G7 Calibration Using the Default Profile assigned in CWS

[Doyle]

Now what is really strange.

With Print Patches I created a new calibration with G7. Before this did as I posted earlier.

Checked measurement and this time it seamed the Cyan and Magenta values were correct. Finished the the G7 calibration and again checked values, and all looked okay. So finished creating the Output Profile. Prints look great. Recalibrated, values looked okay print still look great.

The only conclusion I can think of somehow a new Calibration needed to be started with ColorWise to reset something so that the FCP started with correct values. After this I did go back and made a new calibration with Express Profiler and recalibrated with the values looking correct and they stayed the same. I will keep a very close eye on this.

All this testing made me realize it sure with be nice to view measurement when in the FCP calibrator.

[Doyle]

Possibly if a paper in the Paper Catalog is selected that has a profile assigned to it possibly FCP calibration is picking up some color management some how even when ColorWise Off is selected.

One of the calibration targets that did not give correct values, when viewed in ImageViewer the 100% Magenta patch was only at 91%. When I choose the locked coated profile that came with the RIP (ColorWise Off still selected) and re-ripped the calibration target the 100% Magenta Patch showed 100%.

I mostly only edit the Magenta and Cyan (taking down the 100% sometime up to 90%) in profiles so that might explain why those are the ones that appear off.

I am no not sure if this is happen when recalibrating from FCP. Doesn't really seem that it does as there you are choosing a calibration but does choosing a paper from the Paper Catalog with a profile assigned affect it is some way also. Guess I will have to test this out as well.

[Doyle]

Checked the processed recalibration targets with ImageView and I found a few with the same thing happening.

I this point I can not trust the FCP calibration. This explains why randomly I have found that in the middle of a G7 calibration suddenly the second or third target is way off.

Until this is fixed I have no choice but to go back to ColorWise for Calibrations which forces me to use the same size paper which was one of the advantages of using the FCP calibration.

[Justin_dB]

Hello All;
I just wanted to chime in - as the discussion here has included many screen shots using CWS to inspect calibration targets and measurements.

There is a bug with the way the targets are shown in CWS - the G7 corrections to the goals are not being reflected in the target curves. Please keep that in mind. The G7 target goals are in place on the server after the G7 process has completed, but CWS is not displaying these corrections. Until this is resolved - please ignore the shape of the target goals that you see displayed in CWS. These goals are in fact different after a G7 target is created in CPS.

[Doyle]

I don't think your understanding the problem here. There are output profiles being applied to calibration targets and recalibration targets. It does not matter if G7 calibration is being used or not.

[Justin_dB]

Here are some details that will hopefully help.

Curve edits associated with output profiles are expected to apply when "CMYK Source" is set to "Bypass Conversion". Curve edits are not expected to apply with "ColorWise Off" - if you see something different - please share so we can investigate.

Step 1 of creating a calibration set, and jobs sent to "Recalibrate" an existing cal set use "ColorWise Off"
Steps 2-x of creating a G7 calibration set use "Bypass Conversion" in order to build on the previous steps.

The result of which is - If an output profile has a correction curve associated with it - Step 1 or Re calibration steps will not be affected by the Curve edit - (OK) - but steps 2-x creating a G7 calibration will print with a curve correction associated with the temporary copy of the output profile which brought the curve edit with it. - (NG) This will cause all kinds of havoc - I think that is what we are seeing here.

To avoid - it is important that default output profiles, or media defined profiles on the system - have the curve edits removed before starting the process of creating a G7 calibration. Today you will need to do this manually before starting. We are discussing if there is an automatic way of doing this in the future.

Generally - if you have the tools to create a G7 calibration set - you won't need to apply output profile curve edits to get an initial state of calibration. If you must use output profile curve edits for your workflow, do this after the cal set and profile have been created - don't apply them in the process of creating cal sets and profiles.

Thank you
Justin

[Doyle]

Here are some details that will hopefully help.

Curve edits associated with output profiles are expected to apply when "CMYK Source" is set to "Bypass Conversion"

I would think that should not be the case as it I would think Bypass Conversion should mean all conversions include any curve edits. But if that is the case I can work around that.

Curve edits are not expected to apply with "ColorWise Off" - if you see something different - please share so we can investigate.

Yes I am seeing something different. My first post with attachments about Express Profiler clearly show that curve edits were being applied, knocked what should have been 100% down to 91%. I am also not sure how the profile was selected with ColorWise Off because I did not select that.

Step 1 of creating a calibration set, and jobs sent to "Recalibrate" an existing cal set use "ColorWise Off"
Steps 2-x of creating a G7 calibration set use "Bypass Conversion" in order to build on the previous steps.

The result of which is - If an output profile has a correction curve associated with it - Step 1 or Re calibration steps will not be affected by the Curve edit - (OK) - but steps 2-x creating a G7 calibration will print with a curve correction associated with the temporary copy of the output profile which brought the curve edit with it. - (NG) This will cause all kinds of havoc - I think that is what we are seeing here.

Step 1 is where I am seeing the problem. It is being affected by Output Profile Curve Edit.

Step 2 I have not really investigated enough at this point other than once the Output Profile with edit curves was introduced into Step 1 it likely carried on to Step 2.

To avoid - it is important that default output profiles, or media defined profiles on the system - have the curve edits removed before starting the process of creating a G7 calibration. Today you will need to do this manually before starting. We are discussing if there is an automatic way of doing this in the future.

Generally - if you have the tools to create a G7 calibration set - you won't need to apply output profile curve edits to get an initial state of calibration. If you must use output profile curve edits for your workflow, do this after the cal set and profile have been created - don't apply them in the process of creating cal sets and profiles.

Thank you
Justin

Thanks for taking time to explain this. Most of it I understand but let me state again it is Step 1 where I am seeing the problem which as you state shouldn't exist. This is why I have brought this up because I could not see this a being correct behavior. And it is especially concerning after finding a recalibration target that the output profile curve edits where applied. And I am see this randomly but I will continue to monitor this to see if I can see a pattern.

For the 51 large patch calibration target do you have the percentages that the colors should be if shown in ImageView. It would be helpful in evaluation whether a curve edit was applied. It is easy to tell if the 100% values were dropped but not for the other patches as some edits I have not dropped the 100% values.

[Justin_dB]

More details:
1. "Bypass conversion" is refers to bypassing the CMM conversion from CMYK Source space to Output space.
2. "ColorWise Off" - behaves like "Bypass conversion" with the additional feature of turning off Calibration
3. Curve edits apply to the calibration curve.

As calibration is applied to Bypass conversion, features like curve edits (that are technically edits to the final calibration curve), will apply.

If you have Curve Edits applying to "Color Wise Off" jobs, that would be un-expected - or I have wrong information.

[Doyle]

More details:
1. "Bypass conversion" is refers to bypassing the CMM conversion from CMYK Source space to Output space.
2. "ColorWise Off" - behaves like "Bypass conversion" with the additional feature of turning off Calibration
3. Curve edits apply to the calibration curve.

As calibration is applied to Bypass conversion, features like curve edits (that are technically edits to the final calibration curve), will apply.

If you have Curve Edits applying to "Color Wise Off" jobs, that would be un-expected - or I have wrong information.

Ok, if curve edits apply to the calibration curve then setting "Color Wise Off" should not be applying those curves, BUT that is not what is happening. Those curves are being applied as you see in my three attachment about Express Profiler in post #9 the original curve (earliest time) shows cyan and magenta low and that is picking that up from the curve edit. In the following curves from recalibration you see magenta and cyan where they should be and where they should have been in the original calibration. But obviously the Output Profile was built on top of the original calibration which was wrong. Why this curve was being pickup up in the original calibration (and sometimes in recalibration) I think you do need to investigate. It is not consistent behavior and after looking at processed calibration targets I can not explain way some profiles were chosen, but still from what you are saying and from what I understood is that if "Color Wise Off" in selecting this should not be happening.

I am assuming that best practice is to choose the default supplied output profile when setting properties before the initial calibration target is printed, but after what I have seen I am not sure if this issue will still not show up.

[Doyle]

Engine Is.