[Small Look-Up Table(s)]

Of course, if we wanted to, we could forego the first step (the conversion to grayscale) and just replace the black ink with blue and the paper with light orange as before, but still keep the rest of color inks. The result would be probably the simplest LUT I have posted here, namely modrina.cube,

[Small Look-Up Table(s)]

The opposite of the above is dark azure ink and off-white orangish paper. In other words, orazur.cube.

[Small Look-Up Table(s)]

As nice as a monochrome image may be, the duotones are more interesting (in my opinion, anyway). We could even consider a plain old grayscale image analogous to a duotone. I mean, what happens when we send a grayscale to an inkjet printer? There is no gray ink (at least not in most of them), yet they are perfectly capable of producing a grayscale image. We can think of any shade of gray as a mixture of some amount of black with some amount of white (in fact, black is just the darkest gray an imaging system can produce and white is just the brightest gray a paper manufacturer can bleach up).

The inkjet printer uses translucent ink. To spit out black, it just gives out as much of the black ink as is necessary to reduce the translucency to 0 (in theory). And to give us white, it keeps the ink inside the cartridge. For anything in-between, it sprays out less and less ink for the lighter and lighter gray.

To me, then, to produce a duotone, we first convert the image to some kind of gray, then print it with a different color ink on a different color paper, preferably using two complimentary colors.

Now, to do it in software, we just have to mix the two colors, replacing black with the “ink” color and substituting the “paper” color for white and of course mixing them as needed for the different shades of “gray”. Mathematically, all that is three equations with three variables, something easily done by a matrix. Which we can convert to a simple LUT, such as galazery.cube,

In this case, the “ink” is dark red of sorts, while the “paper” is just slightly off-white toward a light-blue of sorts. Or to be more precise, the “ink” is dark orange and the “paper” is very light azure. And just about every pixel is a mixture of the two.

[Small Look-Up Table(s)]

For some strange reason, people often refer to grayscale images as monochrome. They are not, they are achromatic, i.e., lacking any color. An actual monochrome image has one color, albeit of varying brightness. Just like this one,

Made with erythron.cube.

[Small Look-Up Table(s)]

I also made a palette from the above, Red and Cyan.afpalette.

[Small Look-Up Table(s)]

I made another one of those red/green LUTs, but I combined it with several others. The result was interesting but very contrasty. So I normalized it in such a way that while it reduces the colors to just red and cyan  (which often looks more green even if it is not) but does not change the contrast of the image. I named it MQR.cube . Here is what it does:

Then I thought, if I can normalize it, I now know how to stretch it (to increase the contrast) and shrink it (to decrease the contrast) while keeping the exact same color effect.

So first I added 6% contrast (3% at the black end and 3% on the white end) and result was MQT.cube:

And then I went in the opposite direction and made one that decreased the contrast by 6% percent instead, MQU.cube:

So now we have the same effect thrice, just with an unchanged contrast (MQR.cube), with an increased contrast (MQT.cube), and with a decreased contrast (MQU.cube).

[Small Look-Up Table(s)]

As a strong believer in less is more, I’m trying to make these clearly noticeable on the histogram but barely noticeable in the image. And I always start by asking, what would happen if... I also try to answer the question before I actually make the LUT (I have been doing it long enough that my predictions generally pan out. Just like this one from two days ago.

This LUT only changes two numbers from what the Identity LUT is (the LUT that changes nothing). And both numbers are in the black vertex, which I consider the vertex that affects the effect the most. One is in its blue channel, which it lowers by 5% (from 0 to -0.05). The other is its green channel, which it lowers by 2.5% (from 0 to -0.025). I count the percentages from the fact that all values in the Identity LUT are either 0 or 1, so the maximum stretch in the default is 1.

Anyway, my prediction was it would make the result just a wee bit more orange. Those two percentages total 7.5%. And since the Simple LUTs I make have eight vertices (black, blue, green, cyan, red, magenta, yellow, white, in that order), each of which has three channels (red, green, blue) for a total of 8*3=24 values, the total per-channel change is 7.5% / 24 = 0.3125%, i.e., the subtlety of 99.6875% (based on the assumption that the Identity LUT is 100% subtle, as it changes nothing).

I called this one russet.cube, and here is what it looks like:

[Small Look-Up Table(s)]

8 hours ago, jmwellborn said:

I like the way that it very delicately alters the green (or is it the yellow?) in everything, so that other colors appear in all their glory!

Thanks. I’m glad you like it.

[Color Printer Test]

I needed to test my printer’s handling of color, so I created a vector graphic in SVG, then imported it to Affinity Designer and used it to send the patterns to the printer. And now I know I still need to clean the printer heads.

I figured others might find it useful, so I I uploaded it to the web for anyone to get. It is still in the original SVG, but you can open it in AD as mentioned above. This is what it looks like:

[Vector Ice Cream toon - AD]

12 hours ago, digital_wampa said:

Not to be selfish but you guys really need to create affinity animator! 

Nothing selfish about it. It would help a lot of us.

[Small Look-Up Table(s)]

Most of the time I make these by sampling the color of someone’s skin or hair and then making sure it looks good. I have a set of photographs that I test these on. And none of the LUTs has ever worked right on all of them.

Until just now.

I had a photograph of someone playing a guitar and was about to sample the usual. But the guitar attracted my attention. It was of a fairly dark brown color, darker than your average guitar. So I sampled the guitar’s color and did my usual manipulation with that. Definitely a special effect I thought. After all, it is not derived from the color of anyone’s skin or hair.

I named it gitarista.cube(gitarista means guitar player in Slovak) and ran the usual test with the usual set of pictures. It blew my mind. Every single one of those pictures, including those improperly exposed or poorly manipulated by their authors, looked good. It is a very subtle effect but probably my best yet (so says my worst critic, me).

[SVG Clip Art]

And I just drew a long face in Affinity Designer, exported it to SVG, edited that by hand (Affinity really bloats its SVG exports), and uploaded it to OpenClipart.

[ocean daydream + river]

Very nice. Thanks for sharing.

[vector mandala art]

7 hours ago, AfdolGravt said:

any suggest to improve my skill?

Just keep doing it.

[Small Look-Up Table(s)]

Haven’t done a LUT lately, so here is aurelia.cube , which I have created just now.

[How can I find details about the Affinity Designer file format?]

You’re welcome. And yes, I wish they would make the file format public.

[How can I find details about the Affinity Designer file format?]

Maybe this will answer your question, though it will disappoint you.

[Finally got the daisies nearly right!]

58 minutes ago, Antony parks said:

can I call it work in progress?

First rule of art: It is never finished. 👴

[Small Look-Up Table(s)]

35 minutes ago, jmwellborn said:

@AdamStanislav  Your video is delightful!   

Thanks. 😎 I had fun making it.

[Small Look-Up Table(s)]

Just so you know, I am still working on the software I used to create all these LUTs. Right now I used it to make a brief video using my software library to color-grade some free footage. Here is the footage: https://www.pexels.com/collections/flute-lessons-at-home-oz7ygln/ (the web site offers free photos and videos that we can use for just about anything, a great resource for the use with the Affinity products).

I also used my own fonts that I mentioned on this forum before, to make the impression of the title being etched in glass.

And here is the final video:

https://youtu.be/S6wEHVVN9Lg

[Artwork for greetings cards]

8 minutes ago, MikeW said:

In short, I used it as a "misery loves company" sort of way. 

I understood it perfectly. 😉

[Artwork for greetings cards]

Oh, for crying out loud, William. I am retired. I do not come here to assist people I have never met with their research projects. I only come here for Affinity products. For example, have you noticed they released 1.10.1 of their software today?

[Artwork for greetings cards]

5 hours ago, William Overington said:

I do not understand how that text relates to this discussion.

Then let’s drop it. 👴

[Artwork for greetings cards]

39 minutes ago, AdamStanislav said:

知者不言,言者不知

See how simple that is? Just four glyphs in different order:

知 ( = know) 者 ( = person) 不 ( = not) 言 ( = word)

Those who know, do not say. Those who say, do not know.

[Artwork for greetings cards]

4 hours ago, William Overington said:

I am willing to reconsider my position in the light of evidence.

知者不言,言者不知