Click
HERE
to view it.
Subscribe to:
Post Comments (Atom)
Blog Archive
Didn't find the answer?
Translate my blog into your language (click on flag).
About Me
- Gordon Pritchard (a.k.a. gordo)
- Formerly Print Quality Marketing Manager for eleven years at Creo/Kodak. Presented at print technical conferences, trained printers and buyers regarding print quality issues in Europe, N. America, and S.E. Asia. Articles published in trade journals, co-authored TAGA paper on halftone screening, authored BRIDG's guide to halftone screening. Previously Technical Director of Western Canada's largest commercial sheetfed shop. For several years Professor of Digital Graphic Design at Emily Carr University. Former Creative Director at McCann Ericksson Vancouver. Currently looking for opportunities related to the subjects covered in this blog. Contact me at: pritchardgordon @ gmail (dot) com.
Past Topics - but not every topic covered is listed here.
- AM/XM Screening (27)
- Color alignment (41)
- Color tolerancing (9)
- competition (35)
- DMaxx (1)
- dot gain (12)
- Dot Shapes (5)
- FM screening (21)
- GCR (10)
- Ghosts (2)
- grey balance (6)
- Grey Levels (1)
- Grrrrrrrrrrr (5)
- Halftone Screening (36)
- Hi-Fi color (5)
- Ink (14)
- ink saving (10)
- Ink Sequence (4)
- Investments (13)
- jpeg (3)
- Loupes (2)
- lpi/dpi (5)
- Metamerism (1)
- Misc. (44)
- Moiré (8)
- Optical Brightners (6)
- Pantone (11)
- Paper (3)
- Print 09 (8)
- print buyer (15)
- printing press (35)
- Profit (6)
- proofing (9)
- quality (99)
- RE:Print (11)
- Registration (3)
- Reseparation for ink saving (9)
- Rosettes (4)
- Scanning (4)
- Screen angles (5)
- SID (8)
- Spot color (11)
- Success (64)
- Supplied art (24)
- TVI (10)
- UCR (2)
- Vendors (10)
- Wayback View (35)
Copyright Gordon Pritchard All Rights Reserved
Subscribe To Quality in Print
Posts
Comments
Followers
Links of interest. Don't be afraid to check them out.
I Gordo,
ReplyDeleteI totally agree with you! I am also shocked by this answer!
Louis
That is surprising, however I have encountered quite a few printshops that only use a plate calibration because the prepress guys do not know in advance what press will be used, and hope that at least a generic curve gets them into the ballpark.
ReplyDeleteHi Gordo. I think it makes perfect sense.
ReplyDeleteI use one curve on top of another for a very simple reason. It's faster, easier and cheaper.
As you know, making a proper press curve takes time and money. We had 2 press curves done 3 months ago (for 2 different types of paper), and it took 1,5 day of work for the technician and the press, plus 20 plates and lots of paper. Last week one of the lasers of the CTP broke down and was replaced. I had to linearize the plate again. It took 1 plate and 15 minutes to do. If i dind't have one curve on top of another, it would take another 1,5 day, plus the technician fee, plus 20 plates, plus lots of paper... So, as you can see, theres a good reason to keep using both curves.
All the best.
Tiago
Plate linearization is a good solution to separate plate-curve from press-curve (dot gain). It simplify a lot curves handling. Was implemented starting from Prinergy 4.
ReplyDeleteScenario A (working without plate linearization, 1 curve):
A customer have 1 press with a specific dot gain.
Same customer is using 4 different plate brand.
He create 4 Harmony curves related to both plate linearization and dot-gain compensation.
After some time, for any reason customer needs to modify the dot-gain compensation curve.
He have to put his hands on 4 curves.
Scenario B (working with plate linearization, 2 curves):
A customer have 1 press with a specific dot gain, so he create 1 dot-gain compensation curve.
Same customer is using 4 different plates brand. So he create 4 plate curve linearization. Total = 5 curves.
After some time, for any reason customer needs to modify the dot-gain compensation curve.
He simply modify 1 curve (the dot-gain compensation) and it works fine for every kind of combined plate-brand.
1 edit vs 4: plate linearization won!
Now, think to a customer site where there are 4 presses, 3 plate's brand, 4 or more kind of paper with 3 or more screen rouling output...
Thanks Prinergy to gimme the possibility to separate plate-linearization from dot-gain compensation!
:-)
@Tiago,
ReplyDeleteYour description of the extra work and time you'd have to spend if you only had one curve, rather than the two that you are using, is not correct. This will be explained in a later post on this topic.
Hi Gordo,
ReplyDeleteI have always favoured the 2 curve strategy for the simple reason that it is far easier to linearize your CtP device than to change your trc. If you know your standard settings it is far easier to recalibrate when necessary. There are also too many variables on a press whereas a CtP engine can be calibrated scientifically.
We do 2 plate curves and reason is same/similar to Tiago's.
ReplyDeleteNow I was surprised with your question, but I guess you would not be a Gordo if you did not have good reason :)
Now I can't wait to hear your reasoning and potentially opening our eyes :)
Thanks
@ML
ReplyDeleteIt depends on how you look at it.
Scenario 1 (my way): 4 press curves = 4 curves used in production
Scenario 2 (your way): 4 linearization plate curves plus 1 press curve = 5 curves used in production
Let's say that you want to modify the dot gain curve:
Scenario 1 (my way): 4 new press curves = 4 curves total (45 minutes of work)
Scenario 2 (your way): 4 linearization plate curves plus 1 new press curve (15 minutes of work) = 5 curves total
Because you apply 2 curves per plate you also have at least twice as many potential points of failure and all you've saved is a one time 30 minute time cost.
I think it is possible to have a situation where a two curve strategy makes better sense. However, your case is very, very far from being typical. You are using a two curve strategy to get around a very unique problem, so I don't think that the two curve strategy should be the standard for other shops. This will be explained in later posts.
@Zoran,
ReplyDeleteSimpler is usually better - so a one curve workflow is potentially better that a stacked two curve workflow - less potential for error, and, if you're using high lpi screens, even FM screens - less chance for shadestepping.
I'll try and make my case over the next few posts - but there will always be exceptions where a two curve approach might make more sense - but surely 70% of the market can't be exceptional cases.
@ Stephen
ReplyDeleteI hope to explain this more thoroughly in the upcoming posts - but basically, whether you linearize or change your trc you are building a curve.
With linearizing you expect a 50% request to be 50% on the plate then you apply your trc. In a trc only workflow you expect a 50% request to equal some other value - say 58% and that's it, you're done.
In both workflows if, for some reason, you don't get what you expect on the plate when you request a 50% tone then you make the same adjustments - you to your linearization curve and me to my trc. Same procedure and the same amount of work. Your linearization curve is redundant.
This will be very interesting. My shop is fighting this right now. We have 2 different HeatSet Webs and 2 Sheet Fed 4 Colors but just about every job is built to be inter changed between them when possible. We have a mountain of Process Templates and curves.
ReplyDeleteCurrently we are having big problems from file to proof to press. The part of the solution floated is linearize the plates first.
@Anonymous
ReplyDeleteLinearizing your plates first will have approximately zero to no impact in solving your problems from file to proof to press.
What it will do for your shop is to add complexity to the workflow as well as introduce a new point of potential failure/error to your "mountain of process templates and curves."
Call me thick, but wouldn't two curve approach give you better tool for troubleshooting and consistency?
ReplyDeletee.g. we want to make sure our trc is the same for exact reprints.
If we have some inconsistency in plate output for some reason (new batch of ctp chemistry, new rollers, something, etc..) we can easily just re-linearize the plate without touching our trc, with expectation that exact reprint will be the same on press (this is if something did not change on press itself).
This would also be helpful if you have second location, where you just get their trc and as long as both places have linearized plates, theoretically you can output plates and exchange them.
In one curve approach, we would have to tweak trc to compensate for CTP processor failure, which subsequently mean any exact reprint of jobs that used old trc would not work and you would have to apply new trc, correct?
I guess this could work with rules set for each exact reprint to use the latest trc.
I think we've been there before with our old Brisque workflow and we could never hit exact reprint, but this could have been something else, not curve :(
@Zoran (part 1)
ReplyDeleteLet's suppose that a linear plate provided the tone response on press that we need. Would it make sense to then use two curves - one to linearize the plate (a plate curve) and a second curve (a press curve) to linearize the linearized plate? I doubt it. Makes more sense to just apply the one linearizing curve - based on the uncalibrated natural condition of the plate.
So, if that logic makes sense, why wouldn't it make equal sense if we needed a non-linear press curve? Just apply the one non-linear press curve based on the uncalibrated natural non-linear condition of the plate.
To answer your specific points:
"wouldn't two curve approach give you better tool for troubleshooting and consistency? e.g. we want to make sure our trc is the same for exact reprints."
No, it's not the "trc" that needs to be the exact same for reprints. It is the dots on the plate that need to be the same. This can be accomplished in a single press curve workflow.
"If we have some inconsistency in plate output for some reason (new batch of ctp chemistry, new rollers, something, etc..) we can easily just re-linearize the plate without touching our trc, with expectation that exact reprint will be the same on press (this is if something did not change on press itself)."
You would need to create a new linearizing plate curve then apply your trc. in a one curve workflow you would create a new press curve to deliver the same tones on plate that you had previously. There is no difference in work to this point, except that after linearizing you will apply a second trc curve which I don't need to do.
"This would also be helpful if you have second location, where you just get their trc and as long as both places have linearized plates, theoretically you can output plates and exchange them."
Linearizing is still redundant.
With your two shop workflow both shops will potentially need to have two different linearizing curves on top of which you apply a single trc curve. So now you have three curves being used.
With my two shops, each shop will have their own single press curve so that the final dots on the plates from both shops are the same. That means I have two curves to manage while you will have three. You are increasing the number of variables while I am reducing them. It's much easier to troubleshoot two curves than three just because the points of failure are greater when you add more steps to any process. Just like you, I can interchange plates because the dots on the plates are the same from both locations.
@Zoran (part 2)
ReplyDelete"In one curve approach, we would have to tweak trc to compensate for CTP processor failure, which subsequently mean any exact reprint of jobs that used old trc would not work and you would have to apply new trc, correct?"
Well, I don't know about using curves to compensate for a processor failure - seems like a bit of a stretch. But, that being said, you would need to tweak your linearizing curve and I would have to tweak my trc - we're doing the same work. I'm simply not complicating things by adding a second curve on top of my trc. Your linearizing curve is a trc curve - it's just that it's targeted for the plate rather than the press. So you use two trc curves while I use one.
I doubt that you're thick-headed. If you were you probably wouldn't be asking these questions.
I think it's more about how one looks at a problem. We often do things a certain way because they have a kind of appealing internal logic even though - they make may no sense at all.
For example, many high rise buildings don't have a 13th floor. The elevator goes from the 12th floor to the 14th floor. This is because superstitious folks associate the number 13 with bad luck. So the 13th floor would be potentially harder to rent. But people who live on the 14th floor know for a fact that they're actually living on the 13th floor. They can easily prove it by simply walking up the stairs and counting the floors. However, it just feels better to be on the "14th" floor even though in reality it's the 13th.
Ooops, I may have pushed my analogy too far, sorry:-)
Just to add a little something to Zoran's question.
ReplyDeleteLets assume i have a double curve, and you don't.
We all have our systems under ISO 12647-2 specs, so we both require TVI curves in order to respect ISO 12647-2 standards.
A TVI curve is, as you know, made of 4 distinct curves, one for each process color. And you will need to have a set of curves for each paper. To correct it it takes at least 4 plates plus paper and press time.
So, let's assume our CtP broke down, and is replaced by a new one. We will keep using the same plates and chemistry. In order for me to keep under ISO's specs, all i need is to burn 1 plate, measure it, linearize it and i'm good to go. You my friend, on the other hand, will need to burn at least 4 plates per paper, redo all the TVI curves (4 for each paper) plus paper and press time.
So, with my separation between the linearization and the TVI curves, i took 15 minutes and 1 plate, you will take a few days and lots of plates, paper and press time.
Unless i'm missing something, i strongly believe separating linearization from the rest is the best policy.
Savage
@ Anonymous/Savage
ReplyDeleteNo, I wouldn't need to burn at least 4 plates per paper, redo all the TVI curves (4 for each paper) plus paper and press time.
First off, a TVI curve does not have to be made of 4 distinct curves, one for each process color.
But let's say that's what we have.
As you said you'd burn 1 plate, measure it, modify your linearization curve and you're good to go with your existing 4 press curves. However, I would also burn 1 plate, measure it but then modify my 4 press curves so that I got the same tones I was getting before. No need for me to go to press. I would have 3 more curves than you that would need to be modified - maybe another 15 minutes of work for me compared with you, but in the end you'll be managing 5 curves and I'll only need to manage 4.
Well, i guess you just overlooked one simple problem with your logic. It's 4 curves per paper. Meaning that 10 calibrated papers would require adjustments in 40 curves to you, and just one to me. Thats why i find my 2 curves system better than a single curve one... Although i understand that it will not be the best system to everyone, i think it's the best system for me and my company.
ReplyDeleteAll the best.
Savage
@ Savage
ReplyDeleteI agree with you.
I have difficulty understanding how you arrived at needing 4 curves per ink set to hit ISO 12647-2 since ISO 12647-2 was, AFAIK, based on a linear film workflow - i.e. a one curve workflow. But that is a separate issue.
So, yes, for your shop and your needs a two curve workflow...works best.
best, gordon p