, 2017, 12:25:32 am by Jeff-Grant »
Logged
Cheers,
 Jeff  www.jeff-grant.com

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Re: Comparing spectros
« Reply #42 on: June 20, 2017, 01:15:30 am »

I also discovered today that Colorport puts out an amount of bad spectral data. I don't know how much but on the patch that I discovered it on, it literally makes black white. I've changed to PM5 output, and the problem that I know of goes away. Anyone know of another target reading app that supports the DTP70?
Logged
Cheers,
 Jeff  www.jeff-grant.com

Rhossydd

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3369
    • http://www.paulholman.com
Re: Comparing spectros
« Reply #43 on: June 20, 2017, 01:28:49 am »

I also discovered today that Colorport puts out an amount of bad spectral data.
I've been using Colorport with a DTP70 for years and never had a problem with the combination.
Logged

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Re: Comparing spectros
« Reply #44 on: June 20, 2017, 02:55:02 am »

I've been using Colorport with a DTP70 for years and never had a problem with the combination.

I envy your good fortune or superior knowledge. I'll attach a couple of grabs. The second is from CTP which shows L of 86.79 for patch 1. The first is from the Colorport CGATS file which shows L of 6.28.

If you have a DTP70, would you measure a target for comparison? That would be very interesting for me.
« Last Edit: June 20, 2017, 02:58:17 am by Jeff-Grant »
Logged
Cheers,
 Jeff  www.jeff-grant.com

Alan Goldhammer

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4344
    • A Goldhammer Photography
Re: Comparing spectros
« Reply #45 on: June 20, 2017, 08:23:46 am »

Here's the target that I was looking at initially: https://chromachecker.com/info/en/page/A2

Certainly not cheap, but more trustworthy?
I went to the website and it appears to be part of a kit with a $500 price tag on it.  I couldn't find a link that allowed for separate purchase though I imagine if you contact the company perhaps you can. 
Logged

GWGill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 608
  • Author of ArgyllCMS & ArgyllPRO ColorMeter
    • ArgyllCMS
Re: Comparing spectros
« Reply #46 on: June 20, 2017, 10:15:31 am »

Here's the target that I was looking at initially: https://chromachecker.com/info/en/page/A2

Certainly not cheap, but more trustworthy?
Given that there seems to be no indication as to whether  a reference file is supplied or not, never mind whether it is individually measured or what instrument/standard it is measured to, I don't see why you think it might be used to establish the accuracy of  DTP70 or i1Pro class instrument.

On the contrary, Looking through their marketing material, it seem that all they are offering here is a target for relative measurement - i.e. tracking how a particular instrument changes, or comparing between your different instruments. They aren't providing a high accuracy baseline measurement against a reference.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Comparing spectros
« Reply #47 on: June 20, 2017, 11:00:46 am »

I'm making an assumption that the O/P's objective is to select the profile-making solution that will help him print file values as accurately as his equipment and materials will allow. In that context, to put a bit of perspective into this discussion, I think it's important to be realistic about what's important and hence what's needed. Spot-on accuracy with prosumer or lower-end professional equipment and materials is not likely to be achievable, but for most people using this stuff it also isn't important, because beyond a certain point human visual perception has limits that make some levels of tolerances acceptable; and these levels can vary depending on what colours and what aspects of the colours one is looking at. For this class of equipment and materials, most peoples' needs would be adequately served using methods that are capable of ranking which produce lower dE than others relative to known reference values for the patches on the target being printed. So, for example, if I print and measure a printer evaluation target with a resulting dE average of 0.8 (and low dispersion around the mean) versus another (say, using another profile) with average dE of 2.5 and higher dispersion, I would tend to prefer the 0.8 result, even knowing that part of the difference could also include for variances in the performance of the instrument being used for the measurements. So going back to the O/P, if he were to use his DTP-70 and i1Pro2 to make profiles with which he prints evaluation targets having known references values and measures them with the same equipment used to make the profiles, he can find out enough to make a workable choice for his future work.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Re: Comparing spectros
« Reply #48 on: June 20, 2017, 08:35:00 pm »

Given that there seems to be no indication as to whether  a reference file is supplied or not, never mind whether it is individually measured or what instrument/standard it is measured to, I don't see why you think it might be used to establish the accuracy of  DTP70 or i1Pro class instrument.

On the contrary, Looking through their marketing material, it seem that all they are offering here is a target for relative measurement - i.e. tracking how a particular instrument changes, or comparing between your different instruments. They aren't providing a high accuracy baseline measurement against a reference.

Thanks. That's why I asked the question. It would appear that such a product for measuring accuracy does not exist.
Logged
Cheers,
 Jeff  www.jeff-grant.com

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Re: Comparing spectros
« Reply #49 on: June 20, 2017, 08:55:30 pm »

I'm making an assumption that the O/P's objective is to select the profile-making solution that will help him print file values as accurately as his equipment and materials will allow......  So going back to the O/P, if he were to use his DTP-70 and i1Pro2 to make profiles with which he prints evaluation targets having known references values and measures them with the same equipment used to make the profiles, he can find out enough to make a workable choice for his future work.

Thanks Mark, that sums it up well. I've been down a lot of rabbit holes along the way, mainly, I now suspect due to Colorport rather than the DTP70 itself. I started this thread because I could not reconcile what I was seeing from the two devices.
Logged
Cheers,
 Jeff  www.jeff-grant.com

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Re: Comparing spectros
« Reply #50 on: June 20, 2017, 09:44:35 pm »

I may be missing something very basic here and would love to be enlightened. I have discovered that Colorport gives quite different measurement files depending on the output format. Measuring a 2033 patch target I discovered that each file is different. The worst example is between PM5 output and CGATS. The first two patches have a de of 79 and 87 respectively and then 0 for the other 2031 patches. Other variations produce differences but not as dramatic. I have measured the patch set a number of times and the PM5/CGATS difference is repeatable so it wasn't just one reading.

It's probably my own lack of knowledge at play here but I would not have expected to see differences when saving the same set of data in different formats. I have put the files on Dropbox if anyone want's to see them: https://www.dropbox.com/sh/qv3mu764lrbrmrc/AADSRQl5TlSeaDI5TRMSkL2Va?dl=0
Logged
Cheers,
 Jeff  www.jeff-grant.com

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Comparing spectros
« Reply #51 on: June 21, 2017, 12:07:24 am »

I may be missing something very basic here and would love to be enlightened. I have discovered that Colorport gives quite different measurement files depending on the output format. Measuring a 2033 patch target I discovered that each file is different. The worst example is between PM5 output and CGATS. The first two patches have a de of 79 and 87 respectively and then 0 for the other 2031 patches. Other variations produce differences but not as dramatic. I have measured the patch set a number of times and the PM5/CGATS difference is repeatable so it wasn't just one reading.

It's probably my own lack of knowledge at play here but I would not have expected to see differences when saving the same set of data in different formats. I have put the files on Dropbox if anyone want's to see them: https://www.dropbox.com/sh/qv3mu764lrbrmrc/AADSRQl5TlSeaDI5TRMSkL2Va?dl=0

I looked at the 4 files with PatchTool.

The CGATS Data XRGA 380.txt file differs from the others in that it also has computed LAB values in addition to the spectral values.

The other major difference is that it's spectral numbers are scaled to 100 whereas the others are scaled to 1.

It also appears to be spectrally identical to ProfileMakerCGATS XRGA.txt except for the scaling factor of course.

Other than that they are reasonably close. The CGATS Data XRGA 380.txt median error with ProfileMakerCGATS Gretag.txt is  0.54, and with ProfileMakerCGATS None.txt 0.24
Logged

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Re: Comparing spectros
« Reply #52 on: June 21, 2017, 12:41:26 am »

Thanks Doug. I understand the differences. What I don't understand is why there are differences. These files come from the same scan.
Logged
Cheers,
 Jeff  www.jeff-grant.com

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Comparing spectros
« Reply #53 on: June 21, 2017, 01:18:46 am »

Thanks Doug. I understand the differences. What I don't understand is why there are differences. These files come from the same scan.
The conversion to XRGA adjusts the spectral data slightly so that the product lines that got merged into x-rite would produce more closely aligned data. That's what produces the biggest difference. I'd guess the smaller change is just specific to the instrument with a small adjustment so the instrument more closely matched some internal reference between different products in their line.
Logged

Rhossydd

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3369
    • http://www.paulholman.com
Re: Comparing spectros
« Reply #54 on: June 21, 2017, 01:43:59 am »

I now suspect due to Colorport
Colorport is fine and works well. The problem here is your understanding of the data it produces and what you're doing with it.
« Last Edit: June 21, 2017, 03:04:33 pm by Rhossydd »
Logged

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Re: Comparing spectros
« Reply #55 on: June 21, 2017, 02:11:40 am »

The conversion to XRGA adjusts the spectral data slightly so that the product lines that got merged into x-rite would produce more closely aligned data. That's what produces the biggest difference. I'd guess the smaller change is just specific to the instrument with a small adjustment so the instrument more closely matched some internal reference between different products in their line.

Thanks Doug.
Logged
Cheers,
 Jeff  www.jeff-grant.com

Pat Herold

  • Full Member
  • ***
  • Offline Offline
  • Posts: 161
Re: Comparing spectros
« Reply #56 on: June 21, 2017, 07:23:43 pm »

The kinds of differences you are referring to sound exactly like the kind introduced by XRGA.  I could not find any recent info on xRGA on the X-Rite website.  The link to the white paper referred to here is no longer active:
http://whattheythink.com/articles/53636-x-rite-xrga/

But I will include our own copy here.  You'll have to read through it and decide which flavor of export is more appropriate for your purposes.
Logged
-Patrick Herold
  Tech Support,  chromix.com

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Re: Comparing spectros
« Reply #57 on: June 21, 2017, 07:54:43 pm »

Thanks Pat.
Logged
Cheers,
 Jeff  www.jeff-grant.com

Ethan Hansen

  • Full Member
  • ***
  • Offline Offline
  • Posts: 148
    • Dry Creek Photo
Re: Comparing spectros
« Reply #58 on: June 23, 2017, 12:29:34 am »

Regarding XRGA, this standard came about because of X-Rite's acquisition of GretagMacbeth. Both companies calibrated their instruments to internal reference standards. This led to instruments from CMB producing slightly different readings on the same substrate as did legacy X-Rite devices. As a unified entity, the goal was to define a common reference point and calibration standard that minimized systematic differences between existing instruments and would be the baseline for future spectrophotometers.

If you look at the average post-XRGA measurement differences between instrument families (see the pdf Pat posted above), they are in roughly the same range as inter-instrument variation for each instrument. In other words, don't fret it.

Back to the original topic of this thread, I pulled the measurements we made comparing various spectrophotometers. I only included measurements made directly on BCRA tiles (hauled around the country to different facilities), not secondary standards such as plastics or prints that fit in scanning instruments. We measured eight devices for all but the Spectrolino and I plotted average and maximum average dE for the 14 tiles. We had measurements for 19 Spectrolinos. The chart is attached below.

A general trend emerged: higher-end instruments have somewhat lower average inter-instrument dE and dramatically lower maximum dE. Even with the higher sample size, the old Spectrolino exceeded the capability of newer instruments in terms of maximum measured differences.

Mark brought up whether these differences are significant in the real world. If you use larger than minimum patch sizes with an i1Pro and scan slowly, the argument degenerates into trivial quibbling. Minimum size patches and speedy scanning are another matter, but this comes down to instrument limitations, not inherent accuracy.

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Re: Comparing spectros
« Reply #59 on: June 23, 2017, 03:15:25 am »

Thanks Ethan. That picture is worth a thousand words. I have taken to using my i1p2 on the io in spot mode after reading some of the previous posts. I have increased patch size by 1mm each way. Do you have a suggestion for an optimal size?
Logged
Cheers,
 Jeff  www.jeff-grant.com
Pages: 1 2 [3] 4   Go Up