# Simplified "spectral mixing" or hue transition

**URL:** <https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274>\
**Category:** Feature Requests\
**Tags:** for-filters, for-brush-engine, for-color-mgmt, for-layers\
**Created:** [September 10, 2026, 9:55pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274 "2026-09-10T21:55:06Z")\
**Posts on this page:** 17\
**Page:** 1

<div class="post-metadata">

**Author:** ![AfterCrow](https://krita-artists.org/user_avatar/krita-artists.org/aftercrow/32/241270_2.png) [@AfterCrow](https://krita-artists.org/u/AfterCrow)\
**Post date:** [September 10, 2026, 9:55pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/1 "2026-09-10T21:55:07Z")

</div>

First of all, thank you for making and maintaining Krita. 🙏

I’m looking for a way to blend/mix colors in such a way that the “gradient” between them includes the intermediate hues _along the color circle_.

I’m not asking for “realistic simulation” of color pigments, I know that’s been discussed in the past.

The picture below shows an example of what I want to achieve. For simplicity the example covers smudging colors on a single layer, but a blending mode for brushes or layers could also work for me.

 ![Pigment_blend_example](https://krita-artists.org/uploads/default/original/3X/4/4/443085d55b21c3998164a080e3204496dbeea845.png)

I’ve done some testing and noticed the **MyPaint** brush engine has this field “_Change Color Hue_”, which is able to shift the hue in a clockwise or anti-clockwise direction (on the color wheel), even for smudge brushes.

With that function set to Stroke as input, I can vary the hue over the distance of a stroke. However, it does not seem to respect the color I’m smudging into (or color I’m painting on top of), and also I need to be very cognizant of the Clockwise or Counter Clockwise direction of the hue shift.

 ![Pigment_mixing_example](https://krita-artists.org/uploads/default/original/3X/7/5/75277b15a95dabb010424302d20086b03f2db5ff.jpeg)

Notice in the above picture I was able to feign a transition through **Red** by using a smudge brush with “Change color hue”. The main issue is that this is very rigid in terms of respecting the color I’m mixing into (yellow in the case above). The brush settings in this case will only work from magenta into yellow, not the other way (and it wouldn’t work at another resolution or size because then I’d have to change smudge parameters in the brush preset).

 ![Overmixing_CMYK](https://krita-artists.org/uploads/default/original/3X/2/9/29baf418a8dd590774efac1c3ca5af61ce93d482.jpeg)

In the picture above, my smudge tool is shifting the hue “too far” along the color circle (all the way into cyan, when it should stop at blue). This is something I want to avoid, and I just want to smudge from the Magenta into the Blue in this case (And stop at blue).

 ![Rainbow_Brush](https://krita-artists.org/uploads/default/original/3X/6/0/607054fb3ab34f0fcb311dc7e0910bafd93c1edc.jpeg)

The above picture shows a MyPaint brush with Change Color Hue, This one is not a Smudge brush but you can see it shifts the hue _along the color circle_, creating a “rainbow” gradient.

Is it possible to use this functionality to create a smudging tool (or other blending strategy) that can make the type of transition gradient I’m looking for? I’m not asking for “Realistic pigment simulation”, just a hue shift along color circle. I suppose some would call it Color Temperature shift.

Thanks for reading. Please let me know if anything is unclear.

---

<div class="post-metadata">

**Author:** ![EyeOdin](https://krita-artists.org/user_avatar/krita-artists.org/eyeodin/32/2662_2.png) [@EyeOdin](https://krita-artists.org/u/EyeOdin)\
**Post date:** [September 12, 2026, 7:59am UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/2 "2026-09-12T07:59:35Z")

</div>

I know this is not it just a long work around. The pigment.o plugin does this math calculations with its mixer when in a rgb hue color space like hsv and others with clock wise and counter clock wise paths. With it you at least know the color in the mix and apply it and then force the rgb on the canvas to do it smaller steps. It does not do spectral mixing though.

---

<div class="post-metadata">

**Author:** ![RalMer](https://krita-artists.org/user_avatar/krita-artists.org/ralmer/32/247644_2.png) [@RalMer](https://krita-artists.org/u/RalMer)\
**Post date:** [September 12, 2026, 12:58pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/3 "2026-09-12T12:58:43Z")

</div>

I don’t really think this is possible, I don’t even think what you want makes sense.

What you want seems to be just shifting the hue between two colors. Sounds very simple, but it completely is not. For example, what if you pick two colors on the opposite side of the circle, the complementarities? Say red and cyan directly on the left and right of the circle… should the shift pass through the lower half of the circle (the purples)? Or should it shift through upper half of the circle (yellows and greens)?

You might say that going through the lower half is the correct one obviously, but that is not my point, there are a huge number of colors that a complement of each other. You have to have something that could deal with all that.

And even if you dealt with that, or even just avoid it entirely by not picking colors on the opposite side of the circle, say you just choose color that is more or less on one side of the wheel, there are still lots of problems. Say you choose red and green. Based on the wheel, the hue should be shifting in the oranges and yellows… but that seems kinda wrong right? Red and green results to oranges and yellows? The other fraction of the circle doesn’t look right either? You have to deal with that and every possible combination.

So what you want isn’t simple at all. Simulating realistic pigment mixing would have been easier since there is an objective reality we could refer to and imitate. What you want is arbitrary, or at least, just the same “realistic simulation of color pigments” but with some arbitrary requirements from you like more brighter mixes than actual reality. What you want is not any simpler compared to realistic simulation of color pigments.

The best strategy that I could think of is for you to manually just paint the in-between colors then blend. So if you mix yellow and blue. You paint in yellow-green, green, and then blue-green between the yellow and blue (or at least just the green) and then blend the colors. This is kinda tedious but it would achieve the exact effect that you want.

---

<div class="post-metadata">

**Author:** ![AfterCrow](https://krita-artists.org/user_avatar/krita-artists.org/aftercrow/32/241270_2.png) [@AfterCrow](https://krita-artists.org/u/AfterCrow)\
**Post date:** [September 12, 2026, 1:49pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/4 "2026-09-12T13:49:25Z")

</div>

> [@RalMer](#):
>
> For example, what if you pick two colors on the opposite side of the circle, the complementarities?

Thank you for the comprehensive analysis. I didn’t go into detail on that, in fear that my original post was becoming too long. To answer the question, I would expect transitions between complimentaries to go through grey since they cancel each other out.  
The same could apply at a lesser extent to near-complimentaries, like the Red to Green transition you mentioned. It would go through a somewhat desaturated orange-green gradient. Or the user could simply turn off the function when blending red into green if necessary.

> [@RalMer](#):
>
> So what you want isn’t simple at all. Simulating realistic pigment mixing would have been easier since there is an objective reality we could refer to and imitate.

I’m not so sure that realistic simulation is _easier_, since apparently several people have built systems that achieve this and they never get officially implemented in Krita. I presume due to performance loss and complications related to the multiple channels required.  
But rest assure I have a lot of respect for the programmers that work on this and it’s not my intention to imply that what I’m asking for is easy/simple to implement. Perhaps I could have named the topic better but I wanted to distinguish it from the Realistic Pigment mixing, or Spectral Mixing discussion.

> [@RalMer](#):
>
> I don’t really think this is possible, I don’t even think what you want makes sense.

It’s funny you should say that, as I just had this weird thing occur to me. A few hours after I posted this Feature Request, in a totally different context, I learned of a new painting app called [Kala Paint](https://kalapaint.app/index.php). I nearly spilled my drink as I was scrolling through the features description and got to the point about **Hue Blending**.  
Not only does it achieve exactly what I describe above, the developer also describes the exact same reasoning as I had in mind (though much more eloquently than myself), and he even covers what you brought up about complimentaries.

The discussion of it being possible thus becomes moot, but maybe I should just buy Kala Paint instead.

---

<div class="post-metadata">

**Author:** ![wakocima](https://krita-artists.org/user_avatar/krita-artists.org/wakocima/32/217849_2.png) [@wakocima](https://krita-artists.org/u/wakocima)\
**Post date:** [September 12, 2026, 2:16pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/5 "2026-09-12T14:16:40Z")

</div>

This is an old topic that gets revived every so often, to the point where I feel like a few key things should be pointed out;

- no blending mode is perfect, but being stuck only in rgb really sucks, and I don’t want all my brushes on parallel mode [especially because it works awfully with the pixel engine]\
- a variation of hue blending is nice, and we should strive for more experimental ways to blend. It doesn’t matter if we end up with like, 5 new ways of blending color – I think it’s overall a boon to play around with the concept of “nicer” color transitions in digital, no matter what path you pick to get there
- spectral blending is fine. Linear blending is fine. CSP’s way of handling blending is fine. They’re all just different approaches to the above point of “nicer color transitions”
- again, the more the merrier. I feel like it’d be beneficial if artists get more options and aren’t stuck with something eg, as unreliable as mixbox-esque blends only
- if anyone talks about “pigment blending” one more time I feel like I’m going to explode, rip all my hair out, and simply cease existing. Because it mainly says “I know nothing about pigment”. There’s way too many parameters going into pigments irl, and you will not have a machine that could even scrape the surface of computing true pigment blending for at least a couple of years still. This is, genuinely, an impossible feat. If you can handle trying to compute even 10 variations of red, their reflectivity, pigment granule size, dilution, and medium, and how they would precisely blend with 10 variations of yellow pigments (which have their own innate parameters), and your computer won’t explode? That’s cool. Would be insane to see. Still not possible and no one’s come up with **any real pigment blending** in digital yet.
  - Sidenote for above: real pigment blending would be perhaps possible, but you would never be able to use it on a color wheel, and only from swatches. This is because trying to assign pigments to some arbitrary color wheel, when they behave very differently based on their properties & medium, is, again, impossible.

- this does not mean that creative blending solutions should be discarded. This just means that in digital, we can allow ourselves for weirdo blending projects that are, instead, _impossible in traditional art._ Kala’s hue blend comes to mind.

You should definitely also buy Kala. Just because it’s a very fun program to draw in… even if I only have used it without pressure for now [no linux release yet].

So, to directly respond:

> [@RalMer](#):
>
> So what you want isn’t simple at all.

This is blatantly untrue.

> [@RalMer](#):
>
> Simulating realistic pigment mixing would have been easier

Even mixbox is a false pigment mixing method. Using only 4 hyper-specific pigments, without the context of medium, particle size, so on and so forth, is not how pigments work in real life.

> [@RalMer](#):
>
> What you want is not any simpler compared to realistic simulation of color pigments.

Using Fancy Maths for more vivid color blends + hue shifts is actually easier than trying to fit the [whole pigment range](https://www.artiscreation.com/Color_index_names.html) into a blending mode. Feel free to also try & unify every variation of each pigment depending on pigment properties, as even artist supply brands with the same pigment rarely have the paint/crayon/pencil end up looking the same (when they have different brands).

> [@RalMer](#):
>
> The best strategy that I could think of is for you to manually just paint the in-between colors then blend. So if you mix yellow and blue. You paint in yellow-green, green, and then blue-green between the yellow and blue (or at least just the green) and then blend the colors. This is kinda tedious but it would achieve the exact effect that you want.

This is like recommending people only use parallel for all color smudge engine brushes, even when you really don’t want to be using paralell/ your brush would be better off using a different mode. I’ve had people tell me to use gradient maps (???) for better color blending, and, no, that does not solve anything either.

Digital art shouldn’t be a tedium, but a freeing medium. We just really like boxing ourselves into very… basic, old, and non-experimental spots, for some reason. And it’s fair to assume that not every artist will be up to the task of making their own DIY software, like Kala or OurPaint – otherwise we’d already have hundreds of art softs by now.

---

<div class="post-metadata">

**Author:** ![RalMer](https://krita-artists.org/user_avatar/krita-artists.org/ralmer/32/247644_2.png) [@RalMer](https://krita-artists.org/u/RalMer)\
**Post date:** [September 12, 2026, 2:18pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/6 "2026-09-12T14:18:51Z")

</div>

> [@AfterCrow](#):
>
> To answer the question, I would expect transitions between complimentaries to go through grey since they cancel each other out.

Then in this case, blue and yellow should produce grey. Your example with yellow and cyan should be correct… but that is not what you want right?

> [@AfterCrow](#):
>
> The same could apply at a lesser extent to near-complimentaries, like the Red to Green transition you mentioned. It would go through a somewhat desaturated orange-green gradient.

So in this case it should be de-saturated? What if we mix colors that are just similarly distanced as red and green, like yellow and magenta… then the mixture should be de-saturated reds?

> [@AfterCrow](#):
>
> I’m not so sure that realistic simulation is _easier_, since apparently several people have built systems that achieve this and they never get officially implemented in Krita. I presume due to performance loss and complications related to the multiple channels required.

The point is that, it is not easy. What you want, and I’m even more certain of it now, is the exact thing as realistic simulation but just with more brighter mixes.

> [@AfterCrow](#):
>
> Not only does it achieve exactly what I describe above, the developer also describes the exact same reasoning as I had in mind (though much more eloquently than myself), and he even covers what you brought up about complimentaries.

No, it does not achieve what you describe. Even from their demonstration in their site, yellow and blue mixes to a grayish green. The demonstration was just very deceptive because it went very fast and it uses very desaturated colors in the first place.

> [@AfterCrow](#):
>
> The discussion of it being possible thus becomes moot, but maybe I should just buy Kala Paint instead.

I’m happy that you find what you are looking for.😃

---

<div class="post-metadata">

**Author:** ![EyeOdin](https://krita-artists.org/user_avatar/krita-artists.org/eyeodin/32/2662_2.png) [@EyeOdin](https://krita-artists.org/u/EyeOdin)\
**Post date:** [September 12, 2026, 2:36pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/7 "2026-09-12T14:36:05Z")

</div>

Well I am not gonna argue with you guys.

but considering the complementary situation you can shift your color 1 degree and choose either way, it is not a big problem on that one. Code wise I just choose a path and it was done unless you want a hue shift for complementary to be grey or something. but that would be weirder.

Either way I did what I could to help with this situation and it’s an old topic either way. I am gonna go eat lunch.

---

<div class="post-metadata">

**Author:** ![RalMer](https://krita-artists.org/user_avatar/krita-artists.org/ralmer/32/247644_2.png) [@RalMer](https://krita-artists.org/u/RalMer)\
**Post date:** [September 12, 2026, 2:48pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/8 "2026-09-12T14:48:45Z")

</div>

> [@wakocima](#):
>
> This is blatantly untrue.

Your argument is?

> [@wakocima](#):
>
> Even mixbox is a false pigment mixing method. Using only 4 hyper-specific pigments, without the context of medium, particle size, so on and so forth, is not how pigments work in real life.

Uhh.. I don’t actually mean physical pigments, just the colors. It the color mixing that they want, not the every pigment effects. I might have been very loose with the terms here. I don’t actually know what to call it.

Though, mixbox seems to be exactly what they want. Don’t know what that is and but that might something they ask to implement.

> [@wakocima](#):
>
> Using Fancy Maths for more vivid color blends + hue shifts is actually easier than trying to fit the [whole pigment range](https://www.artiscreation.com/Color_index_names.html) into a blending mode.

This is a total misunderstanding.

> [@wakocima](#):
>
> This is like recommending people only use parallel for all color smudge engine brushes, even when you really don’t want to be using paralell/ your brush would be better off using a different mode. I’ve had people tell me to use gradient maps (???) for better color blending, and, no, that does not solve anything either.

Huh? If they and you don’t want to do it, then don’t do it. Find something else. If you don’t want to use gradient maps or it doesn’t work for you, well no one is forcing you to use it. Do you know what recommendation means?

Seriously, so we can only recommend stuff that works for you? How exactly could we know that? Maybe people should not bother recommending anything to you since now we know, this is your attitude towards them.

> [@wakocima](#):
>
> Digital art shouldn’t be a tedium, but a freeing medium.

Why do you get to define what it should or should be not? What do you even mean by this? This is nothing but a meaningless platitude.

> [@wakocima](#):
>
> We just really like boxing ourselves into very… basic, old, and non-experimental spots, for some reason.

Easy to say… the doing is the problem.

> [@wakocima](#):
>
> And it’s fair to assume that not every artist will be up to the task of making their own DIY software, like Kala or OurPaint – otherwise we’d already have hundreds of art softs by now.

This is completely irrelevant. The discussion was if the exact behavior they want is easy to implement or not. You say it is easy but you didn’t actually explain that. I explained why despite how easy it seems, what he wants is actually incredibly complex. The devil is in the details.

You just say it is easy and blame the developers for “boxing ourselves into very… basic, old, and non-experimental spots, for some reason” as if they are just lazy and not trying or experimenting or anything. They just don’t want to do it somehow.

It is easy for you to talk and demand things but the thing is, implementing these things are not very easy, especially stuff that are vaguely and poorly defined. It would likely end in failure and wasting weeks or months if not years of work. What makes you think that your ideas would actually work and be worthy of hours and hours of work from the developers?

And even then, what makes you think they aren’t working on it? Why? Because you demand them now and not months or even years later as how things typically work?

---

<div class="post-metadata">

**Author:** ![RalMer](https://krita-artists.org/user_avatar/krita-artists.org/ralmer/32/247644_2.png) [@RalMer](https://krita-artists.org/u/RalMer)\
**Post date:** [September 12, 2026, 2:51pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/9 "2026-09-12T14:51:52Z")

</div>

> [@EyeOdin](#):
>
> but considering the complementary situation you can shift your color 1 degree and choose either way, it is not a big problem on that one. Code wise I just choose a path and it was done unless you want a hue shift for complementary to be grey or something. but that would be weirder.

Even you aren’t sure about it. Maybe you can try to implement it though and see if it works in a way that that satisfy AfterCrow.

---

<div class="post-metadata">

**Author:** ![wakocima](https://krita-artists.org/user_avatar/krita-artists.org/wakocima/32/217849_2.png) [@wakocima](https://krita-artists.org/u/wakocima)\
**Post date:** [September 12, 2026, 3:06pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/10 "2026-09-12T15:06:26Z")

</div>

> [@RalMer](#):
>
> No, it does not achieve what you describe. Even from their demonstration in their site, yellow and blue mixes to a grayish green. The demonstration was just very deceptive because it went very fast and it uses very desaturated colors in the first place.

It literally does, though. I’ve been sitting in Kala paint for maybe a week now? That is, literally, what happens if that’s the mode you select. You can easily adjust the hue variation with like, 2 sliders.

Kala has 3 ways of blending as of today, so nothing in the demo is “false”. I also don’t quite understand where you see muddied/gray combos, because the hue blend is … extremely vivid in the gifs?.. Either way.

There’s the default rgb, linear, and “hue” variations in blending. The dev also will not be doing “pigment mixing” due to all of the points I’ve made in my posts are literally the barrier to making pigment mixing real, everywhere. No software exists currently that can simulate pigment mixing in any real way. None of them behave like pigments, because they _can’t behave like pigments._

To quote the literal page on Kala:

> I’ve come up with a new system that still allows complements to cross greys while other colour pairs are almost magnetised away from the centre,meaning blue + red slightly tint _purple_ rather than the usual awkward digital pink-grey! I think it’s a mistake to try and model physical pigments as 4 different reds might mix in different ways, so this isn’t that. Instead it’s I think an even more intuitive-feeling mixing than the arbitrary digital default most artists are used to.

* * *

> [@RalMer](#):
>
> Your argument is?

There’s been multiple tries to do this in Krita. Just because they failed on the final stretch isn’t really something that makes this an impossible, sisyphean task.

> [@RalMer](#):
>
> Uhh.. I don’t actually mean physical pigments, just the colors. It the color mixing that they want, not the actual pigment effects. I might have been very loose with the terms here.

You were the one that suggested simulating realistic pigment mixing, so I’m replying to suggesting realistic pigment mixing.

> [@RalMer](#):
>
> The discussion was if the exact behavior they want is easy to implement or not. You say it is easy but you didn’t actually explain that. I explained why despite how easy it seems, what he wants is actually incredibly complex. You just say it is easy and blame the developers for “boxing ourselves into very… basic, old, and non-experimental spots, for some reason” as if they are just lazy and not trying or experimenting or anything. They just don’t want to do it somehow.

I’m not talking about the devs, but I wish any suggestion of a different, new blending mode, wasn’t constantly being shot down for no actual reason. This discussion keeps happening every few months, so there’s some point to it that makes artists hit a roadblock within Krita. Hell, _I’m_ guilty of it. I don’t use Krita for everything due to that painting roadblock.

> [@RalMer](#):
>
> The discussion was if the exact behavior they want is easy to implement or not.

If other DIY programs like OurPaint and Kala have implemented their own ways, it is fair to discuss other programs in this context, because people are hitting a roadblock in Krita and moving away from it because of not having an easy way to blend.

This is a sentiment I’ve seen in many places, including Kala discussions.

> [@RalMer](#):
>
> “boxing ourselves into very… basic, old, and non-experimental spots, for some reason”

Because this isn’t merely a Krita-centric problem, but Krita is software where we can at least have these types of discussions at all.

> [@RalMer](#):
>
> Why do you get to define what it should or should be not? What do you even mean by this? This is nothing but a meaningless platitude.

It is tedious to switch software every hour or so due to roadblocks. You might think it isn’t, but it really is when you’re just trying to paint.

No, telling me to simply move is not really a valid answer to this. It’s a form of shutting down discussion.

Also telling me that I’m only blaming devs is very tiring and that’s not what I’m doing. You’re free to think whatever you want, but don’t put words in my mouth. I also don’t like the sarcastic tone you’re giving me, so I think my input here is over.

* * *

> [@RalMer](#):
>
> The point is that, it is not easy. What you want, and I’m even more certain of it now, is the exact thing as realistic simulation but just with more brighter mixes.

Hue variation during mixing is not realistic sim with extra steps. I’ve no idea where you’re getting that from. If it’d be truly more impossible than pigment mixing… Kala wouldn’t have it. Hell, CSP wouldn’t have their own custom way of mixing, which… is why I’ve still been using CSP to this day.

> **Disclaimer about Krita in relation to this**
>
> To be clear – I’m not taking a dump on any part of Krita. I’ve been trying to comprehend the code base on my own, and I know how much of an undertaking adding _anything_ to the core is. But dismissing things like this as “not possible” is, frankly, coming from a very limited point of view. I try to stay up to date on a lot of software. Kala does their own way of blending, OurPaint has pseudo-pigment/spectral mixing, CSP has their own way of custom blends. None of their ways are fundamentally incorrect or wrong, they’re just different approaches to mixing in digital. I don’t have a particular favorite, but I do highly appreciate the variety of choices from the above three. Not having to constantly switch would also be appreciated.
> 
> Again, not trying to say “Krita should have all of these and more”, but being more open to new blending approaches makes artists very happy, and I can say that with confidence, because people interacting with Kala have been extremely happy – even if this way of blending isn’t “natural” or “real”, everyone’s still having fun with it, experimenting with it. That’s… kinda one of the points of doing art. Being hostile to these kinds of ideas locks us out of fun experiments, and for others, makes them drop the software. We _should_ have more of these kinds of examinations, and yes, I’m acutely aware the Krita team is small. I’m acutely aware how hard it can be to change anything in Krita.
> 
> This, however, doesn’t have to remove any core part of Krita and is _only an extension_. This isn’t, like, “remake the full UI because I personally don’t like it and think it’s ugly” kind of request. This isn’t also “implement pigments”, which by now we all know is not a realistic task in any kind of way.

If a demo from Kala is needed for its specific way of hue blending, there’s an inbuilt tool there for it showing how the calculations are done. No, I don’t mean “do it this way only and this is the only proper way”. I’m only saying that having more ways of blending colors is beneficial to everyone and not just “tradigital” painters. I don’t know what other way I have to keep repeating this to make myself clear enough that, no, I don’t hate the Krita devs (lolwut) and no, I’m not trying to force anyone into doing some arbitrary way of blending I’ve seen somewhere else (literally I am saying I don’t want this. What else do you expect me to say at this point).

That’s all from me for this thread, although I do wish people stopped ignoring how much more comfortable it would be to have more options in this area.

---

<div class="post-metadata">

**Author:** ![AfterCrow](https://krita-artists.org/user_avatar/krita-artists.org/aftercrow/32/241270_2.png) [@AfterCrow](https://krita-artists.org/u/AfterCrow)\
**Post date:** [September 12, 2026, 3:23pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/11 "2026-09-12T15:23:18Z")

</div>

> [@wakocima](#):
>
> - if anyone talks about “pigment blending” one more time I feel like I’m going to explode, rip all my hair out, and simply cease existing.

100% agreed, and I want to re-iterate that’s not what I’m asking for in this suggestion. 👍  
Core Krita developers seem to be on the same page and are aware that a perfect simulation of reality is tedious both from the programming side and from a user-perspective. Hopefully they will realize that hue shifting is still a desirable thing in digital art. Like you said, having to rely on gradient maps for everything is just not a valid solution. I mean yeah, you could do that but you could also paint a whole picture using a mouse instead of a wacom. Doesn’t mean it’s a good solution.

> [@wakocima](#):
>
> You should definitely also buy Kala. Just because it’s a very fun program to draw in… even if I only have used it without pressure for now [no linux release yet]

Wait, are you saying it doesn’t have tablet support or it doesn’t have tablet support just on Linux? I need pressure 😅

> [@RalMer](#):
>
> Then in this case, blue and yellow should produce grey. Your example with yellow and cyan should be correct… but that is not what you want right?

I am rapidly loosing track of the discussion between you and wakocima, but I wanted to address this point because it’s an important part of the discussion. These are my subjective opinions about the matter;

The thing is, you’re absolutely right in the case of Blue to Yellow, I don’t want full grey. I also don’t need a perfectly brilliant green. But this is a part of a separate, and still very controversial discussion in the art community in regards to what are the “true primary colors”. And my opinion (+experience) is that you can’t actually mix all colors with just three primaries (in the real world). The notion of mixing blue and yellow to get green is particularly outdated. It would be better in that case to mix Cyan and Yellow. There is, after all, a reason that CMYK is the standard for printing.

We can discuss the performance of real world pigments all day long, but at the end of the day, what I want is “just” a seamless hue shift along the color circle. And if you want, we can forget about bringing grey into the mix for complimentaries. Mixing yellow and deep blue really wasn’t something I expected to do to be honest.  
I agree any such system needs to be robust, And I agree it needs to know what to do (mathematically) if you give it a complimentary mix. But I also think Kala achieves it beautifully.

---

<div class="post-metadata">

**Author:** ![RalMer](https://krita-artists.org/user_avatar/krita-artists.org/ralmer/32/247644_2.png) [@RalMer](https://krita-artists.org/u/RalMer)\
**Post date:** [September 12, 2026, 3:36pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/12 "2026-09-12T15:36:11Z")

</div>

Uhh this argument is becoming unwieldy.

> [@wakocima](#):
>
> so nothing in the demo is “false”.

Not false, deceptive.

> [@wakocima](#):
>
> I also don’t quite understand where you see muddied/gray combos, because the hue blend is … extremely vivid in the gifs?.. Either way.

I didn’t say that it was not vivid? Even with the normal blending modes you create vivid mixes. My point is that his complaints also applies there though lesser. Like, the blues and yellow blends into grayish green instead of the vivid one that they want.

> [@wakocima](#):
>
> There’s been multiple tries to do this in Krita. Just because they failed on the final stretch isn’t really something that makes this an impossible, sisyphean task.

How is this an argument? So, it simple because multiple tries failed?

And when did difficult become sisyphean? Or is your argument is that anything that isn’t sisyphean is easy?

> [@wakocima](#):
>
> You were the one that suggested simulating realistic pigment mixing, so I’m replying to suggesting realistic pigment mixing.

And I just admitted that my wording is wrong. Sorry. Just don’t know how to describe it or what term to use. Maybe could offer something?

> [@wakocima](#):
>
> I’m not talking about the devs, but I wish any suggestion of a different, new blending mode, wasn’t constantly being shot down for no actual reason. This discussion keeps happening every few months, so there’s some point to it that makes artists hit a roadblock within Krita. Hell, _I’m_ guilty of it. I don’t use Krita for everything due to that painting roadblock.

But there are actual reasons. Like, I’m putting down some reasons specifically for this one.

And you even said that there are attempts that failed. So clearly they aren’t shooting things with no reason. They already have experience that say that this isn’t gonna work?

> [@wakocima](#):
>
> If other DIY programs like OurPaint and Kala have implemented their own ways, it is fair to discuss other programs in this context, because people are hitting a roadblock in Krita and moving away from it because of not having an easy way to blend.

We are talking about AfterCrows requests which is completely different.

It would be great if we know how they did it or we create an implementation that works, but as you said, they have tried many times and failed in all of it.

It would be great if you can provide an implementation that works. Or at least an idea that has a great of being successfully implemented.

> [@wakocima](#):
>
> Because this isn’t merely a Krita-centric problem, but Krita is software where we can at least have these types of discussions at all.

Then discuss it! Why would you all then point at the developers and label them that way?

> [@wakocima](#):
>
> It is tedious to switch software every hour or so due to roadblocks. You might think it isn’t, but it really is when you’re just trying to paint.
> 
> No, telling me to simply move is not really a valid answer to this. It’s a form of shutting down discussion.

Huh? What does this have to do with anything? You keep going into these irrelevant tangents.

> [@wakocima](#):
>
> Also telling me that I’m only blaming devs is very tiring and that’s not what I’m doing. You’re free to think whatever you want, but don’t put words in my mouth. I also don’t like the sarcastic tone you’re giving me, so I think my input here is over.

But that is exactly what you are saying. Just because your being fancy with your words doesn’t change the root meaning.

> [@wakocima](#):
>
> Hue variation during mixing is not realistic sim with extra steps. I’ve no idea where you’re getting that from. If it’d be truly more impossible than pigment mixing… Kala wouldn’t have it. Hell, CSP wouldn’t have their own custom way of mixing, which… is why I’ve still been using CSP to this day.

Seriously, I just explained what I mean by realistic mixing, that it is not actually accurately simulating physical pigments.

> [@wakocima](#):
>
> If a demo from Kala is needed for its specific way of hue blending, there’s an inbuilt tool there for it showing how the calculations are done.

Again, we are talking about AfterCrow’s suggestion. He didn’t say to copy that feature.

> [@wakocima](#):
>
> I’m only saying that having more ways of blending colors is beneficial to everyone and not just “tradigital” painters.

Seriously, again, we are talking if AfterCrows suggestion/idea would work. No one is arguing if more ways of blending colors is beneficial or not.

> [@wakocima](#):
>
> I don’t know what other way I have to keep repeating this to make myself clear enough that, no, I don’t hate the Krita devs (lolwut) and no, I’m not trying to force anyone into doing some arbitrary way of blending I’ve seen somewhere else (literally I am saying I don’t want this. What else do you expect me to say at this point).

I’m not saying you hate Krita Devs either.

Again, we are talking AfterCrow’s idea not other software’s implementation.

I expect you to not go into various tangents. And while you decry about me putting words in your mouth, then can do the favor and not imply that I’m siding or saying things I didn’t say.

> [@wakocima](#):
>
> That’s all from me for this thread, although I do wish people stopped ignoring how much more comfortable it would be to have more options in this area.

But I’m not ignoring it though? I’m just pointing out that AfterCrow’s idea is kinda half-baked and not in any way implementable in its present form. I would also want some new amazing blending modes but we do need to filter if an idea does or doesn’t work. If could work, we could try. If the idea has issues, we could discuss how to fix it.

---

<div class="post-metadata">

**Author:** ![RalMer](https://krita-artists.org/user_avatar/krita-artists.org/ralmer/32/247644_2.png) [@RalMer](https://krita-artists.org/u/RalMer)\
**Post date:** [September 12, 2026, 4:10pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/13 "2026-09-12T16:10:18Z")

</div>

> [@AfterCrow](#):
>
> Hopefully they will realize that hue shifting is still a desirable thing in digital art.

It is desirable for me too, but we are talking about your specific implementation. I’m pointing out where your idea fall short.

And hue shifting? Normal mode would do that. What I think you mean is a brighter mix, that is, shifting the saturation up from Normal mixing, as I explained before. And again, I’m pointing out, as before, where just “brightening” the mix could fail, like in the red-green mix. Just directly applying your idea would result to red-green mixing into bright yellows. That doesn’t seem right and you agree too. So, how are you going to essentially fix your idea so that it works in cases like mixing red and green? Direct implementation just doesn’t work at all.

> [@AfterCrow](#):
>
> Wait, are you saying it doesn’t have tablet support or it doesn’t have tablet support just on Linux? I need pressure 😅

Oh, what a shame! You could have find just the perfect tool for you.

> [@AfterCrow](#):
>
> The thing is, you’re absolutely right in the case of Blue to Yellow, I don’t want full grey. I also don’t need a perfectly brilliant green. But this is a part of a separate, and still very controversial discussion in the art community in regards to what are the “true primary colors”. And my opinion (+experience) is that you can’t actually mix all colors with just three primaries (in the real world). The notion of mixing blue and yellow to get green is particularly outdated. It would be better in that case to mix Cyan and Yellow. There is, after all, a reason that CMYK is the standard for printing.

It is not really about “true primary colors” or anything. I’m just asking at how exactly you want your blending mode to behave. You provided some description but as I’m pointing out, it fails to actually get what you want. A more detailed implementation would actually be helpful since vague descriptions just doesn’t work.

So how exactly could we make a blending mode that does what you want, or at least something close to it.

> [@AfterCrow](#):
>
> We can discuss the performance of real world pigments all day long, but at the end of the day, what I want is “just” a seamless hue shift along the color circle. And if you want, we can forget about bringing grey into the mix for complimentaries. Mixing yellow and deep blue really wasn’t something I expected to do to be honest.

And I made comment about that too. If you just avoid the colors on the opposite sides of the color circle, yes? That is when I pointed out that red and green aren’t on the opposite sides of the color circle and your kinda naive implementation would result to a mix of bright yellows, which is not what you want.

And even if you don’t expect mixing complementaries, there are just times when it happens. It is not impossible to have complementary colors side by side in an artwork; in fact, it actually happens a lot, especially in the digital medium (for example blue and yellow are complementary colors here and they find themselves side by a lot of times). So, for your blending mode to not act weird in this cases, you really need to define how your blending mode behaves in these cases.

As I said, it is not that simple as it seems.

> [@AfterCrow](#):
>
> I agree any such system needs to be robust, And I agree it needs to know what to do (mathematically) if you give it a complimentary mix. But I also think Kala achieves it beautifully.

Kala did provide a description, but I don’t think it works exactly that way as they say. While their implementation isn’t that perfect, there is clearly something more going on there. Maybe we can ask the Krita team to look into Kala and try to replicate what they did.

What do you mean by spectral rendering anyway? Wakocima was practically tearing my ear off for using “pigment mixing” for something I mean that isn’t actually physical pigment mixing including grain effects and diffusion and what not. I want to know if what I think you mean is what I actually think it is.

Wakocima though said something about mixbox? I look it up and, it seems to do exactly what you want… but it is one of those “simulating pigment mixing” thing so it is not exactly a “simplified” thing. They do have an online paint app. I tried it and the color mixing is awesome! The app is very limited though. The license seems to be incompatible with Krita thus it can’t be implemented in it. So sad.

I look around the forum and there actually seems to be stuff being done about this. They are trying to implement their own spectral mixing, but it seems to be kinda slow so we might have to wait for some time. But that is something to look forward to.

And I think you might also like to try the MyPaint software. Not the MyPaint brush engine in Krita but the MyPaint app. It has its own pigment mixing implementation. Try and see if you would like it: [MyPaint](https://www.mypaint.app/en/)

---

<div class="post-metadata">

**Author:** ![AfterCrow](https://krita-artists.org/user_avatar/krita-artists.org/aftercrow/32/241270_2.png) [@AfterCrow](https://krita-artists.org/u/AfterCrow)\
**Post date:** [September 12, 2026, 4:37pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/14 "2026-09-12T16:37:35Z")

</div>

> [@RalMer](#):
>
> Wakocima though said something about mixbox? I look it up and, it seems to do exactly what you want… but it is one of those “simulating pigment mixing” thing so it is not exactly a “simplified” thing. They do have an online paint app. I tried it and the color mixing is awesome! The app is very limited though. The license seems to be incompatible with Krita thus it can’t be implemented in it. So sad.
> 
> I look around the forum and there actually seems to be stuff being done about this. They are trying to implement spectral mixing, but it seems to be kinda slow so we might have to wait for some time.

Yeah so, this is what both I and Wakocima have been talking about all along. Yes, I do recommend reading up on those older threads because there is a long history of people trying to implement Pigment-esque color mixing, which again, is not what I’m asking for in this suggestion.

You can forget about MixBox because it’s a paid/licensed product that can’t be incorporated into Krita. The Rebelle painting program uses it to great success though.  
MixBox was based on Kubelka-Munk theory of color mixing, and some attempts were made to replicate this process in Krita. I believe [Ronald van Wijnen](https://krita-artists.org/t/paint-like-color-mixing-kubelka-munk/78156) had a working build of Krita that incorporated this, but there were some technical issues leading to it not being implemented in the long run.

**Edit:** Yes, I know about MyPaint. This is also why I mentioned in the OP that maybe we can utilize the MyPaint brush engine infrastructure to achieve the hue shift I’m talking about. In Krita.

---

<div class="post-metadata">

**Author:** ![RalMer](https://krita-artists.org/user_avatar/krita-artists.org/ralmer/32/247644_2.png) [@RalMer](https://krita-artists.org/u/RalMer)\
**Post date:** [September 12, 2026, 5:02pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/15 "2026-09-12T17:02:01Z")

</div>

Brush engine is kinda separate the Color mixing part? MyPaint has it own spectral rendering implementation but it seems Krita didn’t implement it for itself? Is something a problem here?

Based on your description on how your simplified spectral mixing works, I have this idea floating around in my head that could partially do it? My idea is essentially to have more “primary” colors. Color seems to be digitally stored in 3 primaries: red, green and blue. Why not have more? Why not use six (maybe yellow, cyan and magenta too?) or more for the mixing/blending calculations and then interpret the result into RGB? I think it would not be physically accurate at all (I don’t think we really need a physically accurate mixing anyway) but I think it would reduce the problems of just just having the 3 RGB colors (like yellow and blue mixing into gray).

I even get into thinking about how to code so I could try to implement it in Krita (maybe as an add-on) and see the results but I don’t really know if it would work or not. I don’t know if that would result to huge problems either (seems like it would use a lot of resources). I also don’t know if anyone tried it in Krita before or put an idea like that. But it is fun to think about.

Maybe I should try to code it, how to do it though? I really just want to know what the result is even if it fails.

---

<div class="post-metadata">

**Author:** ![Drawpile](https://krita-artists.org/user_avatar/krita-artists.org/drawpile/32/140367_2.png) [@Drawpile](https://krita-artists.org/u/Drawpile)\
**Post date:** [September 12, 2026, 6:17pm UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/16 "2026-09-12T18:17:48Z")

</div>

To re-rail this thread: the way “spectral mixing” works in MyPaint is (over-simplified) that it converts the 15 bit linear RGB representation into a floating point format with 10 channels, does the interpolation there and then converts it back to 15 bit linear RGB. As far as I can see, that’s also how the attempted implementations for Krita work, except they have to deal with Krita having a whole lot of different channel formats and they use a different formula with more channels (36 and 41 respectively, which makes them way slower.)

The advantage of doing it in HSV or similar would be speed. 3 is less than 10, 36 or 41 and the conversion formula is simpler. If the results are okay enough, it may be preferable for weaker devices and/or heavy brushes. Implementation-wise, I’d guess that just taking an existing spectral mixing implementation is easier than creating a new HSV mixing implementation from scratch and trying to figure out the undefined hue problem.

---

<div class="post-metadata">

**Author:** ![RalMer](https://krita-artists.org/user_avatar/krita-artists.org/ralmer/32/247644_2.png) [@RalMer](https://krita-artists.org/u/RalMer)\
**Post date:** [September 13, 2026, 1:23am UTC](https://krita-artists.org/t/simplified-spectral-mixing-or-hue-transition/190274/17 "2026-09-13T01:23:51Z")

</div>

Oh, the MyPaint implementation is pretty much the same as my idea but with more colors (channels). Never mind then.