Dot Gain in Photoshop

[marc3llo]

Hi to everyone, I am not new at the forum but this is my first time posting something.

I am confused with dot gain in Photoshop. And the question is, if I am converting my image from one profile to another or rgb to cmyk, Do I have to compensate the image with the dot gain % in Photoshop? What about if later on I send my job to the rip and apply my compensantion curve that I ´ve got trough a fingerprint? Will I have a double compensation? Or should I set my dot gain % in Photoshop to 0% because, later on I am going to apply via rip? Can someone explain this?

Thanks a lot!

Marcelo

 

[kyle]

Assigning a profile will not change the values, only the appearance on screen (and potentially how the image is converted later if the profile is saved with the image and, e.g., your RIP is configured to honor embedded profiles and perform conversions based on them). If you edit an image, it is best to first assign the most accurate profile available so you have the best preview of the final result.

Converting to a profile, rather than assigning, preserves the appearance by adjusting the values so that the appearance in the new profile matches the appearance in the old profile. This is best for when an image looks fine as-is, but is assigned a profile that will not match the final result.

You will only get redundant conversion if your RIP is configured to automatically perform the same conversion you manually apply to images.

By "dot gain in Photoshop," you could be referring either to Photoshop's default grayscale profiles, or the parameter in the custom CMYK profile creation feature, which I don't think that many people use except to control black generation. If you are referring to custom CMYK profiles, as with grayscale, you are only changing the values if you convert rather than assign.

Ideally, you should have a CMYK profile for the press that is completely ignorant of everything your RIP is doing (like the press curve). It should essentially describe the final color result for CMYK values after ripping, plating and printing. For example, if you had a press that printed with a 30% dot gain (50% prints as 80%), but you want to make it effectively have a 20% dot gain (50% prints as 70%), you would have a press curve that lightens a 50% tint to something like 43%, so that the 43% on plate would become 70% considering your dot gain. With that curve, 50% in the artwork becomes 43% on plate, then 70% on paper, giving you the 20% dot gain desired. An accurate color profile would then describe 50% input as 70% output, and the profile would be completely ignorant and independent of your press curve. If you created a custom profile that matched what a press curve is doing, rather than the final result, it would describe a 50% input as a 43% output instead of 70% as it should. You can never create an accurate profile based upon changes your RIP is making - you have to know what the final result is. If you knew, for example, that your RIP was configured to cause your final printed result to match "U.S. Web Coated (SWOP) v2," then this would be the ideal profile to use, regardless of whatever your RIP is doing to get there.

 

[marc3llo]

First of all Kyle, thanks for the quick reply.

By dot gain in Photoshop I mean, when you are converting from one profile to another, let´s say SWOP to Japan, if you go into "custom CMYK", you will find the "ink options" and here you may choose the type of ink and the "dot gain", standard (btw -10 to 40 %) or "custom" to enter your values. This dot gain compensates in some way my image? Is the same o sort of the same this compensation to the one that I apply in rip (compensation curve).. not an easy topic, specially when my english is not the best!!

Speaking about the best color profile, Would it be the one that is like my compensation curve? (eg, file 50% print 80% - file 50% trough profile 80% - color proof will show 80% when the file is 50%.

Well thakns a lot again!

Marcelo

 

[Mike Strickler]

Dot Gain in Photoshop

Let's keep this simple. Think of dot gain as the color gradient curve. This is already implicit in any CMYK profile, which describes the press's color input/output performance. On one side of the profile's lookup tables you have CMYK going in and on the other side the color value (defined in L*a*b*) going out. Dot gain numbers per se are not written into the profile, but since it was made from measuring actual press sheets those L*a*b* values reflect whatever dot gain was present. One way you can measure the dot gain of a profile easily is to make a good proof with color bars using the profile as the reference. (Be sure that the color bars are also printed through this profile.) Then just measure the dot gain on those color bars. You should have 25%, 50% 75%, and solids to get a decent curve. It's accurate enough for most purposes.

When you convert RGB or CMYK to a CMYK profile the incoming color will be "bent" in the opposite direction to compensate for the dot gain in the destination profile and produce the closest appearance to the source document that is possible on the press or printer described by the output profile.

The dot gain in the Custom Profile option is used when one wants to create a synthetic CMYK profile without actual characterization data, and it is very crude. I would discourage its use and encourage you to use a standard profile from a good source like Fogra, IDEAlliance, or IFRA that best matches your printing conditions, including dot gain. If your print condition is "standard" expect for dot gain I would urge you to investigate and correct that problem at the source rather than tinker with the dot gain of an existing profile or worse, create your own in Photoshop.

Mike Strickler
MSP Graphic Services

 

[kyle]

...converting from one profile to another, let´s say SWOP to Japan, if you go into "custom CMYK"....

If you use a custom CMYK profile, you are not converting to one of the "Japan..." profiles, you're just converting from SWOP to a custom CMYK profile.

The dot gain curve should be an input to output function with image tints as the input and final result on press as the output. It should not bear any relation to the curve applied by your RIP.

The higher the dot gain value entered, the LESS ink there will be after conversion. If you enter a dot gain of 40%, you are describing a final output condition that has very heavy dot gain, and therefore the conversion will lighten the ink so that you preserve the original appearance. You will not see much difference on the screen, because Photoshop automatically applies the profile you converted to, and the lighter ink will appear darker because the 40% dot gain profile says it will. You would definitely notice the difference on press, however - the image would be much lighter.

Are you printing proofs from Photoshop? If not, your RIP or some other part of your workflow should be set up to create color-managed proofs using a profile for the press and a profile for the proofer (or potentially a single device-link profile, which would sort of be like the press profile and proofer profile combined into one file). Both of these profiles should also be unrelated to the press curve in your RIP, unless the proof is based on the exact same raster image that you will plate from. In that case, the curve is already applied to the raster image, and a normal press profile that is based on pre-RIP CMYK values would not be correct and a mirror image of the press curve would have to be multiplied with the normal values in the profile to compensate for that fact.

If proofs are ripped independently of the plate images, the press profile used for proofing should have nothing to do with your in-RIP press curves or dot gain compensation, and should be the same as what you would use in Photoshop.

 

[kyle]

The dot gain in the Custom Profile option is used when one wants to create a synthetic CMYK profile without actual characterization data, and it is very crude.

I definitely agree. A custom CMYK profile can be handy in a pinch when you want to specifically control black generation (how much black versus CMY is used to make gray and dark colors) and the color appearance is not critical. There are certainly better ways of doing that also.

At my last job, I didn't have any decent tools to create accurate profiles. I tried to cheat by sampling CIELAB values for all the inks and overprints that Photoshop allows you to specify when creating a custom CMYK profile and tints of all primaries to approximate the dot gain/TVI. I used all of this data to create a custom CMYK profile in Photoshop, and after many attempts and tweaks was still better off with a standard profile (I think SWOP was the winner).

 

[Luc St-Pierre]

The “Dot Gain Feature” in PS custom profile will only allow you to preview on your monitor the file as it will look at final print, and this feature is very approximate. Total ink coverage is something else. Be aware that Photoshop is definitely NOT the tool to create an ICC profile. Many of the algorythms in there are false. Try this: create a custom profile entering lab values for CMY and K. Imagine a bad ink brand with a blue corruption in magenta. Once your profile is created, make a new document based on this profile, draw a square colored with the lab value of your custom magenta and put the reader on it. YOU WILL NOT READ 100% Magenta but a combo CMYK value. So long for Photoshop's capacity of creating a profile. It would be a good investment to buy some profiling software. They are less expensive than they used to be and you will have a better control on your profiling activities.

 

[marc3llo]

Thanks a lot for your time and knowledge. You really clarify this issue for me.

I am gonna try this things out and see if I can learn something from it.

Thanks!
Marcelo