curves workflow question

daveman

In the Whipping Post, Marc Muench, suggested on a shot that I not convert the mode to CMYK because it could cause the colors to go out of gamut. I have some questions about his comment, but since that was the whipping post I decided to ask them here.

Someone who I respect as a really good photographer and PS user, suggested he typically makes curves adjustments in each of the 10 channels, in this order - RGB,CMYK,LAB - flattening in between modes. He then flattens and converts to RGB for printing.

I followed that workflow on the shot I posted in the whipping post. When Marc and others suggested I enhance the shadows, I assumed I should move to CMYK to adjust the K - then back to RGB.

When Marc said to not convert to CMYK, I then started to wonder if there is something wrong with the workflow I used. The person who suggested this workflow gets fantastic results in PS- but maybe I am missing something.

I realize the responses will probably be way over my head, but I thought I would ask anyway.

Thanks for your help.
Dave

LAB.rat

In most circumstances there should be no risk:
www.curvemeister.com/tutorials/cmyk-rgb/index.htm

If you wanne be safe and are just using CMYK for workflow reasons, you can install a Wide Gamut CMYK profile to convert to.
www.curvemeister.com/tutorials/widegamutcmyk/index.htm

If you use curvemeister, there is no more need to convert. You can access all LAB/WGCMYK curves from the plugin.

His WGCMYK profile is also included on one of the Dan Margulis book CDs, so that should count for something

If you want quick access to all of the 10 channels, this might be of some help:
www.thelightsrightstudio.com/TLRSplitChannels.htm

edgework

daveman wrote:

In the Whipping Post, Marc Muench, suggested on a shot that I not convert the mode to CMYK because it could cause the colors to go out of gamut. I have some questions about his comment, but since that was the whipping post I decided to ask them here.

Someone who I respect as a really good photographer and PS user, suggested he typically makes curves adjustments in each of the 10 channels, in this order - RGB,CMYK,LAB - flattening in between modes. He then flattens and converts to RGB for printing.

I followed that workflow on the shot I posted in the whipping post. When Marc and others suggested I enhance the shadows, I assumed I should move to CMYK to adjust the K - then back to RGB.

When Marc said to not convert to CMYK, I then started to wonder if there is something wrong with the workflow I used. The person who suggested this workflow gets fantastic results in PS- but maybe I am missing something.

I realize the responses will probably be way over my head, but I thought I would ask anyway.

Thanks for your help.
Dave

This subject will never go away, nor will it be resolved to the satisfaction of everyone. Yes, the CMYK gamut is smaller than RGB and so, in theory, you potentially stand to lose some color in a conversion. The wide gamut CMYK is a solution that permits you to take advantage of some capabilities that are unique to CMYK, but will not serve you if you actually need to go to a press. But, if that's the case, you have no issues at all: you take what you get and make the best of it. That being said, I think Dan Margulis has it pretty much right when he states that most colors in the real world are not out of gamut colors. It's rare to find highly saturated reds, blues, greens and pure pastels on city streets or the countryside. Not impossible, or even improbable, but cerainly much less than one would expect, given the inordinate fear of conversion that people bring to the subject.

I find CMYK to be far superior to RGB when it comes to fine tuning face colors and detail. Nothing helps faces like a contrast curve to the magenta channel in luminosity mode. I use it on every face I work with. It's worth the loss of a few blues and pastels. If the rest of the image is that color critical, I'll just mask out the skin and pop it back into my RGB file when I'm done with it. Even though I'm preparing images for print, I keep my layered Photoshop files in RGB, so there are all manner of workarounds to take advantge of both spaces.

Shadows have been covered many times and you seem already to be aware of the advantages of working with them in CMYK.

In general, you will not get the subtle fine-tuning with your curves in RGB that you get in CMYK. Each RGB curve is working on the full range of tones, highlights to shadows, and so comparable moves have considerably more weight when they are made in RGB. CMY curves, by definition, dominate in the areas where color counts the most, losing their usefulness at the three-quarter tone range. It's easier to keep the curves focused on precise ranges.

After all that, just keep in mind that every beautiful image you've ever seen printed in a book or magazine, managed to get the job done with the puny little CMYK gamut and it's rotten cyan ink and dirty blacks.

thegridrunner

hmm, I am listening...

daveman

thanks
Edgework and Lab.rat.

Thanks to both of you for responding. I also appreciate links you provided, LabRat - lots of good information there.

I can see that while there may be color shifts when you convert to CMYK, there are advantages to using that color space that for some people may outweigh the potential of color shifts. I am going to do more reading on this.

Thanks again.

Dave

LAB.rat

My pleasure, Daveman and thank you Edgework for your edge info and the lecture

Marc Muench

daveman wrote:

In the Whipping Post, Marc Muench, suggested on a shot that I not convert the mode to CMYK because it could cause the colors to go out of gamut. I have some questions about his comment, but since that was the whipping post I decided to ask them here.

Someone who I respect as a really good photographer and PS user, suggested he typically makes curves adjustments in each of the 10 channels, in this order - RGB,CMYK,LAB - flattening in between modes. He then flattens and converts to RGB for printing.

I followed that workflow on the shot I posted in the whipping post. When Marc and others suggested I enhance the shadows, I assumed I should move to CMYK to adjust the K - then back to RGB.

When Marc said to not convert to CMYK, I then started to wonder if there is something wrong with the workflow I used. The person who suggested this workflow gets fantastic results in PS- but maybe I am missing something.

I realize the responses will probably be way over my head, but I thought I would ask anyway.

Thanks for your help.
Dave

Daveman,

I find it much simpler to recommend to someone that they use the RGB curves unless going to press. I agree with all that is stated above and have used most of the techniques. However, in this case all you really need to do is darken the 3/4 tones in your Arches image and that could be done very easily in a RGB curve.

BTW, I know of many subjects that turn weird colors when converted to the best of the CMYK profiles and printed on a sheetfed press. Once again, until you know those subjects and or must go to press I find it much better to stay in RGB or LAB.

edgework

Marc Muench wrote:

Daveman,

However, in this case all you really need to do is darken the 3/4 tones in your Arches image and that could be done very easily in a RGB curve.

Then at least dampen the blunt force of the RGB curve by converting a copy of the image to cmyk and using the black plate as a layer mask for your adjustment curve in RGB. It's not quite as subtle as an actual black curve would be, but it will do the job without plugging up everything else.

daveman

Marc Muench wrote:

Daveman,


BTW, I know of many subjects that turn weird colors when converted to the best of the CMYK profiles and printed on a sheetfed press. Once again, until you know those subjects and or must go to press I find it much better to stay in RGB or LAB.

Marc,

I follow you now. I think you are saying that because there is a potential of color shifts when you convert into CMYK, you should only convert when needed. In this case you felt the changes could be made in RGB, so no reason to risk a color shift by moving to CMYK. That makes sense to me.

I appreciate your help.

Dave

arodney

Marc Muench wrote:

Daveman,

I find it much simpler to recommend to someone that they use the RGB curves unless going to press.

I totally agree. CMYK is an output color space. Its based on very specific output conditions. Suggesting you convert to such a space to pull curves is like suggesting you convert to the RGB output color space of your Epson printer to pull a curve. The gamut loss is certainly an issue and there's more data loss due to rounding errors in color space conversions. Lab is no different here. Just what the heck does someone have to do, image correction wise that demands a conversion to CMYK when the final image isn't intended for this specific device? And in todays world, why wasn't this issue, if there is an issue, fixed in the Raw conversion where you have infinitely more data, and control over rendering an image?

IF your final output device is CMYK and you have converted a copy of the original AND you need to apply curves based on this soft proof and output color space, fine. But some arbitrary CMYK space for editing? Seems totally silly, unnecessary and a good way to toss some useful data.