Color Management in Adobe Photoshop® part 4:
The Convert to Profile Dialog Box
(And What You Can Do With It)
With a clear understanding of the differences between assign profile and convert to profile, we can begin to talk about actually using these concepts in Adobe Photoshop. If you do not have a clear understanding of the distinction between the two, I strongly suggest that you read or re-read the previous section, “Assign Profile Versus Convert to Profile”, because the rest of this series assumes familiarity with these concepts. Assign profile and convert to profile can be accessed directly from the top menu in Photoshop. In Photoshop CS and earlier, this was accomplished by selecting from the top menu Image> Mode> Convert to profile or Assign Profile. For CS 2 and 3, choose Edit> Assign or Convert to Profile. The convert to profile dialog looks like this:
Taking the options in this box from the top, we first see the origin space. This cannot be changed because it refers to the color space or profile that the image is in (or is assumed to be in.) Next is the destination space, which is where you want to convert your image to. Any color space or profile visible to Photoshop is available through the destination space pull-down menu.
Conversion options are next. These include the “engine” or color management module (CMM) used to perform the conversion. The CMM is essentially an interchangeable calculator used to translate the color meaning. In the early days of color management, some companies “tuned” their CMMs to take advantage of special “secret sauce” information in the profiles created by their software. This was in direct conflict with the platform-independent interoperability that was a key goal of the ICC, so CMMs that favored a specific vendor were relatively short-lived. Today few profiling software companies bother to create separate CMMs, and Adobe’s Ace Engine has always been one of the best. So unless you have a specific reason to use another CMM, such as a legacy Kodak or Heidelberg profile, Adobe ACE is a better choice than the “operating system freebie” CMMs.
Next down the dialog box is the rendering intent, which is just called “intent” in some versions. Rendering intent mainly tells the CMM how to deal with “out of gamut” colors; colors that exist in the source space but cannot be coded directly into the destination. Depending on the intent chosen, some in gamut colors are usually affected also. It should be noted that when converting into any of the currently used “editing” color spaces such as sRGB, Adobe 98, ProPhoto etc., the only intent currently supported is relative colorimetric. Photoshop CS1 and earlier allowed absolute colorimetric as well, but this was rarely a useful option. Adobe decided that absolute was being chosen by mistake most of the time that it was used and discontinued the option for CS2 and later. So while the other rendering intent choices show up in the menu for these spaces, if you choose one of them and click the preview box you will see that they have no effect.
Rendering intent only becomes a real decision when converting to an output profile such as a printer profile. To give a brief overview of the choices, I will begin with the two most frequently used intents, perceptual and relative colorimetric. The remaining two intents, saturation and absolute colorimetric, are used less frequently and only in special situations. They will be discussed in-depth in the “All About Rendering Intents” (coming soon) section of the website.
Perceptual maps all possible colors (from LAB- it knows nothing about the size of your source profile) to in gamut ones while maintaining smooth transitions to the colors that would have been in gamut any way. This has the advantage of preserving smooth transitions and maintaining relative color relationships, but tends to affect a more significant amount of in gamut color than the other intents. It is widely used for photographic images, where relative color relationships are often more important than absolute color accuracy. To maintain the relationship between in and out of gamut colors however, severe compression is applied to almost all color values, including in gamut ones. This means that some saturated colors that could be reproduced in the destination get somewhat de-saturated. This happens even when no colors in the source would be out of gamut.
Relative colorimetric attempts to leave all in gamut colors unchanged, and map out of gamut colors to the nearest in gamut choice. This makes it an excellent choice if all of the source colors exist in the destination space. If this is not the case however, then the transition to the formerly out of gamut colors can be abrupt. A whole range of out of gamut color may be translated to the single nearest destination color. We call this “saturation clipping” because the end points of a color range are clipped off, or stop becoming more saturated. This changes the relationships between colors and in extreme cases can lead to flat blobs of saturated colors where a gradation would otherwise be. Sometimes this is objectionable, other times it is hardly noticeable.
The choice between perceptual and relative colorimetric is dependent on image, profile and intended use of the converted file. As a general rule, perceptual is often best for translating photographic images into a small printer profile because of it’s ability to scale out of gamut ranges and maintain color relationships. For converting photographic images to large gamut printer profiles, and for accurate color conversion of non-photographic flat color graphics, relative colorimetric is often the better choice. There is no “one-size-fits-all” answer here, so this is where the “preview” check box can be pretty useful. Just keep in mind that you are viewing the color change in the file “through” your monitor profile. This means that if any color ranges lie outside of your monitor’s gamut, you will be unable to see a change even if it takes place. This usually does not present much of a problem, and clicking preview on and off is generally a very good way to ascertain the effects of a profile conversion.
Next in the convert to dialog box is the option of black point compensation. Some profiles translate their darkest possible value directly into the space they are converted into. This means that if you are going from an offset press profile, which is typically not capable of producing a very deep black, to a device profile with darker blacks, the darkest value in the image keep the same appearance as they had in the more limited profile. Thus you lose some advantages of the second device by not being able to print the darker black that it is capable of printing. Not all profiles translate black in this way, and the difference is due to ambiguity in the original ICC specification. Black point compensation is an Adobe proprietary solution which when turned on will always map the darkest possible value in source to the darkest possible value in the destination. Unless you are trying to duplicate a workflow that does not offer the ability to use black point compensation, it is always best to map black using this option by checking the box.
Finally we have “use dither” and “flatten image”. Use dither introduces a tiny amount of noise to the image in the conversion process. It seems counterintuitive, but this is actually desirable for most photographic images. This is because any profile or color space conversion, especially 8-bit ones, will have at least a slight amount of rounding errors in the mathematical conversion. The resulting loss of data, which is usually small, is called quantization. It is almost never noticeable on textured parts of an image, but sometimes can be visible in very smooth gradations (like a cloudless sky) as slight “posterization” or “banding”. The minimal noise introduced by the use dither control will ensure that you do not see this banding as the result of a single transform into a normal profile. So if you have a tonal or photographic image, checking use dither is the way to go. If your image is a flat graphic creation that has no gradients however, then you probably don’t want even the slight amount of noise that this control introduces and you should leave the box unchecked.
The flatten control will collapse all of your Photoshop layers before making the profile or color space conversion. Since some layer options and blending modes are dependent on the profile or space in use at the time, flattening before conversion stands the best chance of keeping the appearance of a file as unchanged as possible. Of course, you lose the advantage of editing any of the pre-existing layers going forward. If this presents problems, try the conversion and use Edit>undo to toggle back and forth to evaluate the visual difference. You must use this method for evaluation because the preview checkbox does not preview the effects of unchecking the flatten box. In other words, preview will always show you the result of flattening before the conversion.
So what can you actually do with the convert to profile dialog? Some very useful things, including:
• Convert to another color space. You could use convert to profile to change a copy of your digital camera RAW conversion in ProPhoto into a smaller space like sRGB to email to your friends. (Most low-end image viewers assume that everything is in sRGB) Or you could convert an sRGB file into a larger space like ProPhoto to achieve more saturated colors than are possible in sRGB. Please note that if you are doing this, the conversion alone to the larger space does not make the color in the file any more saturated. The purpose of convert to profile is to translate to the destination color space without changing the color at all. To saturate the colors beyond those available in a small space you must convert to a larger space and then use adjustment controls like Image>Adjust>Hue and Saturation.
• Convert to a printer profile for printing. If you have a custom profile for your printer, paper and ink combination, this is where you would use it. The best approach is usually to convert to profile and print from Photoshop. Make sure that color management is turned off in the printer driver; you may have to consult your printer manual for this. For an excellent tutorial on printing with Photoshop, click here. Note that you would not want to save the converted version as your only copy of the file because after the conversion it will only contain colors that are printable under that condition, i.e. that paper on that printer with that ink. This would limit your ability to use a more capable printing condition and profile in the future, so if you need to save the converted version, save it as a copy.
• Convert to another color mode. Most people use Image> Mode> and choose the color mode to accomplish this. Convert to profile can be used to convert between most of the commonly used color modes with much greater control. For instance, if Image> Mode> is used, then the default working space for that mode from your Photoshop color settings is always used as the destination space for that mode. So if you were sending a job to a print shop that used a profiled CMYK, press, it might seem to make sense to use Image> Mode> to convert to CMYK and then convert to the press profile. However, this would involve an unnecessary conversion to your CMYK working space that would introduce another level of quantization and could cause additional problems with black generation. Using convert to profile to go straight from RGB to the CMYK press profile would be a better choice. In addition Image> Mode> also uses your Photoshop color setting defaults for rendering intent and dither setting. The only way to independently control these variables for each mode conversion is to use the convert to profile option.