|
|
Note that it won't show up in print preview, but on the canvas and in acroread yes. |
|
|
I was thinking, why not push this further and make this possible with channels? Like, Gimp doesn't allow us to do Spot color images, but working with transparency allows everyone to turn colors into channels. You probably see where I'm heading now. ;) Well, we could have maybe an effect or something that would allow people to map a color to each channel, all set to overprint if desired. Think it would be cool? |
|
|
maybe such simple feature for the beginning: 1. convert grayscale to colorscale. 2. add color to grayscale (colorize) + with balance. others operations should be done in graphical software, like photoshop or gimp (graphics SHOULD BE prepared in dedicated applications first, imo). |
|
|
Yes, but will Gimp export to DCS? No. Until then, there is no way to make spot color bitmap illustrations with Scribus on Linux. I think mapping colors to channels would be a good solution. |
|
2006-05-23 07:41
|
colorize_pic_1.png (345,249 bytes) |
|
2006-05-23 07:42
|
colorize_pic_2.png (361,853 bytes) |
|
|
DCS? what for? ok, i will try to describe my idea better. we have any grayscale bitmap (in examples i have converted colour bitmap, but with grayscale result would be the same), and we want to make it red (c0m100y100k0) for example. in both cases scribus should treat/interprete grayscale bitmap as cmyk - and print it as cmyk also. 1. scribus converts any value of black (0-255 or 0-100%) to value of red (c0m0y0k0-c0m100y100k0); for example 128 (50%) black should be c0m50y50k0 (50% red) - see 'pic_red.png'. this is simple case. 2. there is more difficult graphical filter neccessary. i don't know the filter in gimp, but in photoshop you may do this using hsb values with 'colorize' option on - see 'colorize_pic_1.png', or use two layers: first with main picture, second filled with the solid color (red), and the first layer have to be blended using luminocity - see 'colorize_pic_2'.png (both methods have different results of course). |
|
2006-05-23 10:22
|
pic_red.png (130,782 bytes) |
|
|
Yes, we're pretty much saying the same thing. It's similar to putting a color layer set to Screen over a grayscale layer in Gimp. :) |
|
|
most important is, that we agree with main subject ;) |
|
|
Faced the same issue and found it partially resolvable. Had to print a grayscale image in spot color from Scribus. Steps: 1) Using Gimp, decolorize the image (keeping it RGB). Add alfa. Create layer transparency based on white color. Convert to grayscale. Save as PNG. 2) In Scribus, place the image and fill its frame with the spot color you want (must be created beforehand of course). 3) Output to PDF 1.6 to preserve transparency. Check the file in Acrobat Pro's Output Preview: your image (in fact, its background masked by the image) is there, on the spot plate of your choice. I believe this is only usable when you're not intending to print CMYK inks at all; otherwise, the image will produce output to these plates, and the result will depend on the sutiation. |
|
|
There is actually a way to do this through the Image Effects > Colorize menu in Scribus. You need to first create the spot color you want so you can select it to colorize your image. |
|
|
@louisdesjardins no, Colorize does not work for this, especially not for black/white images. The white part is colorized, too. This is wrong, expected is: black gets a color, white keeps white or gets transparent. There should be a way to reliably colorize just the grey or black parts of an image. |
|
|
I have noticed this shift in color of a colorized photo. It is see in the exported pdf. In the options for creating a pdf the is the option for intended output. Screen is normal but printer results in the shift of colors to much darker and the details in the darkened area are gone. I tried to upload a image of the results. I am not totally sure what all the differences are between screen and printer but it does have to do with cmyk. |
|
|
Hi StefanM, Hi kentonr I see what you mean but still, I believe the colorise fuction works just how it should and the result is expected. There is often not much "pure" white in a picture (shade of 0%). Hence, a conversion to other than black will affect the perceptible contrast between highlights and shadows. Where light shades of black may trick the eyes and seem to be white, a light shade of magenta will look pink. In fact, the colorize function will turn the chosen color to the shades as they are in the picture. One way to check that is to take a black & white pic, make a duplicate of it and colorise one of them. Output to PDF and check the shades for each picture in the same area. Test with as many areas as needed. If the shades are not identical, there is a bug and should be fixed. |
|
|
we should be able to create color from a color + tint % and use this new color to mix it with another just like we do with primaries |
|
|
I went back and tried a few more things.The bug is if you start with a rgb image and colorize it (even to grayscale it) the conversion to cmyk is darkened and details are lost. So wjen "output intended for" is selected to printer the rgb doesn't work, but a grayscale image created in gimp colorizes normal. |
|
|
I am not sure how to upload a image showing this difference in pdf processing. Some time the upload says successful (but nothing shows up in the list of files) other times it fails uploading. But there is a difference in a RGB image that is grayscaled or colorized in image effects and then exported with "intended output for" = "printer" |
|
|
Let’s keep the topic on spot colors, which is what the primary post is about. This discussion is obviously about printing. I see not much use of spot colors if there is no print in the end. There are *many* ways of turning a color image to a greyscale image. See this page for more on that topic: http://www.gimp.org/tutorials/Color2BW/ The result will then depend upon the chosen method. Once we’re satisfied with the result, and this can be tested against a proof we believe to be just as close to the final output as possible, we can now start acting upon the change black to spot, and proof again. That means there will be a specific plate for this color at output. The data for the image should not differ, wether the image is intended to be part of the black plate or part of a spot color. It's the same image, with the same values, dot for dot or pixel for pixel. Considering the final output being a spot and not black, it's then advisable to make some testing on the grayscale conversion method, to see what fits best the given image once it's going to be output in a given color. How it *appears* to the eye is a matter of perception, not a matter of data. If the spot color is light — say yellow — you can bet that the contrasts are going to be difficult to perceive for any human eye. On the other hand, a dark color will tend to keep the perceptible contrast that you see when the image is printed in black ink. |
|
|
Yes I agree with you that spot colors need to be printable, but currently 1.4.1 they are still converted to cmyk (as mentioned in bug 0003743). Maybe mapping them through channels would work for colorize does work to print separations when the color in colorize is set to magenta. This does result in a "black" plate for the spot color. And the darkening or incorrect conversion I have been seeing in the RGB to CMYK conversion is only in pdfs. (not in separations). So maybe this does need to be another bug report. |
- Anonymous
