When did OEMs start using uV cut (M2) for their canned profiles?

[Doug Gray]

I've noticed the profiles for newer printers all seem to be made with uV cut spectros. My old 9800's canned profiles as consistent with M0 but the Pro1000 is all M2. Other, newer printers I've looked at profiles for are also M2.

BTW, I consider this a good thing. There is less and less uV in indoor lighting and even window light significantly attenuates uV and low end spectros like the ColorMunki is M2 only. And the high end iSis scans M2 twice as fast.

But I wonder when this trend started. I don't recall any discussion of this shift.

[Mark D Segal]

Doug, how did you ascertain whether a profile is made including M0, M1 or M2 as the measurement condition? I don't see anything from either ColorSync Utility or ColorThink Pro directly informing on that parameter.

Whether a good thing or not, another question. For printing papers, the issue isn't the amount of UV in the scene, it's the presence or absence of FBAs in the paper. For papers with no FBAs, it doesn't matter whether one profiles with or without UV cut. For papers with FBAs, the question becomes whether we wish the profile to account for them or ignore them. In principle I believe (subject to correction) that including them would help neutralize their impact on printed colours, whereas ignoring them could allow a bluish bias in the colour temperature of the print - not that I've seen such in the several M2 profiles I've used for papers with moderate FBA presence.

[Alan Goldhammer]

Doug, how did you ascertain whether a profile is made including M0, M1 or M2 as the measurement condition? I don't see anything from either ColorSync Utility or ColorThink Pro directly informing on that parameter.

Whether a good thing or not, another question. For printing papers, the issue isn't the amount of UV in the scene, it's the presence or absence of FBAs in the paper. For papers with no FBAs, it doesn't matter whether one profiles with or without UV cut. For papers with FBAs, the question becomes whether we wish the profile to account for them or ignore them. In principle I believe (subject to correction) that including them would help neutralize their impact on printed colours, whereas ignoring them could allow a bluish bias in the colour temperature of the print - not that I've seen such in the several M2 profiles I've used for papers with moderate FBA presence.

When I first started using Argyll several years ago, I was fooling around with all the settings in the Colprof module (the program that creates the ICC profile).  I remember that changing the illuminant value would result in a bluish cast profile for Ilford Gold Fiber Silk.  I cannot remember what value gave the cast but I did make a note to leave the default setting alone and the resultant profile did not have any blue issues.  Perahaps in your profiling software there is a way to deal with this.

[Mark D Segal]

When I first started using Argyll several years ago, I was fooling around with all the settings in the Colprof module (the program that creates the ICC profile).  I remember that changing the illuminant value would result in a bluish cast profile for Ilford Gold Fiber Silk.  I cannot remember what value gave the cast but I did make a note to leave the default setting alone and the resultant profile did not have any blue issues.  Perahaps in your profiling software there is a way to deal with this.

"My profiling software" is simply X-Rite's i1Profiler. One of the normal ways of using this application is to profile in dual reading mode, which reads for three measurement conditions: M0, M1 and M2. The application then saves three profile versions: M0, M1 and M2. One may then make test prints with each of them and use the one that produces the most accurate or most visually pleasing outcomes as the case may be.

I can't relate what i1Profiler does to your comment about Argyll without you being able to recall which setting produced which result.

[howardm]

I think it actually just *saves* the M0/1/2 *data* in various .mxf (?) files and you specify
which dataset to make a profile from (on the Measurement page).

[Mark D Segal]

I think it actually just *saves* the M0/1/2 *data* in various .mxf (?) files and you specify
which dataset to make a profile from (on the Measurement page).

If you are talking i1Profiler, in the Measurement tab if you select "dual scan" it captures M0, M1 and M2 information all at once. But yes, it creates one profile at a time by going back to select the measurement condition before making the profile. (Reading data in Measure Chart saves the data in specified file format for all three conditions at once - different process, and I was probably thinking of this one.)

[Doug Gray]

Doug, how did you ascertain whether a profile is made including M0, M1 or M2 as the measurement condition? I don't see anything from either ColorSync Utility or ColorThink Pro directly informing on that parameter.

As you know there is nothing in the profile itself that tells whether M0,M1,M2 was used to create it. But there are telltale signs. The most blatant one easy to use in Photoshop is converting a "white" in device RGB space to Lab using Abs. Col. and looking for shifts in b*.  It's most noticeable on high OBA paper profiles like Canon's Matte or Epson's Ultra premium glossy, but also on moderate OBA papers. I have a Matlab script that prints the shift for all profiles that match a keyword. For instance `ShowProfileAttributes('Canon`) will print a summary of Lab values for the white and black points for any Profile with "Canon" in the description.

Whether a good thing or not, another question. For printing papers, the issue isn't the amount of UV in the scene, it's the presence or absence of FBAs in the paper. For papers with no FBAs, it doesn't matter whether one profiles with or without UV cut. For papers with FBAs, the question becomes whether we wish the profile to account for them or ignore them. In principle I believe (subject to correction) that including them would help neutralize their impact on printed colours, whereas ignoring them could allow a bluish bias in the colour temperature of the print - not that I've seen such in the several M2 profiles I've used for papers with moderate FBA presence.

M2 profiles with high OBA papers doesn't change the overall print much with Perc. or Rel. Col. because they both scale to the white point which is always paper white or 255,255,255. It can make differences in the mid tones but they are subtle in my experience. It depends on how similarly the uV is attenuated by the inking. Testing M1 and M2 profiles of the same paper I see some differences but they are quite small. One way to visualize it in Photoshop with custom profiles is to soft proof and switch between M2 and either M1 or M0 profiles in either Perc. or Rel. The few times I've done that show almost no change.

OTOH, there is a huge difference printing with Abs. Col of known colors like a CC image. A M0/1 profile print will look quite yellowish with indoor low or no uV content lighting while an M2 print will look pretty normal. The reverse occurs viewing the prints outdoors. However, Abs. Col. is rarely used outside of hard proofing or reproduction work.

[Doug Gray]

When I first started using Argyll several years ago, I was fooling around with all the settings in the Colprof module (the program that creates the ICC profile).  I remember that changing the illuminant value would result in a bluish cast profile for Ilford Gold Fiber Silk.  I cannot remember what value gave the cast but I did make a note to leave the default setting alone and the resultant profile did not have any blue issues.  Perahaps in your profiling software there is a way to deal with this.

Argyll has an option to synthesize M2 from M0 measurements. This was useful back in the day when I1Pro's were either M0 or (uV cut M2) and most people owned a regular I1Pro (M0). XRite's M0 has significant amounts of uV, about 70% as much as the full, D50, M1.  I have not tested this option in Argyll. Most all of my Argyll work was done directly with M2 data.

[Doug Gray]

If you are talking i1Profiler, in the Measurement tab if you select "dual scan" it captures M0, M1 and M2 information all at once. But yes, it creates one profile at a time by going back to select the measurement condition before making the profile. (Reading data in Measure Chart saves the data in specified file format for all three conditions at once - different process, and I was probably thinking of this one.)

Yep, one cool thing is that I1Profiler stores M0,M1,M2 from dual scans in the profile so it's easy to go back and make different M profiles by dragging the ICC profile into the measurement tab, selecting the M desired, and making a new profile. No need to re-scan charts.

[dehnhaide]

Yep, one cool thing is that I1Profiler stores M0,M1,M2 from dual scans in the profile so it's easy to go back and make different M profiles by dragging the ICC profile into the measurement tab, selecting the M desired, and making a new profile. No need to re-scan charts.

Brilliant pieces of information, Doug! Thanks a lot for sharing!


Sent from my iPhone using Tapatalk

[howardm]

+1

I didn't realize that but I'm assuming that the little checkbox that says (approx) 'include cxf data' has to be on?

[Doug Gray]

+1

I didn't realize that but I'm assuming that the little checkbox that says (approx) 'include cxf data' has to be on?

Yes, that's the default in I1Profiler. OEM's generally don't include spectral data in the profiles and it makes the profiles somewhat smaller. Especially those with smaller 3D LUT cubes.

[Alan Goldhammer]

Argyll has an option to synthesize M2 from M0 measurements. This was useful back in the day when I1Pro's were either M0 or (uV cut M2) and most people owned a regular I1Pro (M0). XRite's M0 has significant amounts of uV, about 70% as much as the full, D50, M1.  I have not tested this option in Argyll. Most all of my Argyll work was done directly with M2 data.

I have the standard i1Pro so I guess I am getting M0 data out of it.  I think the M2 approach is used when addressing OBA rich papers as per the discussion Graeme has on the 'colprof' command.  I've not looked at the option to synthesize M2 data as the profiles I get with the standard setting give very good results when using standard test prints.  Also, I stay away from papers with significant OBA amounts.

I did go back and recreate the 'blue cast' profile.  If you set the illumination spectrum to D65 from the standard D50 you get a distinct blue cast when profiling Ilford Gold Fiber Silk (a paper that has small amounts of OBAs).

[Doug Gray]

I have the standard i1Pro so I guess I am getting M0 data out of it.  I think the M2 approach is used when addressing OBA rich papers as per the discussion Graeme has on the 'colprof' command.  I've not looked at the option to synthesize M2 data as the profiles I get with the standard setting give very good results when using standard test prints.  Also, I stay away from papers with significant OBA amounts.

I did go back and recreate the 'blue cast' profile.  If you set the illumination spectrum to D65 from the standard D50 you get a distinct blue cast when profiling Ilford Gold Fiber Silk (a paper that has small amounts of OBAs).

Or any other paper. Argyll has a lot of cool options that can create special effects.

As for M2 v M0 profiles using standard settings, I've never seen anything materially different in a Perc. or Rel. Col. print where the paper didn't have really high OBAs. Only printing with Abs. Col.  Big differences there. Even with relatively small amounts of OBA.