Hi, your data seem a bit strange, especially the SpectroDens measurements. Please note that the older firmwares of the SpectroDens require separate white calibration per mode, to my knowledge newer versions calibrate M0,M1,M2 at once if the device is in any of this modes, only M3 is treated separately (since one can hear whether the polarization filter is activated and deactivated). With the big differences also in L* and a* I can only speculate that calibration may be the culprit. Anyway I have never found papers that have a b* M0-M1 mode difference of 2 with one device and of 5 with another. Actually, the M1 paper white of FOGRA52 is 93.5 2.5 -10. Even closer... Best regards Hanno
Am 14.02.2016 um 19:36 schrieb Roger Breton <graxx@videotron.ca>:
Dear Herr Hoffstadt,
I think es ist die first time I see a post from you on the ColorSync List -- danke shön! I want to thank you for personally taking the time to post a detailed reply und walking us through your lab at GMG, und share your experience with various devices. The Color Scout table seems like a very versatile instrument indeed :-)
First, I better get hold a CIE Publication 51.2 copy for myself to study...
Second, you say you found "systematic photometric differences between instruments (Minolta somewhat darker, X-Rite lighter and Techkon in between)" but you don't say what differences, if any, you found with respect to "hue" and "chroma", a* and b*?
I made some comparative measurements earlier this week between i1pro2, eXact and Spectrodens III, on Domtar Husky Opaque, an ISO-12647-2:2013 "PS5"-like uncoated substrate, and found the following :
Spectrodens M0 93,28 1,01 -8,21 Spectrodens M1 92,62 2,47 -13,51 Color eXact M0 93,73 2,03 -8,31 Color eXact M1 93,87 2,24 -10,56
As you can see, not much photometric difference between M0 vs M1 between the two instruments (I don't have an FD-7 to compare) but quite a difference with respect to M0 vs M1 between the two instruments in terms of b*?
For comparison, Fogra52 aka "Fred's project" substrate is reported as 93.00 3.00 -10.00, not far from the readings I made on this Husky paper. (The measurements supplied with the ICC profile on the ECI site don't say which instrument they were made with...)
Incidently, will GMG be offering a "Fred52-like" proofing paper in the future?
MfG / Roger Breton
-----Original Message----- From: Hanno Hoffstadt [mailto:Hanno.Hoffstadt@gmgcolor.com] Sent: 14 février 2016 12:21 To: ColorSync Mailing List <colorsync-users@lists.apple.com> Cc: graxx@videotron.ca Subject: RE: Spectrolino repair
Well, for M1 measurements you want to get the effect as if you had used a D50 light source. This can be just a single source (M1 type 1), or multiple sources with a model or compensation (M1 type 2).
For type 1, the "D50-ness" in terms of fluorescence is evaluated as in CIE 51.2, referred via ISO 3664 for daylight simulators. There are three (virtual) fluorescent samples which respond to different peak excitation wavelengths (340, 370, 390 nm). For all three the resulting total radiance factor is calculated, then CIELAB and the Delta E (1976) to a nonfluorescent reference. Actually when you calculate this for D50 itself, you get not only DE=0, but a spectral match. So you see that 365 nm, while practical to use from e.g. a mercury discharge lamp, is only part of the answer. 365/370 nm excitation peak is however quite typical for most OBAs in papers that I have seen. So it makes sense to focus on this particular aspect to get most graphic arts applications - the paper-based ones - "right". - Interestingly, the visual part of the spectrum of the light source "should match" D50, but this is not evaluated in ISO 13655 to my knowledge. So it could be a different source like a white LED or tungsten lamp or a multi-LED mix as used by Barbieri in the Spectropad, which would affect the excitation of fluorescent inks that are excited in the visible range too (like many fluorescent yellows, oranges, reds). Maybe Ray can tell more about the spectral correspondence of the M1-filtered eXact light source to D50 in the visible range?
For type 2, the light source can have any spectrum, and the "D50-ness" of the compensation method in terms of fluorescence "shall be evaluated" using fluorescent certified reference materials, probably similar to the IR3 standards issued from Innventia (Sweden) or Paprican (Canada) for the UV calibration in the paper industry. I personally haven't seen such a D50 evaluation yet.
Anyway, from my experience with the eXact as a type 1-device and with several FD-7, i1pro2, and SpectroDens III, I would say that these M1 devices have a good inter-instrument agreement on most OBA-containing substrates. We at GMG have measured several dozens of test chart sheets with all four devices on the same sheet, thanks to the versatility of the ColorScout A+ xy tables with different adapters. Many of those were sheet-fed and web CMYK offset test charts (like ISO12642 / IT8.7/4) on coated and uncoated paper, e. g. for the creation of FOGRA51 and FOGRA52 in ECI's FRED15 project. Usually we find more of a systematic difference from the photometric scale (Delta L* > 1 on the substrate, with X-Rite being brighter and Konica Minolta being darker, Techkon in-between) than from the OBA, except if it is very strong.
With respect to the subject line (the original Spectrolino/Spectroscan topic), I personally prefer the combination of a universal table like the Color Scout with the device I need for my work (like the M1 ones, or a multi-gloss meter, or devices with spherical geometry etc., or even a Spectrolino). Measurement in spot mode with true lifting up and down is necessary for many packaging materials (e.g. PET and OPP films, also some tissue paper). With so many instruments comes an increased need to focus on a good mechanical or height adjustment for each device in its holder. You may have to experiment a bit. The shipped software is a bit rough but supports all instruments, even a SpectroPlate which can then be used as a automated microscope. Support from third-party suppliers is increasing, not yet from GMG as I am sorry to say, but e.g. from Heidelberger (recently) and basICColor (already for a longer time), for a limited, but probably growing selection of instrument types. (As you can see, I thought this solution deserves mentioning and could be useful to others on the list.)
Best regards
Hanno
Message: 12 Date: Fri, 12 Feb 2016 06:34:49 -0500 From: Roger Breton <graxx@videotron.ca> To: colorsync-users@lists.apple.com Subject: RE: Spectrolino repair
The way I understand the "problem", one light source is needed to cover the non-fluorescing part of the spectrum. That could be a white LED or a tungsten lamp. But, in my opinion, a second lamp is needed at 365nm to excite fluorescence from the substrate. Does ISO-13655 accurately describe this or does it only gloss over the detail of the required illumination, stating the general terms that it "ought" to be a D50-like SPD and leaving the details of the implementation to manufacturers?
I used a Techkon Spectrodens II that has M0, M1, M2 flavors. I used the i1pro2. I could be using an FD-xx. I could be using a Barbieri.
Presumably, none of the manufacturers quite use the same components and methods. Which is "best"? Newer characterization data all specify M1 but it's obvious that "M1" comes in many flavours.
/ Roger