@Vincent I feel that multiply blending mode makes the color in the same tone as the color below it. For example, in case of making a shadow on red, we draw black shape over it, and I change the opacity to 50%. But it looks little black + dark red. I would want it dark red. So I thought multiply would do the trick. But it didn't. I guess I should manually use dark red for shadow in this case.
@Vikas when you multiply RGB black on top of red the black covers the red entirely and the result is just black. White disappears and the result is red. Everything in between darkens the red. So you can use multiply black to make shadows but you need to decrease the opacity.
@Wolff See above. Now, if you switch from Multiply to Darken or Normal, all are same for Black over red. But all have slight difference for Blue over red. This is what I want to know why. Similar experience is with White over any other color.
It's all math (as I know you love). You can see the formulas on wikipedia.
Pure red is RGB(255,0,0) and pure blue is RGB(0,0,255).
If you divide the numbers by 255 all values goes from 0 to 1.
RGB(255,0,0) => RGB(1,0,0) and RGB(0,0,255) => RGB(0,0,1)
Multiply just multiplies the values of the 3 channels:
R: 1 * 0 = 0
G: 0 * 0 = 0
B: 0 * 1 = 0
So if a channel has the value 0 the resulting value will also be 0. And if a channel has the value 1 the resulting value will be unchanged. So pure black RGB(0,0,0) multiplied on another color always gives black and pure white RGB(1,1,1) never affects the other color.
Darken is different. It retains the lowest value from each channel.
According this, if I make the blue is perfect blue and red as perfect red, it should make the result black in Multiply mode? I just tried and it is not black, but kind of mix.
Yes. With Normal the top colro wins and the lower color is disregarded. With Darken the lowest value of each channels wins and the other are disregarded. With Multiply the values are multiplied and anything multiplied with 0 is 0.
The general formulas for blending RGB(r1,g1,b1) with RGB(r2,g2,b2) is:
If RGB(r1,g1,b1) = RGB(0,0,0) all three equations will yield RGB(0,0,0)
A big difference between these 3 equations is that Normal is the only one where the order of the colors matters.
All this said, when I work with images it's a much more intuitive thing where I play around with blend modes. I don't think about this math every time I use Photoshop.
@Wolff Yes. It will be black. But when you decrease opacity, it behaves differently than blue. For blue, the Multiplied and Normal have different views. But in case of Black, it doesn't change. Remains constant. I just can't understand it.
@Vikas Ha. That's a good question. Strictly speaking black is the total lack of color and white is all colors. But in daily speech I would call them colors.
Someone: "What color is you new car?" Graphic designer: "It's not really a color."
OK, sorry. I mean this is often a clash between web and print designers. On a screen you kinda start out with RGB black (all pixels turned off) and in print you start out with CMYK white (no ink applied).
@Wolff Ah sure, probably helps for management... I did it all on one layer for speed typically, I used gradients a lot so switching layers would have been a hassle
The nature of the work is different if you're working on books for artists :) In my case, it was "Here's the news, have some images, we're going on air in 15 minuntes" lol
@Wolff Yeah I guess they didn't think I would spot it since the assignment was to recreate the thing as closely as possible
@curious Yeah i understand. It's good to work non-destructively but sometimes it's liberating to just paint away without a thought of being able to revert.
@curious When I started at the print house about 12 years ago we still only had a CTF machine (computer to film). So to make offset print plates I had to first make a film, take a plate from a drawer, place it in a exposure cabinet, place the film on top, start the suction device, smooth out air bubbles, close the cabinet, turn on the light for a certain amount of seconds, place the film in a drawer, run the plate through the fixating machine with chemicals, place the plate on a holder,
check the plates for errors with a giant magnifying glass, remove hair and dust exposed to the plate with a special pen, if too many errors: repeat.
Back then as now we need to first make the imposition, but when that's done I turn on the CTP (computer to plate) and press "go" on the server. When the machine is out of plates I have to fill it again. Basically like a giant ordinary printer with aluminium plates instead of paper.
Looks similar to this:
No chemicals anymore. The image is burnt on the plate with laser and the excess material is washed off in the printing press.
@curious You can't actually because the image is so faint that you can barely see it.
The plates are covered with tiny teflon balls which are burnt onto the plate where the laser hits. But the area not hit by the laser is still on the plate and is washed off in the printing machine. So the printer has to check as he prints. But there are seldom errors nowadays. Once a year there might be a bad batch of plates. Once there was a guy outside drilling away some concrete and we got diagonal stripes on the print.
(Sorry to fill you with random unrequested info - just wanted to say that a fun thing about our generation is that we have experienced the pre-digital era while also (partly) understanding the digital era)