Could you post lores of some the images you want to combine thusly?
Put each of the images on individual layers, and set the Blending mode of all but the bottom to Screen. Then, for each of the layers, go to the Channels panel, and fill the colors you don't want with black.
In this screen shot, Layer 1 is a copy of your image, which I made three copies of, and then did as I described.
Hey thanks-I think the issue with this for my image is it's grey scale and the contrast thresholds are not as high. In this image her silhouette is darker. HOWEVER, then you see the whites of her eyes so maybe those areas were masked and painted back in?
Thanks for the great details though-much appreciated!
I did try something similar but I was still acheiving the purples and less saturated greens and reds, BUT I will try duplicating and then just adding a color overlay.
Also, the image I provided the silhouette is blackish (or gray scale) figure. WIth the gradient masks, the images are only the colored gradient maps. When you try to blend the B&W image with the R,G,B channels (in my case I just used the channels) the colors are less saturated and really show the greys creating a washed out look.
Then when I try to paint or create create an alpha channel for hiding some of those areas it becomes an abrupt cut off.
I will post some pictures shortly to see if you have any other ideas.
Thanks so much!
This is exactly what I was looking for. Thanks so much! Just two more questiond...
I attached my grey scale image-well a portion of it.
1) When I adjust the contrast to for the black and whites I lose a lot of detail or am unable to effectively adjust the gray tones so that the color can be seen after adjusting the channels.
For instance, the ideal image is not muddied like mine, but incredibly saturated.
2) I also attached another image from the same source which shows another idea of duplicating the channel and selecting the part (I think)-you will see the image is only colored unless blended. Do you think this was used some how for the technicolor?
3) Also-in your example it shows you taking the original image which is already technicolored and then making copies and following the procedure. But what did the original image look like to achieve the end result of the techni color? B+W I would think, but again mine is not satured (and other blends and saturations do not help.)
Starting with question #3:
Took the image you posted and transformed the Green and Blue channels to match the size and position of the Red. In the first, the background added from the transfomations is Black, and in the second one it is White.
Here, I took your image, used a reather drastic Curve to bring out tonality in the face, and then Transformed the Green and Blue channels.
#2 - To better understand what is going on here, load the image into Photoshop, and look at each of the channels in isolation. You will also notice that there is a fair amount of "bleed" or "crosstalk" between the channels, so the original creator was doing something slightly different to what I did.
you rock Semaphoric
Actually right after I sent this I noticed the channels in the original image. I will play with it tonight and hopefully figure it out with your additional knowledge.
Sem what is cross talk or bleed? If you google those key words with photography nothing really identifies the terms. Is it the same idea that is used in audio?
Ok starting out I was along the same path you described. However, I went back into the original image and studied the channels, but am very confused now.
1) I can not transform the B/G channel to match the size of the red without making a copy. When I make the copy I can not turn the original composite channel on without seeing the color adjustment in the copy channel.
2) When you indicate you fill
the background added from the transfomations is Black, and in the second one it is White do you mean filling the blue channel black and green channel white?
3) How did you seperate the channels into layers showing the RGB? When I used my own personal image that was different because I had 3 seperate images. You took 1....
4) Are the high pass lines and eyes I'm seeing in the image the cross talk you were referring to?
Sorry-I just really enjoy learning and right as I was getting it I completely confused myself.
Thanks so much for your great help!
ha ha I figured that out shortly before you responded-second time I have forgotten that
As Noel says, Select > All. and then you can transform individual channels. If you transform it smaller than the original, the surrounding area will be filled with your current background color value for the selected channels. White is 255 for all channels, and black is zero for all channels.
And yes, by bleed and crosstalk I was refering to the ghostly traces of eyes and whatnot from othe channels.
Ok one more question and then I'll stop - although I did finally figure out my mistake
How does one achieve the 'bleed and crosstalk'? Do photographers use these terms and are they achieved through post processing or is set up just as important?
Again to me it looks as though a sharpening technique was performed maybe by using a high pass and then masked and blended, but....looking at the details I know there is a lot of missing.
1 person found this helpful
Well, for that, you would have to ask the person who created the images. I think I see traces of a brush stroke texture as well,
Bleed and crosstalk are more audio terms, but the same concept: where information from one channel appears faintly in another (usually unintended). But you can also do it intentionally, via the Channel Mixer, Apply Image, or Calculations.
1 person found this helpful
How does one achieve the 'bleed and crosstalk'?
Though I wouldn't say "achieve" is exactly the right word ("have happen unexpectedly" seems more appropriate), it seems to me any operation on the image after the channels have been "separated" that affects pixel values where a pixel's color is taken as a whole might do it. For example, sharpening the image with Smart Sharpen, or adjusting the color saturation, or even just compressing it as a JPEG...
Ah ok I haven't heard those terms but makes sense now