Myshkin wrote:

that's a good point, if everyone was submitting their own cmyk files. But most mag submissions want rgb, unless someone can clarify that for me?

a design firm that i do retouching for wants cmyk files. fashion photographers want rgb files. stock photographers want rgb files (with their agencies' black point and white point values)

Robert Randall wrote:
....I will bet you everything I own, including all cash assets, that my picture will look better to everyone than yours will. Mine will be brighter that yours, it will have better contrast than yours, and the colors will sing compared to yours. And all of this will be due to how I manipulate the file after it has been separated. That is why they don't automate the workflow from RGB right to a plate rip.   

However, as soon as guys like you, that don't have a clue about the process, demand in a loud enough voice that they do implement the process you desire, it will happen. And then images will get just a tad crappier looking.

+1  !

The conversion from RGB to CMYK is quite complex, even expensive color servers like GMG's can not do as good of a job as a human can on every image, but they do a pretty good job at converting thousands of graphics.

Please read the pdf I linked to in my earlier repy, it takes skintones as an example and shows how the conversion leaves the cyan plate bland in a lot of fleshtone instances. The parrot in the first PDF is a prime example of how you can do a better job of separating.

http://teched.vt.edu/GCC/HTML/VirtualTe … umbers.pdf

as well as this:

http://teched.vt.edu/GCC/HTML/VirtualTe … ending.pdf

Because that is how the art director I work for preps his files for layout.

Lol art directors are a pain in my ass, in general they are great with creative vision, but lack foresight into how the nitty gritty printing process works (and changes!). The second biggest pain in my ass are print salesmen....aarrrrr lol.

I would bet money that his files where altered before output. I run into this on a daily basis when people repeatedly make some sort of mistake, and is fixed before printing.....they see the finished product and assume they are doing everything perfectly! A recent example was a black box over a high contrast blue and white star background, seems fine except indesign overprints by default leaving rich black over the blue areas and just black over the stars....leaving a black and gray box instead of solid black.

Comes with the territory I guess, if no one notices we did anything, were doing our job

that's a good point, if everyone was submitting their own cmyk files. But most mag submissions want rgb, unless someone can clarify that for me?

Most larger print companys prefer RGB because it usually makes a better conversion into there color servers.

so they are going from

RGB ---> Custom CMYK

instead of

RGB -----> Generic CMYK -------> LAB ------>Custom CMYK

then they proof you on there in house separation, which hopefully you see on a matchprint proof.

I highly prefer doing my own conversions due to my first point above. A lot of print shops will give you a discount for supplying them with rip ready work as well.

Studio 144 wrote:
There is nothing better than closed loop color! Tweaking an image for a given output is the finest way to do color. The issue is in today market place and with today deadlines it is almost impossible to close the loop.

Let try to send your file to a half dozen printers to meet the deadline and then try to send my file and average the quality of the final images.

Oh, by the way I grew up over my family's print shop and was deviling presses before I started kindergarten. I have worked as a process camera man, scanner operator, pressman and etc. I have my BS in printing management and I teach graphic commutations at a university. So you are probably right about me not knowing anything!

I've been working in pre press since I was an engravers apprentice in 1975. So, we probably know enough to kill each other with the forgotten knowledge. In speaking to the folks in this place, I always take the high road, meaning that I believe they all aspire to a better level of work than that of the masses. When I tell someone how something can or should be done, I assume they are thinking national ad or editorial. I never assume they are thinking of producing the kind of crap Jay Leno makes fun of on Monday nights, which is what it appears you are referring to. Anecdotal to this, I had a meeting yesterday with a retoucher to discuss among other things, the Harley Davidson catalog. I learned from him that on average they spend about 40 man hours of retouch on each motorcycle in their book. Do you suppose they work in RGB and send their files to a direct to plate rip.

When I provide files to my clients, they are in CMYK, and they are tweaked, as you put it, to the highest possible fidelity I can deliver. If I can do that based on my impossible schedule, why can't these people learn to do it? Why does the best workflow have to cater to the lowest common denominator?

What the hell is graphic commutations?

What university are you dispensing this crap at, I want to call the chancellor and tell him to get his money back. Quit teaching parity and start teaching quality.

Robert Randall wrote:

One of the smarter questions actually.

The worst thing you can do in RGB is show your client a full blown wide gamut saturated to the nines file that can't be reproduced in print because the colors are so completely out of gamut. You will look like a complete moron by the time the prepress house is done with you, and your client will likely never return for a repeat performance.

Never show a top end file. Same with transparency they look so pretty sharp and colourful on a light box.

c_h_r_i_s wrote:

Never show a top end file. Same with transparency they look so pretty sharp and colourful on a light box.

Typically, if you don't indulge in distortion practices, a transparency's tonal range and color pallet is pretty easy to match on system, as long as you start with a good scan. Velvia might be one of the exceptions.

c_h_r_i_s wrote:
who ever sets the print press.

This could be true 15+ years ago when presses weren't half as automated as they are today. However, with any decent company today it doesn't matter who sets up the press or really what press it's on.

Granted I'm sure there are some minor differences, but they pail to what can be screwed up at the file level.

IMO

Andrew Thomas Designs wrote:

This could be true 15+ years ago when presses weren't half as automated as they are today. However, with any decent company today it doesn't matter who sets up the press or really what press it's on.

Granted I'm sure there are some minor differences, but they pail to what can be screwed up at the file level.

IMO

I printed a promotional brochure about a year ago. It was printed on a brand new 8 color Komori. The pressman had the first form up to color in about 8 sheets. It used to take at least a few hundred sheets to get up to contract color on the older machines. The newer equipment is absolutely phenomenal!

RSM-images wrote:
.

Embed the printer’s ICC profile into a copy of the image and go from there.

Also, inquire about other parameter settings and incorporate them.

Typically, however, printers are not enthralled about the CMYK efforts of amateurs -- there is a lot more to such an output conversion from RGB than meets the eye.

.

Can you be more specific? What would you say is the single biggest problem when converting to CMYK, and do you have a method for working around the problem?

Robert Randall wrote:

I've been working in pre press since I was an engravers apprentice in 1975. So, we probably know enough to kill each other with the forgotten knowledge. In speaking to the folks in this place, I always take the high road, meaning that I believe they all aspire to a better level of work than that of the masses. When I tell someone how something can or should be done, I assume they are thinking national ad or editorial. I never assume they are thinking of producing the kind of crap Jay Leno makes fun of on Monday nights, which is what it appears you are referring to. Anecdotal to this, I had a meeting yesterday with a retoucher to discuss among other things, the Harley Davidson catalog. I learned from him that on average they spend about 40 man hours of retouch on each motorcycle in their book. Do you suppose they work in RGB and send their files to a direct to plate rip.

When I provide files to my clients, they are in CMYK, and they are tweaked, as you put it, to the highest possible fidelity I can deliver. If I can do that based on my impossible schedule, why can't these people learn to do it? Why does the best work flow have to cater to the lowest common denominator?

What the hell is graphic commutations?

What university are you dispensing this crap at, I want to call the chancellor and tell him to get his money back. Quit teaching parity and start teaching quality.

I would not say the procedures that I suggested are just for "crap" work. I would say they are more for higher volume and modern economic factors. No mater if you and I like it the new work flow it is going to happen. It does not mean you can't tweak a image to make it print as well as the old work flow you just have to commit to the new ways and use them to the fullest. The best fine art book I have seen has been done with color management! (Bill Atkinson's Within the Stone) There is nothing wrong with the process just most people implement it badly!

The day of the craftsman journeyman printer are sadly over. Today people just don't understand the concept of spending time to truly learn a craft. Even if they did management does not understand the need to pay them a wage to make it worth the time. Management would prefer to spend money on hardware which is "suppose" to make the process bullet proof and does come close. People today simply want things to work and don't care how it is done.

Graphic Communications is kind of what use to be printing management. The title has change to reflect the modern programs. To day we not only teach ink on paper but pixels on the web.

You asked if the Harley Davidson catalog would be sent direct to plate. YES it would! It probably would be sent direct to plate as a CMYK but it will go direct to plate. In the near future it will go direct to plate as a RGB (if it is not now.)

For the longest time I thought that printers and buyers would never accept non-halftone approval proofs but they are now doing it all the time.

If you ever print to a 6 color ink jet printer you are using color management. It is the same as having the RIP on the platesetter do the conversion to CMYK. I don't think you have written a custom print drive so you can separate into Hexachrome and tweak the seps.

I teach at the university what my customers (employers and student) and employer want. There is simply no reason or time to teach the obsolete technology. I wish my university was a really an ivory tower which did not worry about the real world.

Robert Randall wrote:
I've been working in pre press since I was an engravers apprentice in 1975. So, we probably know enough to kill each other with the forgotten knowledge. In speaking to the folks in this place, I always take the high road, meaning that I believe they all aspire to a better level of work than that of the masses. When I tell someone how something can or should be done, I assume they are thinking national ad or editorial. I never assume they are thinking of producing the kind of crap Jay Leno makes fun of on Monday nights, which is what it appears you are referring to. Anecdotal to this, I had a meeting yesterday with a retoucher to discuss among other things, the Harley Davidson catalog. I learned from him that on average they spend about 40 man hours of retouch on each motorcycle in their book. Do you suppose they work in RGB and send their files to a direct to plate rip.

When I provide files to my clients, they are in CMYK, and they are tweaked, as you put it, to the highest possible fidelity I can deliver. If I can do that based on my impossible schedule, why can't these people learn to do it? Why does the best work flow have to cater to the lowest common denominator?

What the hell is graphic commutations?

What university are you dispensing this crap at, I want to call the chancellor and tell him to get his money back. Quit teaching parity and start teaching quality.

Studio 144 wrote:
I would not say the procedures that I suggested are just for "crap" work. I would say they are more for higher volume and modern economic factors. No mater if you and I like it the new work flow it is going to happen. It does not mean you can't tweak a image to make it print as well as the old work flow you just have to commit to the new ways and use them to the fullest. The best fine art book I have seen has been done with color management! (Bill Atkinson's Within the Stone) There is nothing wrong with the process just most people implement it badly!

The day of the craftsman journeyman printer are sadly over. Today people just don't understand the concept of spending time to truly learn a craft. Even if they did management does not understand the need to pay them a wage to make it worth the time. Management would prefer to spend money on hardware which is "suppose" to make the process bullet proof and does come close. People today simply want things to work and don't care how it is done.

Graphic Communications is kind of what use to be printing management. The title has change to reflect the modern programs. To day we not only teach ink on paper but pixels on the web.

You asked if the Harley Davidson catalog would be sent direct to plate. YES it would! It probably would be sent direct to plate as a CMYK but it will go direct to plate. In the near future it will go direct to plate as a RGB (if it is not now.)

For the longest time I thought that printers and buyers would never accept non-halftone approval proofs but they are now doing it all the time.

If you ever print to a 6 color ink jet printer you are using color management. It is the same as having the RIP on the platesetter do the conversion to CMYK. I don't think you have written a custom print drive so you can separate into Hexachrome and tweak the seps.

I teach at the university what my customers (employers and student) and employer want. There is simply no reason or time to teach the obsolete technology. I wish my university was a really an ivory tower which did not worry about the real world.

I know the plates will be made directly from e files, but they won't be made directly from files that didn't go through monster tweaking in CMYK, and they certainly won't be made from RGB files that just went through a rip.


I had Image Print do it for me. I wanted my printer to profile as closely to an Approval as possible.

I thought perhaps you were being Freudian.

com⋅mu⋅ta⋅tion  [kom-yuh-tey-shuhn]  Show IPA
–noun
1.    the act of substituting one thing for another; substitution; exchange.


I don't think you understand my position, and possibly its my fault. I'm not against a managed workflow, I employ a very strict managed workflow. What I am stating is that a push button conversion only gets you close. You can choose the correct sep for the correct press and substrate combination, and your image is still only half way home. How many times have you had to tweak a red because the sep browned it out. How many times did you have to remove yellow from a blue, or add magenta to the same blue, because the sep didn't realistically convey the correct color. My position is that ignorance of these issues is causing people like Star and her art director to offer less than spectacular work because they believe their job is done once the button is pushed. Your high volume work suffers for the same reason. Can you tell me that software is coming that will allow for automatic correction of these issues. I don't think so, but what I think you can offer me is that soon, no one will know enough to care about these issues.

Studio 144 wrote:
I teach at the university what my customers (employers and student) and employer want. There is simply no reason or time to teach the obsolete technology. I wish my university was a really an ivory tower which did not worry about the real world.

I have a degree in graphic communications management, it's a small world!

Studio 144 wrote:
Today people just don't understand the concept of spending time to truly learn a craft. Even if they did management does not understand the need to pay them a wage to make it worth the time. Management would prefer to spend money on hardware which is "suppose" to make the process bullet proof and does come close. People today simply want things to work and don't care how it is done.

This attitude is defeatist and not contributory towards helping others to learn.  You're in a thread where someone is clearly expressing interest in just the type of knowledge you're saying should be pissed away because 'people want things to work and don't care how it is done'.  I'd give my left nut to intern for some of the folks offering advice in this thread, to learn just what you're saying no one wants to know and I know I'm not the only one.  It's utter hogwash to suggest no one cares about doing things right.

Sean Baker wrote:

This attitude is defeatist and not contributory towards helping others to learn.  You're in a thread where someone is clearly expressing interest in just the type of knowledge you're saying should be pissed away because 'people want things to work and don't care how it is done'.  I'd give my left nut to intern for some of the folks offering advice in this thread, to learn just what you're saying no one wants to know and I know I'm not the only one.  It's utter hogwash to suggest no one cares about doing things right.

Anytime you generalize about people there will be exceptions! There are some truly great people who want to excel. I do however feel that the "walmart" mentality has accelerated in the last few years. Most people today really don't want to spend the time and energy to truly understand how thing work.

Oh, perhaps I am just a grumpy old man who is in a bad mood!

Studio 144 wrote:

Anytime you generalize about people there will be exceptions! There are some truly great people who want to excel. I do however feel that the "walmart" mentality has accelerated in the last few years. Most people today really don't want to spend the time and energy to truly understand how thing work.

Oh, perhaps I am just a grumpy old man who is in a bad mood!

I've found the contrary is the rule, that most people I run into that express an interest in these issues, generally do want to learn. In here in the professional forums, I am exceedingly hostile to the people that spout off about stuff they know nothing about as if they were experts. Unfortunately, you got caught in that hostility because you weren't very helpful in your original posts. You will find that if you take the time to dissect an issue and provide accurate answers, you will have more people than you can count sending you thank you notes in PM's.

I still want to know if you think automated work flows will ever overcome the anomalies that occur from a push button conversion. And how do you think they will do this. I've heard it explained that if you have enough sampling information, the cure is a slam dunk, but I've yet to see it. There are some rips that come pretty close, but they aren't packaged in with PS.

Don't be grumpy, get even!

Creative Works LLC wrote:
I really appreciate this thread.  I haven't been responsible for getting images to print since a stint as Editor-In-Chief of a college newspaper back in 2003.

We did send images converted to CMYK and I was also the photo editor 2003, so part of my job was to manage those conversions.  We were going to news print, so my workflow involved adjusting levels and curves, converting color to b/w, converting to halftone, conversion to CMYK and a final lightening, sharpening and contrast adjustment of each image (not to any formula -- my eyes were the formula).

We sent our layout PDFs and photos to an outside printer via ftp.  Our printing was paid for by advertising, so we were frugal with the newspaper and saved enough to produce two special-edition magazines that year.

Overall, we did pretty well, winning 9 Northern California Journalism Association of Community Colleges (JACC) awards:

LINE ILLUSTRATION
3rd Place - Steve Scheit, San Jose City College
Honorable Mention - Steve Scheit, San Jose City College

PHOTO STORY/ESSAY
1st Place - Bill Clearlake, San Jose City College
4th Place - Lauren Adams, Jesse Van Vleck and Bill Clearlake, San Jose City

INSIDE PAGE LAYOUT - TABLOID
3rd Place - Bill Clearlake, San Jose City College
Honorable Mention - Rodney Cortez, San Jose City College
Honorable Mention - Rodney Cortez, San Jose City College

NEWS FEATURE STORY
3rd Place - Dawn Carlson Zajdel, San Jose City College

SPORTS GAME STORY
Honorable Mention - Sergio Serna, San Jose City College

With that in mind, I've got some upcoming projects that will require me to have a much better handle on preparing images for print than the rather simple stuff I was doing for a college newspaper 6 years ago.

I'm taking it all in, and so far, a lot of it is actually making sense.

Please keep it coming if you've got it.

I'm trying to figure out how to provide an example of my work flow that will be understandable and also, maybe most importantly, show before and after images utilizing different sep setups. Hopefully this will show the failure of bush button conversions. The problem is how to make the files appear as they would in CMYK noting they need to be in RGB to display properly in here. Hmmmm... maybe download zip files.

Robert Randall wrote:

Robert Randall wrote:
I've been working in pre press since I was an engravers apprentice in 1975. So, we probably know enough to kill each other with the forgotten knowledge. In speaking to the folks in this place, I always take the high road, meaning that I believe they all aspire to a better level of work than that of the masses. When I tell someone how something can or should be done, I assume they are thinking national ad or editorial. I never assume they are thinking of producing the kind of crap Jay Leno makes fun of on Monday nights, which is what it appears you are referring to. Anecdotal to this, I had a meeting yesterday with a retoucher to discuss among other things, the Harley Davidson catalog. I learned from him that on average they spend about 40 man hours of retouch on each motorcycle in their book. Do you suppose they work in RGB and send their files to a direct to plate rip.

When I provide files to my clients, they are in CMYK, and they are tweaked, as you put it, to the highest possible fidelity I can deliver. If I can do that based on my impossible schedule, why can't these people learn to do it? Why does the best work flow have to cater to the lowest common denominator?

What the hell is graphic commutations?

What university are you dispensing this crap at, I want to call the chancellor and tell him to get his money back. Quit teaching parity and start teaching quality.

I know the plates will be made directly from e files, but they won't be made directly from files that didn't go through monster tweaking in CMYK, and they certainly won't be made from RGB files that just went through a rip.


I had Image Print do it for me. I wanted my printer to profile as closely to an Approval as possible.

I thought perhaps you were being Freudian.

com⋅mu⋅ta⋅tion  [kom-yuh-tey-shuhn]  Show IPA
–noun
1.    the act of substituting one thing for another; substitution; exchange.


I don't think you understand my position, and possibly its my fault. I'm not against a managed workflow, I employ a very strict managed workflow. What I am stating is that a push button conversion only gets you close. You can choose the correct sep for the correct press and substrate combination, and your image is still only half way home. How many times have you had to tweak a red because the sep browned it out. How many times did you have to remove yellow from a blue, or add magenta to the same blue, because the sep didn't realistically convey the correct color. My position is that ignorance of these issues is causing people like Star and her art director to offer less than spectacular work because they believe their job is done once the button is pushed. Your high volume work suffers for the same reason. Can you tell me that software is coming that will allow for automatic correction of these issues. I don't think so, but what I think you can offer me is that soon, no one will know enough to care about these issues.

Yes I am saying that with right profiles and procedures you don't need to tweak the color manually. The profile does the "tweaking" for you. With a large enough profile the difference between the Delta E of an RGB image and one converted to CMYK is very low. Issues do develop where the RGB image is out of CMYK gamma. I check to make sure when I am editing in RGB where it falls in the CMYK color space. I do think that the algorithms for moving the colors into gamma are getting better but not perfect. I also don't believe in the "modern" approach of using "Relative Colorimetric" when compressing the color space, I still prefer "perceptual".

Darn I was trying not to get too technical with this and just keep it as a grumpy old man argument. ;-)

Studio 144 wrote:

Yes I am saying that with right profiles and procedures you don't need to tweak the color manually. The profile does the "tweaking" for you. With a large enough profile the difference between the Delta E of an RGB image and one converted to CMYK is very low. Issues do develop where the RGB image is out of CMYK gamma. I check to make sure when I am editing in RGB where it falls in the CMYK color space. I do think that the algorithms for moving the colors into gamma are getting better but not perfect. I also don't believe in the "modern" approach of using "Relative Colorimetric" when compressing the color space, I still prefer "perceptual".

Darn I was trying not to get too technical with this and just keep it as a grumpy old man argument. ;-)

What is a large enough profile?

What is Delta E?

Did you mean gamut, not gamma?

Are you simply punching up command Y to check how RGB falls into CMYK color space, or do you have a more technical approach?

Perceptual is prettier, but it is procedurally less perfect due to the admission of human error, in your work flow based on technology, how do you justify this divergent practice?

If you sense hostility in my questions, there isn't any. I'm either genuinely interested in your answers, or I know others are interested in them. So please, take the time to answer. If it keeps you up past your bed time, you can be grumpy with others in the morning.

Robert Randall wrote:
What is a large enough profile?

What is Delta E?

Did you mean gamut, not gamma?

Are you simply punching up command Y to check how RGB falls into CMYK color space, or do you have a more technical approach?

Perceptual is prettier, but it is procedurally less perfect due to the admission of human error, in your work flow based on technology, how do you justify this divergent practice?

If you sense hostility in my questions, there isn't any. I'm either genuinely interested in your answers, or I know others are interested in them. So please, take the time to answer. If it keeps you up past your bed time, you can be grumpy with others in the morning.

Delta-E (dE) is a single number that represents the 'distance' between two colors.
http://www.colorwiki.com/wiki/Delta_E:_ … Difference

Yes you are right I meant gamut and not gamma!

There are more technical ways I am sure to see what is out of gamut I simply use command Y and shift command Y with proof setup set to my output profile.

Yes I know that "technically" Relative Colorimetric is more accurate, however sometime good color is not accurate.  Color is the perception of wavelengths of light. "Perception" is the big word in color! I don't even see color the same out of my left eye as I do my right. With relative you can move to colors to the same CMYK value or "clip the color". Perceptual rendering is based on how the humans perceive color. I prefer some inaccuracies over clipping colors.
__________________
http://www.cambridgeincolour.com/tutori … ersion.htm

Relative colorimetric maintains a near exact relationship between in gamut colors, even if this clips out of gamut colors.  In contrast, perceptual rendering tries to also preserve some relationship between out of gamut colors, even if this results in inaccuracies for in gamut colors.
__________________

Robert Randall wrote:
What is a large enough profile?

What is Delta E?

Did you mean gamut, not gamma?

Are you simply punching up command Y to check how RGB falls into CMYK color space, or do you have a more technical approach?

Perceptual is prettier, but it is procedurally less perfect due to the admission of human error, in your work flow based on technology, how do you justify this divergent practice?

If you sense hostility in my questions, there isn't any. I'm either genuinely interested in your answers, or I know others are interested in them. So please, take the time to answer. If it keeps you up past your bed time, you can be grumpy with others in the morning.

Studio 144 wrote:
Delta-E (dE) is a single number that represents the 'distance' between two colors.
http://www.colorwiki.com/wiki/Delta_E:_ … Difference

Yes you are right I meant gamut and not gamma!

There are more technical ways I am sure to see what is out of gamut I simply use command Y and shift command Y with proof setup set to my output profile.

Yes I know that "technically" Relative Colorimetric is more accurate, however sometime good color is not accurate.  Color is the perception of wavelengths of light. "Perception" is the big word in color! I don't even see color the same out of my left eye as I do my right. With relative you can move to colors to the same CMYK value or "clip the color". Perceptual rendering is based on how the humans perceive color. I prefer some inaccuracies over clipping colors.
__________________
http://www.cambridgeincolour.com/tutori … ersion.htm

Relative colorimetric maintains a near exact relationship between in gamut colors, even if this clips out of gamut colors.  In contrast, perceptual rendering tries to also preserve some relationship between out of gamut colors, even if this results in inaccuracies for in gamut colors.
__________________

You missed one!

Studio 144 wrote:
Whoops, I missed answering the question on how large of a profile.

As large as you can get! I normally use about 5,000 with the xrite software. Some people will say you can get away with only about 128 but I like a lot more than that. If I could I would use 16.7 million!

So if I understand this correctly. The larger profile contains the color conversions from color space to color space?

Michael Magers wrote:

So if I understand this correctly. The larger profile contains the color conversions from color space to color space?

Not exactly but close. Look at it more like a graph the more sampling points the accurate the graph.

Studio 144 wrote:

Not exactly but close. Look at it more like a graph the more sampling points the accurate the graph.

Ok I get that part.

Questions:

1.  Do you have a profile for each color space?  Not one that manages them all?

2.  How does someone go about building said profiles? Is it something in addition to color managing the monitor?  (I have Spyder2 Pro).

3.  Is it something that is constantly changing?  Or is it something that once it's done its done?  Because the color shifts should not really change right?  Just how the color is interpreted/viewed.

Sorry about all the questions. Just fascinated by the topic.

Michael Magers wrote:
1.  Do you have a profile for each color space?  Not one that manages them all?

The profile essentially either defines the color space or describes the behavior of a device; there can't be one to manage them all.

Michael Magers wrote:
2.  How does someone go about building said profiles? Is it something in addition to color managing the monitor?  (I have Spyder2 Pro).

It's the same kind of function: you display ("print") colors, and measure how they display ("print"), then use calculations to determine what inputs give the desired outputs. Because inks aren't as pure as light coming from a screen, more measurement points are typically needed for prints than for display.

Michael Magers wrote:
3.  Is it something that is constantly changing?  Or is it something that once it's done its done?  Because the color shifts should not really change right?  Just how the color is interpreted/viewed

In theory, they don't change--that's how we have 'standard' color spaces: they describe how theoretical constructs like sRGB and Adobe RGB should behave.

In the real world, things DO drift; displays shift over time as the backlights get old, or the CRT phosphors wear out; ink formulations can vary from batch-to-batch, temperatures can cause changes, etc. Profiles for devices describe how they do behave.

Michael Magers wrote:
Ok I get that part.

Questions:

1.  Do you have a profile for each color space?  Not one that manages them all?

2.  How does someone go about building said profiles? Is it something in addition to color managing the monitor?  (I have Spyder2 Pro).

3.  Is it something that is constantly changing?  Or is it something that once it's done its done?  Because the color shifts should not really change right?  Just how the color is interpreted/viewed.

Sorry about all the questions. Just fascinated by the topic.

You don't have a profile for a color space. The profile describes the color space of a device (printer, display, scanner, camera). sRGB and other standards are just "general" profiles for common devices.

To build a profile you have a device produce a known target. You then use a spectrophotometer read each color patch on the target. The color values are then put into the profiling software.

The color space of a output device should not change unless something changes with the device (type of paper, ink and etc.).

Creative Works LLC wrote:
An analogy is forming in my fatigued and slightly Port-wine addled brain:

Music.  Think of a multi-channel high-end surround-sound system with metallic cones and a  massive sub-woofer.   On this system, you can play music at its full range and get clear sound, even at full volume.

Then, hook up a pair of cheap, paper-cone speakers.  At full volume, the speaker cones will rip and the only sound they'll produce will be a weak buzz.

But if you turn the volume down, the bass fades to nothing, the high tones are lost and the mid range is flat and lifeless.

So, the music has to be equalized to match the characteristics of the speakers.  With a bit of tweaking, you can bring up the bass and treble to match the limits of the speakers.  Then bring up the middle tones to fill out the sound.  In the end, you have the best approximation of the original sound that the cheap speakers can produce.

So, the aim is to adjust the input to get the best reproduction the output device can handle.

Am I on the right track conceptually?

Not a bad analogy at all. Color profiles are kind like the settings a multiband equalizer!

photoboykane wrote:

Michael Magers wrote:
Ok I get that part.

Questions:

1.  Do you have a profile for each color space?  Not one that manages them all?

2.  How does someone go about building said profiles? Is it something in addition to color managing the monitor?  (I have Spyder2 Pro).

3.  Is it something that is constantly changing?  Or is it something that once it's done its done?  Because the color shifts should not really change right?  Just how the color is interpreted/viewed.

Sorry about all the questions. Just fascinated by the topic.

1-A color profile has uses beyond defining a single space.  You can profile a device and ASSIGN that profile to an image from that device, correcting chromatic and tonal inaccuracies in the way that device renders your known profiling target (IT8, ColorChecker SG, etc.)  You're then using the profile to define the errors in the image.  Then you CONVERT to your working color space.  That's a fairly push-button approach, and run as an action will "correct" hundreds or thousands of images in the course of a couple of hours to a day.  I use a single corrective profile for each camera body, bringing them into line with each other.  The flow is "camera RAW" > "ProPhoto" > Assign EOSxx.icc > Convert to AdobeRGB.  After placement with InDesign, and when the press has been awarded the contract to print the next step is Convert to PressProfile.icc.  Since each press profile is CMYK, the conversion to that profile converts to a CMYK image as well.  Also, the Press Profiles are provided by the 4-color houses that print the books.  Tweaking the image after conversion MAY screw up their ink limits and push colors out of gamut.  It's crucial that if you do that, you are armed with the level of knowledge of ol' Bobbie Randall, cuz printers bid and schedule jobs based on setting up, printing and billing, not on fixing problems created by overenthusiastic PhotoShop users!!!

2-I use Gretag Macbeth's ProfileMaker exclusively for creation and editing of all profiles.

3-The profiles in my color folder are as follows: ProPhotoRGB, AdobeRGB, sRGB, monitor profile, seven individual press provided profiles, three individual camera profiles created by me, and output profiles for each proofing printer/paper combination created by me.  All the other crap provided by Adobe, Epson, et al has been deleted.  I only update profiles when something changes for a given job, such as a different cover stock or varnish/lamination.  Otherwise it's an assembly line process for the most part.  What we have to keep in mind is that the image on the page only REPRESENTS REALITY, it is not intended to be a carbon copy duplicate of the original item or scene being depicted.

I know it's recommended to profile the camera for different lighting conditions, etc.  After exhaustive testing it became clear that a single profile would adequately quantify the issues that particular chip had with how it rendered color and tone.  Also, my "level" is different from others who are answering your posted questions.  I shoot and "color manage" 50-60 books a year, mostly all how-to books dealing with knitting, crochet and quilting.  So the fact that I have to use a tightly controlled "push button" workflow doesn't mean that's the only way, or even the best way, to handle CMYK conversion.  It means that it's what works for me/my publishing company, no more and no less.

I use RIPs on large Epsons to proof the PDF's that the presses use to print our books.  You can do an "adequate" job of proofing in PhotoShop if you have a profile for your press, and have properly profiled your own inkjet printer/paper combination.  "Adequate" being subjective based on exactly how critical your requirements for a perfect match versus a reasonable semblance of what your sending to press.

Let me know if that helps at all, or just adds confusion, or just fills up space between more valuable responses...

You are absolutely right on #1. I simply did not want to go into the can of worms of input profiles. The original question was about conversion to CMYK so there for I only gave information dealing with output profiles. It is amazing to me how close different input devices (scanner & cameras) can be with the right profiling.

photoboykane wrote:

Michael Magers wrote:
Ok I get that part.

Questions:

1.  Do you have a profile for each color space?  Not one that manages them all?

2.  How does someone go about building said profiles? Is it something in addition to color managing the monitor?  (I have Spyder2 Pro).

3.  Is it something that is constantly changing?  Or is it something that once it's done its done?  Because the color shifts should not really change right?  Just how the color is interpreted/viewed.

Sorry about all the questions. Just fascinated by the topic.

1-A color profile has uses beyond defining a single space.  You can profile a device and ASSIGN that profile to an image from that device, correcting chromatic and tonal inaccuracies in the way that device renders your known profiling target (IT8, ColorChecker SG, etc.)  You're then using the profile to define the errors in the image.  Then you CONVERT to your working color space.  That's a fairly push-button approach, and run as an action will "correct" hundreds or thousands of images in the course of a couple of hours to a day.  I use a single corrective profile for each camera body, bringing them into line with each other.  The flow is "camera RAW" > "ProPhoto" > Assign EOSxx.icc > Convert to AdobeRGB.  After placement with InDesign, and when the press has been awarded the contract to print the next step is Convert to PressProfile.icc.  Since each press profile is CMYK, the conversion to that profile converts to a CMYK image as well.  Also, the Press Profiles are provided by the 4-color houses that print the books.  Tweaking the image after conversion MAY screw up their ink limits and push colors out of gamut.  It's crucial that if you do that, you are armed with the level of knowledge of ol' Bobbie Randall, cuz printers bid and schedule jobs based on setting up, printing and billing, not on fixing problems created by overenthusiastic PhotoShop users!!!

2-I use Gretag Macbeth's ProfileMaker exclusively for creation and editing of all profiles.

3-The profiles in my color folder are as follows: ProPhotoRGB, AdobeRGB, sRGB, monitor profile, seven individual press provided profiles, three individual camera profiles created by me, and output profiles for each proofing printer/paper combination created by me.  All the other crap provided by Adobe, Epson, et al has been deleted.  I only update profiles when something changes for a given job, such as a different cover stock or varnish/lamination.  Otherwise it's an assembly line process for the most part.  What we have to keep in mind is that the image on the page only REPRESENTS REALITY, it is not intended to be a carbon copy duplicate of the original item or scene being depicted.

I know it's recommended to profile the camera for different lighting conditions, etc.  After exhaustive testing it became clear that a single profile would adequately quantify the issues that particular chip had with how it rendered color and tone.  Also, my "level" is different from others who are answering your posted questions.  I shoot and "color manage" 50-60 books a year, mostly all how-to books dealing with knitting, crochet and quilting.  So the fact that I have to use a tightly controlled "push button" workflow doesn't mean that's the only way, or even the best way, to handle CMYK conversion.  It means that it's what works for me/my publishing company, no more and no less.

I use RIPs on large Epsons to proof the PDF's that the presses use to print our books.  You can do an "adequate" job of proofing in PhotoShop if you have a profile for your press, and have properly profiled your own inkjet printer/paper combination.  "Adequate" being subjective based on exactly how critical your requirements for a perfect match versus a reasonable semblance of what your sending to press.

Let me know if that helps at all, or just adds confusion, or just fills up space between more valuable responses...

Up until this thread, I've never seen your name before. In looking back through this thread, I can't find any interaction between the two of us. That leaves me wondering why you decided it was in your best interest to call me ol' Bobbie Randall.

photoboykane wrote:
I use a single corrective profile for each camera body, bringing them into line with each other.

Is that sufficient, or is it not useful to profile each lens-body combination?  It is well known that lenses have color casts.  Is it feasible to have a profile of each lens and marry it to the profile of the body used for the photo to create an input profile for that combination?  I suppose it would be more accurate and possibly actually easier to measure a profile for each body-lens combination directly.

The reason I ask is because I'm seeing very noticeable differences between at least three combos:  1) 100-400 / 5D, 2) 50 / 20D, 3) 10-22 / 20D.  I'm sure I'd find differences between the other combos I could try.

Of course this could be a bit of a nightmare for a photographer with five bodies and ten lenses: 50 profiles.  And then one would have a workflow problem applying the profiles.

Ultimately I want to master and completely control color in my workflow, even if I have to take the long way around to learn it and even though I might not be able to afford all the equipment immediately.

I love threads like this one!