Wierd Color (HELP)

[Tovis]

A bit ago we had a job that had multiple 4 color gradients with a 50% opacity used to create a nifty texture over a background that was also a 4 color gradient.

Samples that we printed for the client looked nice, but they had small yellow tinge near the top of the gradients with the 50 percent opacity.

We use Color Burst 8.2 on a Vutek Fabrivu, and I profiled the printer using an xrite i1.

The proof was okayed, so we ran the larger scaled job. That yellow tinge got bigger and was more prevalent in other sections of the print.

What is mysterious is that the yellow tinge seems to occur at the same color points.

Originally I came to the conclusion that the rip was failing because

a) it was a HUGE file, the rtl was near the max (4GB)
b) it had gradients over gradients with opacities

We ran it as vector because it was so large.

We did a sample of the file saved as a Tif afterwards and the same thing happened.

I need to figure out how to potentially fix or explain this. Any suggestions.

 

[KR3signguy]

a 4 gig vector file?
Howly cow, sounds like your right..overkill that your colorburst can't handle.

 

[MikePro]

opacities over opacities over such a huge file... BOOM!
just photoshop the colorations in the background, and then import back into illustrator to add the vectored artwork.
the ONLY way I've gotten my gradients to work EVERYTIME. Everything else has me running back/forth between design/production files til i get it right.

 

[Rooster]

If the same thing happened when you rasterized the file prior to ripping then it's obviously something inherent in the file.

From what you describe the problem has always been there, it just gets worse as it scales up, which problems tend to do.

Without seeing the problem file it's really impossible to try and help any further.

 

[eye4clr]

As Rooster mentions, check the numerical mixes in the file to rule that out or in.

It may also be an issue with UV phosphorescence in the media causing the profiling software to over react and think you media is more blue than it really is. The symptom of this is yellow highlights and quarter tones and usually some pretty messy banding in gradients.

If the EyeOne or ProfileMaker software made the icc, I'm off the mark and it is something else. If your RIP made the ICC, it is a possibility.

 

[Tovis]

The profile was made by colorburst. So far everything has worked just peachy. I've zoomed up in the file and its not that.

Whats happening is defiantly happening at the rip.

The RTL was 4 gig. One of the largest prints was 366" x 98" (we do grand format printing)

 

[eye4clr]

measure the white of the material in LAB. If the "b" value is more negative than -4 or -5, you may have UV issues in the ICC.

A fun trick is to shine a black light on the paper. If it glows, you've got optical brighteners in the material.

 

[Rooster]

Try another profile and see if that's not the problem.

If you sent the file as a tiff and the same problem occurred, but you're saying the problem was NOT evident in the tiff file on screen, then all signs point to a profile issue/possible corruption.

If you can see the issue in the tiff file, but it just gets worse as it scales up, then you need to fix the file. Double check to ensure the layer properties aren't set to something like "overlay" or "softlight" that could be messing with the colors when they get flattened.

 

[Tovis]

The i1 is a UV i1

Also, if I use another profile I'd have to make another one. If I make enother one its going to likely be the same issue. I've printed a full CMYK gammut with 10% increments and everything looks wonderful. I've also printed all the PMS on a PMS chart and the all match within about 98% accuracy. (Some colors are just not printable)

The Factory profiles we got are not really worth a hoot. But I'll give it a shot.

 

[jc1cell]

My ignorant 2cents here

How about trying a canned profile and see how it handles the file.

Also, you can crop out a section of the file that's causing the problem in the art and rip that only reducing the overall size of the file. T help you see if the size is not affecting how the rip is reacting. Also using your profile and the canned profile.

jc

 

[Tovis]

Canned meaning profile we no longer use - being done now.

Yes, We've also isolated an area that we are testing.

 

[rfulford]

Another interesting thing you may try is to try handling the color management on the client side. Copy your profile over to your workstation. Rasterize the background as others have suggested using the profile you created in ColorBurst. If you still see the line, then you know it is the profile that is causing the error. If so, clone the anomoly out and try converting the whole job to device CMYK and re ripping in ColorBurst with color management turned off.

Frankly, errors like this are not uncommon when creating profiles for grand format devices. Taking more readings per swatch can help with this. ProfileMaker has a feature called "Low Test Chart Resolution" that can help with this greatly.

 

[Tovis]

Interesting - I think our other division has profile maker pro. Would it be worth using that instead or colorburst to create profiles? Has anybody else tried this that has colorburst?

I remember when I created the linearization I tried using more steps but came up with funky results. I called Colorburst and they said to stick with the 20 step one. I wondered why they put the others in the program if they were goofy.

 

[rfulford]

I am not familiar with the fabrivue but very familiar with ColorBurst and Vuteks in general. My experience with both the 5300s and the PressVu UVs was that they were difficult to read in strip mode with an Eye-One using ColorBurst's SpectralVision Pro.

I think ColorBurst profiles well but reading with an i1 is much easier with Measure Tool and ProfileMaker. The bottom line is that if the device you are profiling is low res, it is better to take multiple samples of each pass. The "Low Resolution Test Chart" option in ProfileMaker does this automatically when in strip mode. Of course the biggest advantage that ProfileMaker has over SpectralVision Pro is the number of tools and options within the package. If the problem is indeed with your profile, the simplest thing would be to open the profile, edit it and move on. This is not an option that is available with the bundled ColorBurst profiling. Also, ProfileMaker allows you to create your own test charts and specify the number and size of the swatches. This comes in really handy with grand format profiling.

Here are the tricks I learned for profiling the 5300s. Make your swatches as large as you can tolerate. If you are strip reading, be sure you are scanning across the carriage direction and not with it. Take at least 3 readings per patch

 

[Rooster]

Frankly, errors like this are not uncommon when creating profiles for grand format devices. Taking more readings per swatch can help with this. ProfileMaker has a feature called "Low Test Chart Resolution" that can help with this greatly.

Good point. Is there an aperture adjustment for the i1 that will take readings over a larger area to account for the lower screen density of the grand format devices?

 

[rfulford]

Good point. Is there an aperture adjustment for the i1 that will take readings over a larger area to account for the lower screen density of the grand format devices?

Unfortunately no. Actually, I dont recall what the aperture size is on the i-1. I think it is 6mm.

 

[Tovis]

So you'd recommend this ProfileMaker?

Have you ever toyed with i1 share?

 

[rfulford]

ProfileMaker is my program of choice since I left ColorBlind for version 3 of ProfileMaker. I have taken a cursory look at i1 share but not much more. Xrite has announced i1 Profiler which with replace both ProfileMaker and Monaco Profiler. Supposedly, the application will provide the best parts of both packages. IRRC any purchase made between now and release will be eligible for a free upgrade.