Multiply blending mode

[sPretzel]

Hello,

I thought I understood how this worked but apparently I don't.

Illustrator's user guide states: How to edit artwork using transparency and blending modes in Illustrator

Multiply

Multiplies the base color by the blend color. The resulting color is always a darker color. Multiplying any color with black produces black. Multiplying any color with white leaves the color unchanged. The effect is similar to drawing on the page with multiple magic markers.

In CMYK space, if I have an object filled with 30 0 0 0 set to multiply on top of another object filled with 30 0 0 30, I get for the overlap 51 0 0 30 !! I am no longer sure what I was expecting but it would have probably been 30 0 0 30. My question is:

1- how does "Multiplies the base color by the blend color" actually operate on the numbers to get to 51 0 0 30?

I had to check this in Photoshop to get the resulting value. So my second question is:

2- how can I check the resulting colour value directly in Illustrator?

This is for CS4.

Thanks.

[Doug A Roberts]

i can tell you how it works in RGB.

each blending mode is an equation with the colour value of each channel converted to a figure between zero and 1 (zero is 0, 1 is 255)

so a yellow value of 100 is input as 0.39...

multiply, for example, is (background colour x foreground colour).

because both values will be less than 1, you end up with a lower number than you started with. thus, the colour value will be darker (closer to zero).

this doesn't work with the CMYK values, because they're percentages with the darkest value at 100%. I'm not sure, but i suspect the blend mode is actually doing the maths on an RGB conversion of the CMYK colour.

[Doug A Roberts]

either that, or there's some internal formula for the value of overprinting inks (which is the real-world analogue of Multiply in CMYK).

[c.pfaffenbichler]

either that, or there's some internal formula for the value of overprinting inks (which is the real-world analogue of Multiply in CMYK).

Equating »Multiple« with »Overprint« would be a mistake; in »Overprint« any Channel of the overprinting object that contains non-0 values takes precedence, those that contain 0 are ignored.

An object of the CMYK color 1/1/1/1 set to overprint on any other CMYK background (like 0/100/100/0) would result in 1/1/1/1.

[Doug A Roberts]

of course, my mistake. i don't often have a CMYK head screwed on.

[sPretzel]

Hi Doug,

I gave your suggestion a spin in Illsutrator, with the values I had provided. I assume this would depend on the colour profile.

CMYK: 30 0 0 0; RGB: 171 225 250; 0-1 scale (RGB/255): 0.67 0.88 0.98

CMYK: 30 0 0 30; RGB: 128 169 189; 0-1 scale: 0.50 0.66 0.74

Result when multiplying each 0-1 scale value together:

RGB (0-1 scale * 255): 86 149 185

AI converts this RGB value to CMYK (in my case):

CMYK: 68 31 16 0

It doesn't match the numerical result I observed (in Photoshop though, which might be using a different profile). However, the resulting colour does not match in Illustrator, visually.

How about question #2? 

[Doug A Roberts]

yeah that theory doesn't hold up, i've been checking myself.

for 2, the only ways i can think off off the top of my head are destructive, i.e. flattening transparency. not so useful.

[Ton Frederiks]

If you are interested in how Blend modes work, see the PDF Blend Modes Addendum

Blend modes addendum

It is not simple (at least for me) and it seems to convert the colors (RGB or CMYK) to some  intermediate HSL representation.

[Doug A Roberts]

I think that just applies to the Hue, Saturation, Colour, luminosity modes Ton. the other blend modes contain the usual formulas I've seen before.

the key bit here is this:

"The above formulas apply to RGB spaces. Blending in CMYK spaces (including

both DeviceCMYK and ICCBased calibrated CMYK spaces) is handled in the following

way:

• The C, M, and Y components are converted to their complementary R, G, and B

components in the usual way. The formulas above are applied to the RGB color

values. The results are converted back to C, M, and Y.

• For the K component, the result is the K component of Cb for the Hue, Saturation,

and Color blend modes; it is the K component of Cs for the Luminosity

blend mode."

So it does do an RGB conversion.

[sPretzel]

Hello Ton,

Thanks for this link. I had an inital read-through. The document is not very clear to me either. In particular, I am not sure if the section Doug quoted from page 6 applies to all blending modes or just the non-separable blending modes of Table 2. This is because for the K component, there is no reference to how separable blending modes (which include mutliply) are handled.

Also, there is a reference on page 1 to what appears to be some pictures, but I did not see those: "Plates 18 and 19 illustrate the resulting visual effects for RGB and CMYK colors, respectively".

It's still all clear as mud to me! Perhaps a good soul from Adobe or anybody else can walk us through, with numbers, and explain how Illustrator gets the result it does in the scenario of the original post.

[Ton Frederiks]

sPretzel  wrote

Also, there is a reference on page 1 to what appears to be some pictures, but I did not see those: "Plates 18 and 19 illustrate the resulting visual effects for RGB and CMYK colors, respectively".

It is an addendum, the plates are in the reference manual on page 1148:

PDF Reference, Sixth Edition, version 1.7

[c.pfaffenbichler]

In CMYK space, if I have an object filled with 30 0 0 0 set to multiply on top of another object filled with 30 0 0 30, I get for the overlap 51 0 0 30 !! I am no longer sure what I was expecting but it would have probably been 30 0 0 30.

You might be confusing the setting »Overprint« and the Blend Mode »Multiply«.

This is not the actual math but it might help illustrate nonetheless:

The base Layer’s 30% are the base, you multiplying 30% onto it means that

• 30% of 70% (the difference between the base’s 30% and 100%) get added to the 30%. 

(0,3 x 70) + 30 = 51

[angie_taylor]

Also, I seem to remember that with Multiply mode the top layer’s numbers are inverted before being multiplied with the bottom number?

[c.pfaffenbichler]

Both numbers are translated to a value between 0 and 1 with 0 being »dark« and 1 being »light«.

So for 30% on 30% that would be (1 - 0,7 x 0,7) x 100.

[ceyhun_akgun]

This is a very difficult issue to find with normal values. If you interpret your color and your work because of the ICC Profile you are using, a different value arises. When you multiply, the color values ​​in the blend vary. So, the upper color and the lower color do not mix at 100%. I'm having trouble with Layer Mode. I use manual methods while straightening the blend colors in pre-press preparation for packaging.

"This is your choice for a darkening effect. It multiplies the blend layer’s colour channel values with those of the base layer. Black in either layer gives black; white leaves the other layer’s colour unchanged. Lighter greys will give softer shadows, while darker greys will give deeper shadows. Why? Well, black has a value of 0, and white has a value of 1, and all other multiplied values are divided by 255. “Class, get your calculators out!”

[sPretzel]

Thank you very much c.pfaffenbichler and theangietaylor (the angie! for a reason) for your input. This really turned my (mis)understanding of multiply and overprint upside-down.

Let me summarise what I gathered from what you said. Please correct anywhere that I may be wrong.

MULTIPLY

-First of all, multiply applies channel to channel (each channel is independent). This is in line with the Adobe document referenced earlier by Ton (Blend Modes Addendum) which refers to multiply as a "separable" blend mode.

-To determine the result of multiply, there is no conversion to any other space. This too is in line with how I read the Adobe document.

-In CMYK, say the the Base object B has values Bc Bm By Bk, and the Top object T has values Tc Tm Ty Tk. T is set to multiply.

For the C channel (Cyan), the operation is:

To simplify, all values are first divided by 100 to get a C value between 0-1. The final result Rc is then muliplied by 100 to get a C value between 0-100.

Rc = Tc*(1-Bc) + Bc

The same operation applies to the M, Y, and K channels.

The Adobe User Guide states for multiply:

"The resulting color is always a darker color". Let's check. Rc = something positive + Bc >= Bc. True.

"Multiplying any color with black produces black". For black, Tk=1. Rc=Tc*(1-Bc) + Bc = some value. Likewise for the M and Y channels. For the K channel, Rk=Tk*(1-Bk)+Bk = 1*(1-Bk)+Bk= 1. So the result is Rc Rm Ry 100. True.

"The effect is similar to drawing on the page with multiple magic markers". That, I can't say!

OVERPRINT

-Overprint applies channel to channel (each channel is independent).

-To determine the result of overprint, there is no conversion to any other space.

-In CMYK, say the the Base object B has values Bc Bm By Bk, and the Top object T has values Tc Tm Ty Tk. T is set to overprint.

No need to alter values (divide by 100) here.

If Tc > Bc,  Rc = Tc,   else Rc = Bc

The result R would then be composed of values from both B and T, depending on how they compare for each channel.

Another notable aspect is that both effects are profile/colour space -independent. This brings me to cyhun-akgun's statement "When you multiply, the color values in the blend vary. So, the upper color and the lower color do not mix at 100%".

Is what ceyhun is observing a result of printer intervention (RIP), for example, to limit the amount of ink used? That would then not be related to multiply as such but to the printer. ceyhun, do you observe that in Illustrator too or simply in print?

Is there a document you (c.pfaffenbichler or angie) can point me to that fully describes these two operations?

[c.pfaffenbichler]

If Tc > Bc,  Rc = Tc,   else Rc = Bc

The result R would then be composed of values from both B and T, depending on how they compare for each channel.

No, it only depends on the top object’s values in each Channel.

If it is 0 the underlying content determines the output, if it is >0 it determines the output.

Rc = Tc*(1-Bc) + Bc

I think it should be

Rc = (1 - (1 - Tc/100) * (1 - Bc/100)) * 100

Example: 50%Cyan atop 25% Cyan

(1 - (1 - 0.5) * (1 - 0.25)) * 100 = 62.5

[sPretzel]

That's right, I forgot about your 1 1 1 1 overprint example.

So for overprint:

If Tc > 0, Rc = Tc else Rc = Bc

I find that counter-intuitive because effectively, in case Tc < Bc then Rc = Tc, the top layer channel is effectively "knocking out" the base layer channel. However, I do observe what you are saying in Illustrator! (Overprint Preview)

For multiply, the formula I proposed is the same as yours. Bear in mind that I do the divide/multiply by 100 prior and after to keep the formula light.

[c.pfaffenbichler]

I find that counter-intuitive because effectively, in case Tc < Bc then Rc = Tc, the top layer is effectively "knocking out" the base layer.

Well, I guess you are not alone in that but Overprinting is an old concept and in its way an efficient one, too, while Blend Modes (such as Multiply) actually involve a higher »computational work-load«.

Overprinting is like a simple boolean decision for each object and Channel, no further Math necessary.

I suppose it works as many people would expect it to work primarily when applied to 100% Spot or primary/secondary Colors.

With halftones the behaviour is not truly illogical just not what people are accustomed to from Blend Modes like »Darker« or »Multiply«.

[c.pfaffenbichler]

Is there a document you (c.pfaffenbichler or angie) can point me to that fully describes these two operations?

None comes to my mind.

But the two things are hugely different: The one is a print relevant setting, the other a Blend Mode, thus »Transparency«.

Which is prohibited in pdf X-1 and X-3 and therefore has to be flattened in those.

A »popular« overprinting issue are overprinting white (0/0/0/0) elements … if somebody views a pdf with a white overprinting text on a colored background in some application that does not honour overprint settings they can see the text.

But in print it would not be visible and that is already evident when viewing the pdf in Acrobat with »Simulate Overprinting«.

(Similarly in Illustrator if »Overprint Preview« is checked/unchecked in the Separations Preview Panel.)

[ceyhun_akgun]

In blending modes, MULTIPLY is the clearest what will be the result.

Multiply and Overprint use are mixed. They are not the same, they do not give the same result.

Apart from digital and offset printing, the small values in the color values for flexo and gravure printing are very important. Points that are never seen due to dot gain values can print out.

Photoshop does not display pixels below 1 in the info panel. But at 0.5% color value, 30% point gain is a very different color value. This is caused by the operations in the CURVE curve in RIP.

For all these reasons, I prefer to print out the blend modes flattened. All color values must be in my control.

[Ton Frederiks]

ceyhun_akgun  wrote

Multiply and Overprint use are mixed. They are not the same, they do not give the same result.

Multiply will work the same as Overprint when the Overprint rules are used.

If the Overprint color does not contain a percentage of the color below it, it will give you the same result as Multiply.

Most modern rips will treat Multiply in those cases the same as Overprint.

Where Multilpy differs; it will always give a darker color, even if the same Multiply color is used multiple times on top of each other or when the top Multiply color contains a percentage of the color below it.

Left two colums where Overprint and Multiply give the same result. The right two columns differ.

The AI file:

Adobe Creative Cloud 

[ceyhun_akgun]

Two different functions can give the same result in some cases.
This does not mean that one of the two functions is an alternative.
I share two different results in the same color I share. This is a natural result. It does not give the same result.
In graphic work we have to use very complicated photographs and textures.

[Ton Frederiks]

What are you exactly showing here?

What color is used on top of the (CMYK) image?

[ceyhun_akgun]

You can try it too.

Prepare two different color boxes to print Stunning and Upper Print on a photo. The color values ​​of these two boxes differ.

The color combination tables you have prepared have solid colors. It is suggested that we prepare this picture about the photograph, the results can not be reached.