Cross rendering question

[lauphoto]

Hello all,

I am a commercial photographer who shoots food as a specialty. Recently the studio obtained a packaging account thru an ad agency, a good thing. During a number of preproduction meetings we learned that the studio had to function as a prepress provider as well as a photo studio. We are expected and required to deliver press ready files along with a cross-rendered proof that is client approved. This is new for us as we always delivered rgb files as finals and left the separations and any color tweaks to the people who knew the environment for the final printing. The press that will print the project has a fellow with the title “Color Systems Manager” and we have been working with him to get our “ducks in a row”, the ducks being our respective proofing systems.

This is where my question for this forum comes in. We have a new Epson 3880 that has a bundled ColorBurst rip, but we don’t have time to learn how to use it effectively for this job. We have decided to use what we have experience and good results with, the Epson driver. I have built a profile for an Epson Proofing paper, using a Bill Atkinson 1728 patch target and reference file, thru Profile Maker 5.08. The “Color Systems Manager” has sent us two targets and a press profile, so we can cross-render them thru our printer and he will read these target-prints and compare the numbers to his output. We did as he asked and received an e-mail that the delta E CMC 2:1 were averaging around 2 and he would like them below 1. It seems to me that this is like comparing apples to oranges, don’t you think? Shouldn’t we be using the same ink sets and paper?

Here is some additional info:

Targets were right out of Profile Maker, ECI2002CMYK i1_iO_1_2.tif and ECI2002CMYK i1_iO_2_2.tif, both untagged as they should be.

My procedure was to assign the press profile to each target and then print them thru our profile on the Epson Proof paper.

Thanks for any observations.

John

[terrywyse]

We have a new Epson 3880 that has a bundled ColorBurst rip, but we don’t have time to learn how to use it effectively for this job. We have decided to use what we have experience and good results with, the Epson driver. I have built a profile for an Epson Proofing paper, using a Bill Atkinson 1728 patch target and reference file, thru Profile Maker 5.08. The “Color Systems Manager” has sent us two targets and a press profile, so we can cross-render them thru our printer and he will read these target-prints and compare the numbers to his output. We did as he asked and received an e-mail that the delta E CMC 2:1 were averaging around 2 and he would like them below 1. It seems to me that this is like comparing apples to oranges, don’t you think? Shouldn’t we be using the same ink sets and paper?

I'm not sure what you're saying is apples and oranges here. The "Color Systems Manager" was, I'm assuming, simply measuring your proofs and comparing them against the original measurement data that was used for the press profile. Perfectly legit.

First things first, did you print the "cross-rendered" targets using *absolute colorimetric* rendering intent? That's how *proofs* must be printed. As a photographer, you may be used to use perceptual or relative colorimetric. Absolute basically adds paper simulation for the press stock which could easily account for the delta e difference you report.

By the way, "cross-rendered" isn't exactly the correct term for what you're doing. Cross-rendering generally implies *3* profiles and 2 conversions. Cross-rendering generally uses a *simulation* profile that goes between the *source* and *destination* profiles. Examples:
Cross-rendered with RGB as source:
"working space RGB" -> "press CMYK" -> "inkjet printer"
or...
"press A CMYK" -> "press B CMYK" -> "inkjet printer"

If you check out your ColorBurst RIP, you'll see that it has an option to plug in a Simulation profile for "cross-rendering".

Some may disagree with my definition of what cross-rendering is but that's the way I understand it.

I would also suggest that you measure the target yourself (you should be able to find a version of the ECI2002 chart for your instrument) and submit your measurements along with the chart that the other guy wants to use. If you used a different instrument to create your printer profile (say and iSis or DTP70) than what he's using to measure the final results, there can be enough of a difference between the instruments to cause enough of a delta e shift to make your results look worse than they really are. Inter-instrument agreement can be a factor when looking for average delta e values that are in the range of 1 delta e or less.

Regards,
Terry

[lauphoto]

[quote name='terrywyse' date='Apr 12 2010, 07:03 PM' post='359696']
[I'm not sure what you're saying is apples and oranges here. The "Color Systems Manager" was, I'm assuming, simply measuring your proofs and comparing them against the original measurement data that was used for the press profile. Perfectly legit.]

Yes, this is what he is doing. It just didn't seem legit to me, like something was missing. I don't see how the numbers could match (almost match) between our two systems.

[First things first, did you print the "cross-rendered" targets using *absolute colorimetric* rendering intent? That's how *proofs* must be printed. As a photographer, you may be used to use perceptual or relative colorimetric. Absolute basically adds paper simulation for the press stock which could easily account for the delta e difference you report.]

This is how I intended to proceed, "Color Systems Manager" wanted us to use relative colorimetric, so that is what I used. I looked at the press profile information in ProfileEditor and it indicated that it was corrected for optical brighteners,  so I assume he didn't use a proofing paper or press stock.

[By the way, "cross-rendered" isn't exactly the correct term for what you're doing. Cross-rendering generally implies *3* profiles and 2 conversions. Cross-rendering generally uses a *simulation* profile that goes between the *source* and *destination* profiles. Examples:
Cross-rendered with RGB as source:
"working space RGB" -> "press CMYK" -> "inkjet printer"
or...
"press A CMYK" -> "press B CMYK" -> "inkjet printer"]

Sorry for the miss use of terms, but when this is all figured out that is what we will be doing to make the simulated proof.


[I would also suggest that you measure the target yourself (you should be able to find a version of the ECI2002 chart for your instrument) and submit your measurements along with the chart that the other guy wants to use. If you used a different instrument to create your printer profile (say and iSis or DTP70) than what he's using to measure the final results, there can be enough of a difference between the instruments to cause enough of a delta e shift to make your results look worse than they really are. Inter-instrument agreement can be a factor when looking for average delta e values that are in the range of 1 delta e or less.]

Great suggestion Terry, I will do that on the next round. I checked for ECI2002 chart for the i1 (thats what I have) and they are very different from the ECI2002 chart for the i1 IO. How would the data be compared?




John

[Marco Ugolini]

The “Color Systems Manager” has sent us two targets and a press profile, so we can cross-render them thru our printer and he will read these target-prints and compare the numbers to his output. We did as he asked and received an e-mail that the delta E CMC 2:1 were averaging around 2 and he would like them below 1. It seems to me that this is like comparing apples to oranges, don’t you think? Shouldn’t we be using the same ink sets and paper?

No. Cross-rendered proofs produce a visual match. It does not matter at all whether or not the same inks and paper are being used — as long as there is a close visual match (also called a metameric match, meaning that the proof looks very close in appearance to what the final press output looks like, when both are viewed under the same D50 controlled lighting).

As Terry Wyse rightly points out, it's also important to make certain to use the Absolute Colorimetric in producing your inkjet proofs, closely to match on your inkjet proof the effects of the destination paper tone on the press output.

Here is some additional info:

Targets were right out of Profile Maker, ECI2002CMYK i1_iO_1_2.tif and ECI2002CMYK i1_iO_2_2.tif, both untagged as they should be.

My procedure was to assign the press profile to each target and then print them thru our profile on the Epson Proof paper.

That sounds like what you were being asked to do — as long as by "printing through" you mean converting to your custom inkjet profile (using an Absolute Colorimetric intent).