• I want to thank all the members that have upgraded your accounts. I truly appreciate your support of the site monetarily. Supporting the site keeps this site up and running as a lot of work daily goes on behind the scenes. Click to Support Signs101 ...

Gradient mesh

Andy_warp

New Member
We had a weird one happen.

Customer supplied vector .ai file with the clouds rendered as masks created with gradient meshes. We were stoked because it was 43 megs for 72 foot x 24 foot layout. Digital proof rendered perfect, however when ripped (Caldera) half of the detail went away. We went the rasterize route and were running into serious upper limits of Illustrator memory limits. The file would have been 18 ppi at size. (boo)

No Corel chimers in, please! We all know...

The color in the file was just a linear gradient...so I just tried changing the document color mode to RGB instead of CMYK. We typically avoid this because it can shift grays and REALLY affect how transparency and effects render.

It rendered correctly and no shift in the gradient at print.
 

Attachments

  • 01.jpg
    01.jpg
    2.3 MB · Views: 327
  • 02.jpg
    02.jpg
    3.6 MB · Views: 288
  • Screen Shot 2018-04-04 at 2.06.19 PM.png
    Screen Shot 2018-04-04 at 2.06.19 PM.png
    578.6 KB · Views: 284

ColorCrest

All around shop helper.
We typically avoid this because it can shift grays and REALLY affect how transparency and effects render.

This might be a clue because this is not normal behavior at all.

What process produced the first digital proof that rendered correctly?

18 PPI for a display that size is oftentimes usual.
 

Andy_warp

New Member
This might be a clue because this is not normal behavior at all.

What process produced the first digital proof that rendered correctly?

18 PPI for a display that size is oftentimes usual.
CMYK vector file displayed correctly, ripped incorrectly.
I was excited because every upsample raster "sky" I see is garbage with compression artifacts and bands.
Glad we figured this asset out!
 

Andy_warp

New Member
Yes, but you still may need to figure out your workflow as to why CMYK grays shift or RGB grays shift. You're not profiled properly. Grays should be grays. It's a basic.
Seriously?

My profile is good, man. I've built some decent ones including one at Delta e 1.6 on an old ass Mimaki with j-teck ink.
Typically when there is a color shift it comes from a designer not paying attention.
We do have to run some spot color adjustments from time to time to account for Caldera adding it's flavor...and black generation.

The dye sub ink we use has a low gamut compared to an epson on photopaper, so honestly I have to be more dialed in considering I can hit less colors.
I have a PMS 485 C that will burn your face off!
My whole "Cool Gray" set is the bee's knees, too.

This is an Illustrator issue...if you change the color mode IN an Illustrator layout with effects and transparency...it does change the appearance.
The dilemma was the file ripped differently than the pdf was displaying.

Here's a backlit booth I printed for a client that uses 485C. This is for PAX east in Boston last weekend.
I found the photo online...it's kind of raunchy.

It consisted of a single sided 50' x 8', a double sided 40' x 8' and a double sided 29' x 8'
I printed the hanging sign too, but last year.

It's all direct print on fabric. 15 hour shift.
 

Attachments

  • 001a.jpg
    001a.jpg
    545.8 KB · Views: 270

ColorCrest

All around shop helper.
Seriously?

My profile is good, man. I've built some decent ones including one at Delta e 1.6 on an old *** Mimaki with j-teck ink.
Typically when there is a color shift it comes from a designer not paying attention.
We do have to run some spot color adjustments from time to time to account for Caldera adding it's flavor...and black generation.

The dye sub ink we use has a low gamut compared to an epson on photopaper, so honestly I have to be more dialed in considering I can hit less colors.
I have a PMS 485 C that will burn your face off!
My whole "Cool Gray" set is the bee's knees, too.

This is an Illustrator issue...if you change the color mode IN an Illustrator layout with effects and transparency...it does change the appearance.
The dilemma was the file ripped differently than the pdf was displaying.

Seriously?

Yup, seriously.

My profile is good, man.

You should actually be referring to much more than a single profile here. Input profiles, rendering intents and, at least, both a CMYK and RGB profile because you tried to RIP the file in question with the two different modes. One mode worked and the other did not. Why do you suppose that happened? EDITED ADDITION: Yes, you probably use a single output profile but other shops often use different workflows to handle CMYK and RGB.

Typically when there is a color shift it comes from a designer not paying attention.

Designer? Exactly what would you say they're missing? Are designers also responsible for the RIP process as some shops who have in-house designers are?

We do have to run some spot color adjustments from time to time to account for Caldera adding it's flavor...and black generation.

Caldera's "flavor" is controlled by the profiling process and settings. Using spot adjustments are fine when the file has a discrete spot color but when the same values are used elsewhere as a process color then what? Caldera receives a certification from Pantone for meeting certain specs from a single ICC output profile and not by using spot adjustments. EDITED to say "certified as well as a particular printer brand model."

The dye sub ink we use has a low gamut compared to an epson on photopaper, so honestly I have to be more dialed in considering I can hit less colors.

What does one do to become MORE dialed in? Read more or less color swatches? Average those readings more than usual? Use a more expensive spectrophotometer? A polarized light source? Calibrate 2 times per shift using a densitometer? What do you do exactly?

I have a PMS 485 C that will burn your face off!

Is this a product of your profiling or a spot color adjustment? (Remember my comment above about Pantone certification.)

My whole "Cool Gray" set is the bee's knees, too.

Is this a product of your profiling or a spot color adjustment? (Yup, remember my comment above about Pantone certification.) Are your prints using Cool Gray the same from both CMYK and RGB and spot color callouts because earlier you said your grays were different from one mode or another.

This is an Illustrator issue...if you change the color mode IN an Illustrator layout with effects and transparency...it does change the appearance.

Yes, if you change the color mode IN an Illustrator layout it does change the appearance and this behavior is normal. However, you make this statement immediately following your statement about Cool Gray and earlier you said your grays were different from one mode or another. The grays in all modes should print the same. If not, there is a workflow issue and / or a profiling issue. Remembering calibration is a very important stop in the profiling process.

Seriously, I don't expect you to answer any of these questions. I've just posted these as some areas you might consider and explore if you ever find the time. Also know that Illustrator is very much a PDF and if you're using Caldera on a Mac, the Mac uses PDF technology to draw the image to the screen. So, the preview looked correct but the file did not initially RIP properly. Illustrator file good, screen good, RIP area maybe not so much.

Good luck.
 
Last edited:

Solventinkjet

DIY Printer Fixing Guide
I would go into the RIP and check what your rendering intents are set to. Usually the default has CMYK and RGB interpreted differently. I would make the CMYK rendering intent the same as he RGB one and see if it changes for the better.
 

spooledUP7

New Member
You might want to try different output file types. I have had seemingly simple things go from strange to perfect by saving out from PDF to EPS.
 

Andy_warp

New Member
Yup, seriously.



You should actually be referring to much more than a single profile here. Input profiles, rendering intents and, at least, both a CMYK and RGB profile because you tried to RIP the file in question with the two different modes. One mode worked and the other did not. Why do you suppose that happened? EDITED ADDITION: Yes, you probably use a single output profile but other shops often use different workflows to handle CMYK and RGB.



Designer? Exactly what would you say they're missing? Are designers also responsible for the RIP process as some shops who have in-house designers are?



Caldera's "flavor" is controlled by the profiling process and settings. Using spot adjustments are fine when the file has a discrete spot color but when the same values are used elsewhere as a process color then what? Caldera receives a certification from Pantone for meeting certain specs from a single ICC output profile and not by using spot adjustments. EDITED to say "certified as well as a particular printer brand model."



What does one do to become MORE dialed in? Read more or less color swatches? Average those readings more than usual? Use a more expensive spectrophotometer? A polarized light source? Calibrate 2 times per shift using a densitometer? What do you do exactly?



Is this a product of your profiling or a spot color adjustment? (Remember my comment above about Pantone certification.)



Is this a product of your profiling or a spot color adjustment? (Yup, remember my comment above about Pantone certification.) Are your prints using Cool Gray the same from both CMYK and RGB and spot color callouts because earlier you said your grays were different from one mode or another.



Yes, if you change the color mode IN an Illustrator layout it does change the appearance and this behavior is normal. However, you make this statement immediately following your statement about Cool Gray and earlier you said your grays were different from one mode or another. The grays in all modes should print the same. If not, there is a workflow issue and / or a profiling issue. Remembering calibration is a very important stop in the profiling process.

Seriously, I don't expect you to answer any of these questions. I've just posted these as some areas you might consider and explore if you ever find the time. Also know that Illustrator is very much a PDF and if you're using Caldera on a Mac, the Mac uses PDF technology to draw the image to the screen. So, the preview looked correct but the file did not initially RIP properly. Illustrator file good, screen good, RIP area maybe not so much.

Good luck.
I really am glad I'm not expected to answer all of this...very gracious considering the broad stroke you slathered all over my color workflow and the twenty years of knowledge it took to build it. My favorite part was when you explained Pantones to me as if I haven't utilized their system since the 90's. Thanks!

I can assure you if I send a grayscale image from RGB or CMYK they are one in the same. Neutral. My rendering intents are the same for each colorspace.The color shifts I was referring to are more in line with when a designer converts a spot color to cmyk and it translates with THEIR working space into cmyk builds to several decimal points. I absolutely think the designer should take some stake in preparing workable art. Many of the things that create rip errors. I partly blame adobe for making things possible in their software that don't really work in the real world...like pantone to pantone gradients. We handle artwork for the same brands but different designers, and I can tell you they are seldom in compliance with the brand's guidelines. We feel it is more important to honor the brand than the designer.

I responded about my "profile" in a general term because that's what I was told I should check. I've created a print condition that is consistent and repeatable. I also consider being a printer more than tweaking rip settings as a kludge to a deeper problem, be it file construction or a poor linearization. Consistently on this forum I see people looking for shortcuts through software when the issue is hardware. Or process control. Not sure if you're familiar with grand format dye sublimation, but I can tell you it's a different beast than other more traditional printing methods.

Speaking to my awesome 485C...yes it is a corrected formula, but how can that not speak to the success of my output profile? Isn't it direct proof that the print condition is solid?

We do admittedly run only one output profile on one printer. We've found this helps consistency which is above all a major concern in my market. Many of the fabric covers we print last for years, and I'm constantly called on to "hit the same color" to either seam to an old piece or stand next to older prints. Poking and tweaking with rip settings can complicate this...although I have full access to 5 years of print log.

The "flavor" I referred to with Caldera really is a thing. I can tell you because I have built profiles with Onyx and Caldera with the same hardware, substrates and environment. They were hugely different. I was honestly bummed when Caldera won out because I prefer Onyx.

When I talk about being dialed in it plays on my complete workflow. Like equipment maintenance, controlling the environment and research and development on the best products to use as a baseline. I would like to get a nicer spectro as I only have and eyeone. When it comes down to brass tacks, unpredictable show lighting can nullify ALL of the efforts I've taken to assure good color.
But yes we average several readings, and in some instances have to rotate the fabric grain between. I also spot measure every patch instead of swiping a whole row. The texture on the fabric can affect readings.

Now....pdf's are very much Illustrator? Not sure I can get on board with that. A pdf generated from Indesign is vastly different than one created natively in Illustrator. On an object to object basis they are handled completely different. As far as throwing MAC technology into the mix...the previews I showed were through Caldera's viewer.

This post started as a simple bug I found in Illustrator, but turned into some kind of test about WTF I know. I've built several color departments from the ground up, and have been a service technician for inkjets large format scanners. I am well versed in most things involving digital arts and imaging and it didn't come from forums, it came from boots on the ground.
 

AKwrapguy

New Member
We had a weird one happen.

Customer supplied vector .ai file with the clouds rendered as masks created with gradient meshes. We were stoked because it was 43 megs for 72 foot x 24 foot layout. Digital proof rendered perfect, however when ripped (Caldera) half of the detail went away. We went the rasterize route and were running into serious upper limits of Illustrator memory limits. The file would have been 18 ppi at size. (boo)

No Corel chimers in, please! We all know...

The color in the file was just a linear gradient...so I just tried changing the document color mode to RGB instead of CMYK. We typically avoid this because it can shift grays and REALLY affect how transparency and effects render.

It rendered correctly and no shift in the gradient at print.


Maybe not a fix for this particular issue but could yo install more memory? Is your system maxed out?

On another not what happens if you save as a flattened .pdf?
 

ColorCrest

All around shop helper.
This post started as a simple bug I found in Illustrator,

So Andy, thanks for your reply and know I'm on your side.

Back to Illustrator; you believe it may be a Ai bug but I believe it's on the RIP side. Are you aware that some RIPs can use more than a single RIP engine such as an Adobe engine for some things and then the RIP can revert to use another engine for some other processes? It might help if you can find out if Caldera uses this method or is your version of Caldera using the latest, greatest version of Abobe's print engine exclusively.

Currently it seems, you now have an Illustrator asset that you need to treat unusually "special" in your work flow the next time in production, if I understand your first post correctly. That's unfortunate to say the least. I wonder how Photoshop would rasterize the file to a 2 GB size.
 

Andy_warp

New Member
So Andy, thanks for your reply and know I'm on your side.

Back to Illustrator; you believe it may be a Ai bug but I believe it's on the RIP side. Are you aware that some RIPs can use more than a single RIP engine such as an Adobe engine for some things and then the RIP can revert to use another engine for some other processes? It might help if you can find out if Caldera uses this method or is your version of Caldera using the latest, greatest version of Abobe's print engine exclusively.

Currently it seems, you now have an Illustrator asset that you need to treat unusually "special" in your work flow the next time in production, if I understand your first post correctly. That's unfortunate to say the least. I wonder how Photoshop would rasterize the file to a 2 GB size.
Yes there is a Ghostscript engine as well.
I believe it is basically like the older Jaws interpreter in onyx.

It works...but definitely not as well as the Adobe's APPE.

The problem we were having at Photoshop had to do with Illustrator's memory ceiling. Photoshop is constrained to Illustrator's "pdf rules" when rasterizing, hence the low resolution.
Honestly I would be fine if nothing was EVER cmyk before it ripped. Part of the problem I have is with the dye sub ink. The cyan is akin to 300c or royal blue.
(no I can't make it "brighter" hahaha)

It gets interesting when a client has a brand color that is just 100% cyan...or 100% magenta. It makes things equally difficult when people use Indesign since it is tailored for press.
And with a lower gamut I have the "it's a simulation of PMS blah blah blah" conversation often. There are some valleys in the color range, but some peaks too.
And the process itself can create some really great floods of color. The sublimation process is cool because banding and machine artifacts can gas out.

Resolution is really the elephant in the room for grand format printing. Everything has to be upsampled, and if it's not done with care there is no going back. It's really why I am so clingy to said asset!

I ran old versions of onyx where you had to stand on one leg and cover the opposite eye in order to get things to rip, the software has come a long long way!
We rarely have errors which is why I was so surprised. Then again we know what we can feed the rip. On stuff this big we have to know it's right before we print.

I am glad you are on my side.

Andy
 

Andy_warp

New Member
Yes there is a Ghostscript engine as well.
I believe it is basically like the older Jaws interpreter in onyx.

It works...but definitely not as well as the Adobe's APPE.

The problem we were having at Photoshop had to do with Illustrator's memory ceiling. Photoshop is constrained to Illustrator's "pdf rules" when rasterizing, hence the low resolution.
Honestly I would be fine if nothing was EVER cmyk before it ripped. Part of the problem I have is with the dye sub ink. The cyan is akin to 300c or royal blue.
(no I can't make it "brighter" hahaha)

It gets interesting when a client has a brand color that is just 100% cyan...or 100% magenta. It makes things equally difficult when people use Indesign since it is tailored for press.
And with a lower gamut I have the "it's a simulation of PMS blah blah blah" conversation often. There are some valleys in the color range, but some peaks too.
And the process itself can create some really great floods of color. The sublimation process is cool because banding and machine artifacts can gas out.

Resolution is really the elephant in the room for grand format printing. Everything has to be upsampled, and if it's not done with care there is no going back. It's really why I am so clingy to said asset!

I ran old versions of onyx where you had to stand on one leg and cover the opposite eye in order to get things to rip, the software has come a long long way!
We rarely have errors which is why I was so surprised. Then again we know what we can feed the rip. On stuff this big we have to know it's right before we print.

I am glad you are on my side.

Andy
When I started here a tech told me dye sub could not even be profiled. They weren't even properly linearized, and using a canned profile. A bad one.
The grays were all freaking over the place. It's rewarding to put in the time and do it right. It went from 50% of our work being emergencies down to maybe 5%.
I will never use a canned profile again.
 

Andy_warp

New Member
This might be a clue because this is not normal behavior at all.

What process produced the first digital proof that rendered correctly?

18 PPI for a display that size is oftentimes usual.
@ Colorcrest
This screenshot shows the error better than my vague words.

Another issue we are having is when people use blending modes in a format other than screen or multiply. They rip incorrectly.
Color - Exclusion - Hard Mix

It's multiply (which works awesome in my beloved) or screen.

I am a proponent of letting the rip do all of my compositing from transparency. I absolutely hate the flatten and move on days.
I'm sorry I gnashed my teeth. Nothing you typed was unintelligible. This was Illustrator acting weird.

The JAWS engine wavy hands the transparency.
 

Attachments

  • clouds.jpg
    clouds.jpg
    2.4 MB · Views: 222
Top