indexed color spaces

[assinippi]

When I try to edit an image in an indexed color space (indexed CMYK, indexed RGB, indexed Grayscale) using the Acrobat touch up tool, Photoshop can't open it, giving an unsupported color space error. Is there a way to use Pitstop to change the color space to a format editable in Photoshop? Currently I am rasterizing the entire page in Photoshop (which then has no issue dealing with the indexed color space), cropping the image I want and reinserting it back into the PDF. I can change the color mode of the image in Pitstop Inspector but it retains the indexed format which prevents Photoshop from opening it using the touch up tool.

My understanding is the indexed color space is a result of compression.

 

[gordo]

My understanding is the indexed color space is a result of compression.

I'm sure I've completely misunderstood your post. AFAIK an image in "indexed" color mode is neither RGB nor CMYK. It is a single channel 8-bit image where each pixel can have one of 256 values. It's not really appropriate for print applications.

best, gordo

 

[assinippi]

Unfortunately (at least for my purposes) there appears to be another definition of indexed color where colors not used in a pixel somewhere in an image are eliminated and the color space becomes "indexed". This is a result of compression. I guess the closest analogy would be subsetted fonts where characters not used are eliminated from the set, except in this case it appears to be colors. I have been running into this for years with client supplied pdf files. Unfortunately many of the images they supply need to be edited further than just converting color space. When you select an image using Pitstop and check the color space it reads "indexed on CMYK color", "indexed on RGB color", "indexed on Gray Color" or variations of these including spot color. When you use the object edit tool in Acrobat 7 or 8 you get the message from Photoshop that "Could not complete your request because a color was specified using an unsupported color space." Converting the color space does not eliminate the "indexed".

 

[leonardr]

Unfortunately (at least for my purposes) there appears to be another definition of indexed color where colors not used in a pixel somewhere in an image are eliminated and the color space becomes "indexed". This is a result of compression. I guess the closest analogy would be subsetted fonts where characters not used are eliminated from the set, except in this case it appears to be colors.
.

You are 100% correct about your understanding of indexed colors (and Gordo is incorrect) - and I had never thought of that analogy to subset fonts (LOVE IT!)

Unfortunately many of the images they supply need to be edited further than just converting color space. When you use the object edit tool in Acrobat 7 or 8 you get the message from Photoshop that "Could not complete your request because a color was specified using an unsupported color space." Converting the color space does not eliminate the "indexed".

Yes, this is a long standing issues between Photoshop and Acrobat...someday we might actually get it resolved...

The only tool that I know of that can "unindex" an image is PDF Enhancer from Apago - it has an option to do that specifically to enable editing as you describe (even though it will now increase the size of the final PDF).

 

[michaelejahn]

Index Color in a PDF

Index Color in a PDF

Hi Assinippi,

When your customer saves from Adobe InDesign to PDF - and that InDesign application contains an image that can be rendered using 256 different colors - indeed, the PDF Library will convert 24 bit or 32 bit into an index color space - even if you turn of "Optimise for Web, even if you turn of all downsampling and turn of all compression, there is not methoud I have found that will prevent this conversion to Indexed color space - this is the case for both Adobe Distiller or the Adobe PDF Library (used when you 'Save As" PDF) .

I ran into this same issue when I had created a 'faked' version of the Macbeth Color Checker in Photoshop, (buy painting filled solid squars, as opposed to taking a picture of the actual item) - I then placed these images (as 1 bit, 8 bit, 24 bit RGB and 32 bit CMYK), exporting them to PDF (using PDF/X variants) and processing them using PDF color separation approaches - then i wanted to open the images up in Photoshop (using the method you descibed - that is, right clicking on the image using the Acrobat Touch up Object tool sto open the image in Photoshop) - so that I could quickly measure new CMYK densities.

- as you know, this "convert to index" fouled me up....perhaps the Adobe Photoshop team and the Acrobat team work on different floors at Adobe and don't lunch together or something.

One method that I discovered to force the images BACK into a Non indexed color space was to open the PDF file up in Acrobat, then under the File menu, select the "Reduce File size..." menu item, and when the "Reduce File Size" Dialog box is revealed, in the drop down menu to the right of "Make compatible with:" --

select "Acrobat 4 and later" - the images are no longer indexed and the image open in Adobe Photoshop CS2 (which is what I use)

This works for me when I used Adobe InDesign CS3 on Windows, saving as PDF/X-4 and "High Quality PDF" when I turned all compression off - and used Acrobat 7.

I have not tested this on any other platforms or versions - perhaps it is worth a try, perhaps not.

Maybe this helps !

 

[gordo]

You are 100% correct about your understanding of indexed colors (and Gordo is incorrect) - and I had never thought of that analogy to subset fonts (LOVE IT!)

Please explain what was incorrect in what I wrote about indexed color.

thx, Gordo

 

[mattbeals]

Indexed images have a base color space, RGB or CMYK. Grayscale is inherently indexed because it contains 256 or less unique colors. RGB and CMYK indexed images have less than 256 unique colors. Which is why I like to compare them to flate compressed images. You can think of it as a losless form of image compression.

 

[gordo]

Indexed images have a base color space, RGB or CMYK. Grayscale is inherently indexed because it contains 256 or less unique colors. RGB and CMYK indexed images have less than 256 unique colors. Which is why I like to compare them to flate compressed images. You can think of it as a losless form of image compression.

Hi Matt,

Still not clear to me. Perhaps more clarity in what you mean by "base color space" ?
Here is an RGB image:

Here is the same image converted to indexed color mode in PShop:

Is this what is being discussed as indexed color?
Also, I don't understand how an RGB image that has been reduced to a palette of 256 colors can be thought of as a lossless form of image compression. Please explain.
thx, g

 

[mattbeals]

It's not that someone has intentionally reduced the palette to 256 colors. The image in its original state has less than 256 unique colors. The sample you have above obviously contains more than 256 colors so it would not be "indexed". Think of the same object as above but with solid concentric circles (rather than the blend you have). When a PDF from InDesign would be exported the image would be Indexed RGB. Another example. Years ago I had an insurance company who created a CMYK PhotoShop image of the front of their brochure. On this panel they had a large flat area of some sort of brown taking up the bottom 2/3rds of the page. The top 1/3 of the page had a orange area. Where the brown and orange met they had a "swoosh" that was about 9pts in size going across the page. This particular CMYK image contained three unique colors, orange, green and brown. When the art was placed into the layout the resulting PDF had this panel as an indexed CMYK image.

You could create an image in PhotoShop and "index" it without seeing any noticeable shift in the image.

 

[gordo]

I think this may be a case of you're right but I'm not wrong.

This seems to be a peculiarity of Adobe's PDF format. From Wikipedia:

"... indexed color is a technique to manage digital images' colors in a limited fashion, in order to save computer's memory and file storage (snip) the color information is not directly carried by the image pixel data, but it is stored into a separate piece of data called a palette: an array of color elements, in which every element, a color, is indexed by its position within the array. This way, each pixel does not contain the full information to represent its color, but only its index into the palette. (snip) The palette in itself stores a very limited number of distinct colors, up to 4, 16 or 256 are the most common cases. (snip)
For example, the PDF file format does support indexed color in other colorspaces, notably CMYK, and Adobe Distiller by default will convert images to indexed color whenever the total number of colors in an image is less or equal than 256."

In PShop,, indexed color mode is a single channel image - not associated with RGB or CMYK (other than perhaps through and imbedded profile?) (Which s what I originally wrote)
In PShop if I convert an RGB (or CMYK) image that contains more than 256 colors into indexed color mode then I see artifacts.
In PShop if I convert an RGB (or CMYK) image that contains less than 256 colors into indexed color mode then I don't see artifacts.

So Adobe Distiller does the conversion to indexed color (automatically?) if the image contains less than 256 colors while maintaining its CMYK or RGB association. Because the image uses a range of colors that fall within the (practically) 256 limit of indexed color mode, you can compress the image without introducing artifacts. Correct?

 

[mattbeals]

I don't know how much I would trust Wikipedia, but basically. That's why I equate indexed RGB or indexed CMYK as a lossless form of image compression. You're basically "indexing" the bits of data and creating a look up table to save space. It's my understanding that the PDF library itself, whether it be Distiller or not, decides to index images with less than 256 unique colors when generating the PDF.

Obviously for print production there would no reason to use indexed mode in PhotoShop.

 

[gordo]

I don't know how much I would trust Wikipedia, but basically. That's why I equate indexed RGB or indexed CMYK as a lossless form of image compression. You're basically "indexing" the bits of data and creating a look up table to save space. It's my understanding that the PDF library itself, whether it be Distiller or not, decides to index images with less than 256 unique colors when generating the PDF.
Obviously for print production there would no reason to use indexed mode in PhotoShop.

The definition of indexed color in Wikipedia is the same as the definition you'll find in any graphic arts book - including the educational ones from vendors such as Heidelberg and Agfa. None that I have found (and I've got a ton of reference material) ever mentions conversion to indexed mode in PDF creation.
So, I have only experienced indexed color in PShop where it's primarily used to create GIFs or reduce image size for apps like Powerpoint. When I read the original post, (and I did say then that I may of misunderstood the question) I thought that he was referring to the original images being supplied in indexed format rather than a conversion being applied "under the hood" during PDF creation.

Hopefully "leonardr" from Adobe will clarify this issue a bit more.

thx Matt.

gordo

 

[leonardr]

Wikipedia and Matt are both correct -0 so let's separate things out and maybe it will help.

First - wikipedia is correct as far as the definition of how an indexed image (color space) is stored (regardless of file format - PDF, GIF, etc.). Instead of having color values in the actual image, you have color values in a "palette" (or lookup table) and then the image is just a set of indices into that palette/table. The only thing worth noting is that every other use for indexed color is RGB only - which is why many folks (as Gordo did) ASSUME that indexed == RGB. PDF (and Postscript!), however, support a MUCH MORE flexible definition of indexed, in that the indexing works on ANY (or almost any) of the 11 different color spaces of PDF. So you can have indexed RGB and indexed CMYK as well as indexed ICC, indexed Spot/Separation and indexed DeviceN.

The second part of the discussion is what Adobe's products do when creating and/or optimizing PDF documents. As Matt (and Michael) both said, when our products detect that use of an image that has <256 values - it will automatically be converted to the indexed form in order to reduce file size. There is no option for this because it is 100% lossless - the color VALUES aren't changed, they are just moved out of the image and into a palette. What you did in Photoshop, with >256 colors being reduced to 256 is called quantization - and we do NOT do that when converting.

Hopefully that helps clarify this issue.

 

[michaelejahn]

Indexing images

Indexing images

So Adobe Distiller does the conversion to indexed color (automatically?)

jahn coments - yes - and so does the Adobe PDFLib (for example, when you place an image that contains less than 256 colors - which could be an Orange on an Orange background, or a red apple on a red background, for example) - the image will jump from RGB or CMYK to an index RGB or an Index CMYK image..

if the image contains less than 256 colors while maintaining its CMYK or RGB association. Because the image uses a range of colors that fall within the (practically) 256 limit of indexed color mode, you can compress the image without introducing artifacts. Correct?

ABSOLUTELY - there, you have it exactly - and then, the issue becomes, that Adobe can't eat its own dogfood. That is, now - i can no longer use that Adobe Touch-up Object tool to open and edit an image in Adobe Photoshop.

In you example (the circular rainbow - gee, I never knew you were from San Fran!) - anyway, this image cannot be represented properly using only 256 distinct colors, so you will not get that image to automatically index using the Adobe PDF Lib - or Distiller.

Hope this helps !

 

[gordo]

Thanks all,

This auto-indexing of color by Adobe is something I hadn't read about anywhere and, since I'm not a big user of PDFs, I haven't come across it in my work.
So, thank you. I've learned something.

best, gordo

 

[michaelejahn]

@ Gordo,

Yes, indeed. It was the bane of my existence for a while. This also is an issue when people scan black and white documents in grayscale, then use the Adobe Acrobat PDF Optimizer tool (in Acrobat 7, under the Tool menu > Print Production > PDF Optimizer, in Acrobat 8 and 9, under documents menu > Optimize Scanned PDF) - you can no longer use the touch up object tools (depending on certain settings - like Deskew for example) - people can either delete on object, but the can't open and clone a spec of dirt off an image in photoshop, for example.

Other vendors like ELAN GMK and IoFlex provide deskew and other clean up tools that then create PDF files that can still be edited using the Acrobat touch up object tool > open in Photoshop method.

 

[gordo]

Thanks again Michael. And I do appreciate that you didn't use all CAPs in your responses to me.

best, gordo

 

[assinippi]

michaelejahn

Unfortunately your trick doesn't work on OS X 10.5.4.

 

[Charles]

Indexed Color Space

Indexed Color Space

You can create an action that converts indexed images or graphics in a PDF to their base color space. Once the action has been performed, you can edit the image in Photoshop. Of course, I am sure that everyone has figured that out by now since the original thread is 4 years old.
Thanks

 

[Stephen Marsh]

Any old Scitex Systems operators out there missing editing in LW format?


Stephen Marsh