Rendered at 21:18:13 GMT+0000 (Coordinated Universal Time) with Vercel.
S0und 4 days ago [-]
As an "expert viewer" of Baumgartner Restoration, this site usefulness is questionable. If you look at those color palettes, most of them brownish that is because of dirty & old oxidised varnish. These are not the intended look of these paintings. So these color palettes has nothing to do with those 3000 masters.
It was short but I really enjoyed this little thread this morning, added much color to my life!
> Starting in the Renaissance, artists made sculpture and architecture that exalted form over color, in homage to what they thought Greek and Roman art had looked like. In the eighteenth century, Johann Winckelmann, the German scholar who is often called the father of art history, contended that “the whiter the body is, the more beautiful it is,” and that “color contributes to beauty, but it is not beauty.” When the ancient Roman cities of Pompeii and Herculaneum were first excavated, in the mid-eighteenth century, Winckelmann saw some of their artifacts in Naples, and noticed color on them. But he found a way around that discomfiting observation, claiming that a statue of Artemis with red hair, red sandals, and a red quiver strap must have been not Greek but Etruscan—the product of an earlier civilization that was considered less sophisticated. He later concluded, however, that the Artemis probably was Greek. (It is now thought to be a Roman copy of a Greek original.) Østergaard and Brinkmann believe that Winckelmann’s thinking was evolving, and that he might eventually have embraced polychromy, had he not died in 1768, at the age of fifty, after being stabbed by a fellow-traveller at an inn in Trieste
Man, what a line. What a horror, this projection of opinion! From the "Father of Art History"! To rob the world so! I feel this way all the time, that anti-sentiment, that the pure marble world just stately and so is art and perfection, over the colors of the universe & it's possibility!
> "the whiter the body is, the more beautiful it is"
This should make your blood run cold, imo. A world locked in amber view of reality, static, sedate. Whew.
molf 4 days ago [-]
Agreed. I absolutely adore the idea of it! But all the brownish colours tell the same story.
For some additional context; many old pigments were not stable at all.
Is there enough color data left in the brown to correct it?
Or do you need to infer it based on location, budget, time, climate etc?
molf 4 days ago [-]
This specific painting was reinterpreted based on specific descriptions of the colours in a letter from the painter.
As far as I'm aware there is no way to know for sure what colours originally looked like, especially if the information is limited. There are so many variables, we can only guess.
tapland 4 days ago [-]
Its absolutely lossy and you'd have to know a lot about each piece to know how big the loss is.
ouli 4 days ago [-]
This is a very interesting perspective. I'd thought the muted, brownish colors in these paintings had to do with the quality and availability of pigments during that period.
mbivert 4 days ago [-]
There's most likely multiple aspects at play: high-chroma pigments were historically limited and/or expensive; varnish yellowed over time; pigments faded. The digitization process probably wasn't perfect as well (I'd expect modern scans should be fairly good though).
nekooooo 4 days ago [-]
hundreds of years of oxidation will make everything brown.
pzoln 4 days ago [-]
so is there a formula that can be automatically applied to restore the original colors? at least some reasonable approximation, based on the painting's age?
tudelo 4 days ago [-]
I seriously doubt it. Degradation would be in some part related to the conditions the painting was held in, which would be nearly impossible to backtrack outside of one-off case studies. Imagine a painting that was stuck in a room full of smoke -- or was put on some less than good backing paper/framing.
There has been some research on what causes degradation on paper/pigment but as far as I know much of it ends up as a mystery, a fact of time...
ouli 4 days ago [-]
I would like to know that formula also. this can be an interesting tool to reveal true colors of paintings as the painters intended...
9dev 4 days ago [-]
It’s the colours you will see today when looking at the paintings, however. Your point is valid, but even the somewhat "chromatically degraded" versions of many of these are gorgeous.
senexex 3 days ago [-]
An additional problem is that it looks like the colors are from the intersection of 3 random paintings.
The masters already picked their colors on one painting. If you randomly average them it is like a kid mixing finger paints.
You aren't going to get a better color scheme from 3 random masterworks.
ouli 3 days ago [-]
Actually I group artworks based on artist/ganre/style similar color palette and showcase up to 3 images from that group.
Daub 2 days ago [-]
A few points on how a painter's intentions do not always align with how we see their work many years later:
Generally, a painting is best viewed and photographed in as close a lighting environment as it was painted. I have seen many paintings 'blasted' by unnaturally bright gallery lights. There is a reason why a gallery lighting designer is a real job. However, in my experience effective gallery lighting designers are as rare as rocking horse shit.
It is true that the paint the artists applied many years ago will often bear little resemblance to that which we now see. This is less true of earth browns and very true of paintings done at the beginning of the pigment revolution, when wonderful colors were produced which were later discovered to be very 'fugitive'. However, the relative relationship between these paints remains more or less intact, and IMHO this is the most important factor in aesthetic evaluation.
Another factor is how a photo flattens such differences as rough vs smooth (and their consequent light reflection properties). In a Titian painting, huge areas are untouched rough red oxide primer on rough canvas, vs the slight gloss of oil paint. Importantly, old masters would often apply their lights as (smooth) thick paint and their dark as thin glazes (or scumbles) above a thin primer (red or green or yellow ochre or whatever). The frisson between these layers gave the darks their depth that they would otherwise have lacked. This is mastery of dynamic range at its finest. Googles art project photos comes close to capturing this. For an example, check out any portrait by Durer in Google art project.
camillomiller 4 days ago [-]
It‘s vibecoded slop that turned an idea that should have stayed a bad idea into poor execution, convincing the vibecoder that the idea is validated. The vibe coder has absolutely no domain specific knowledge to understand what you picked up in a second, yet no AI ever mentioned that, they didn’t learn, and instead reinforced their bias by producing another piece of forgettable slop.
Yes, I’m fun at parties.
sixtyj 4 days ago [-]
Honey, get used to it :)
But seriously. I am not against ai slops in general, people explore what’s possible to make. But I don’t really appreciate them at Show HN - there should be a new thread for ai-made or -assisted projects, or there should be a disclosure in project description.
camillomiller 3 days ago [-]
Agreed. I created plenty of sloppy but useful personal one-page tools with Claude myself. Faster than googling them most of the time. The problem is when you get convinced they‘re good enough to present them this way.
KevinMS 4 days ago [-]
> is because of dirty & old oxidised varnish
And the pigments fade. And even worse, they fade at various rates and some are almost completely gone.
DeathArrow 4 days ago [-]
So we can use AI or even deterministic algorithms to recover the original palette.
echoangle 3 days ago [-]
Is that a question? Because that doesn't follow at all. 2 different pigments with different colors could fade to the same color and be indistinguishable from looking at them now, without analyzing the chemical composition.
hparadiz 4 days ago [-]
I used to run an art social network 15 years ago that did this automatically with every piece of art uploaded and then it let you search for art by color that way.
Basically
$average = new Imagick( $file );
$average->quantizeImage( $numColors, Imagick::COLORSPACE_RGB, 0, false, false ); //Reduce the amount of colors to 10
$average->uniqueImageColors(); //Only save one pixel of each color
anthonycoslett 9 hours ago [-]
This is a gorgeous site and beautiful concept. I'm interested to learn what inspired you to create this contribution? I look forward to exploring it more especially as I decide what colors to paint the outside of my house as I attempt to "standout" in my New Orleans neighborhood.
woolion 3 days ago [-]
Not sure about the use for it, is it supposed to be design or painting?
I've thought about making this for a long time to help me with painting, but in that case I think to be useful you need a bit more ways to see the data -- mostly, the thing that is the most important is value. So to get something useful out of it you need a distribution of the hues conditioned value.
And for design, the problem is a bit different. You may have a good looking palette, but 'inverting it' for dark mode is not trivial, and neither are gradients, getting intermediate colors, or getting a shifted hue.
It's called inspiration so it's fulfilling its promise, I'm just curious what are your thoughts on these since you obviously thought a lot about it.
assimpleaspossi 3 days ago [-]
I see these palettes shown everywhere and see no value in them whatsoever. As a designer or painter, one needs to develop their own palette for what they are trying to create. So what difference does it make what someone else used no matter how famous? I just don't get it.
eudamoniac 3 days ago [-]
Kind of a stupid comment tbh. Taste requires breadth of experience. "As a creative, one needs to develop their own <creative thing/technique> for what they are trying to create. So what is the value in looking at how other creative things were made?"
woolion 3 days ago [-]
I don't think the question is stupid, because I kind of consider that to make use of a palette like this, you still need a lot of skills that if you do have, you can probably as easily do your own palette directly from a painting by yourself.
>As a creative, one needs to develop their own <creative thing/technique> for what they are trying to create. So what is the value in looking at how other creative things were made?
That's not the how of the process at all, that's the end result. The "how" for traditional art is completely different; it is taking pigments as a base rather than light, and the algebra of composition is not the same either.
ouli 3 days ago [-]
The intention of this site is to showcase time tested palettes (with empirical evidence) of greatest visual artists instead of listing endless sea of neon bright color palettes we see everywhere in popular palette sites like coolors or colorhunt. Consider it using like a color grading tool for you art/design project similar to cinematographers who use contain color combinations to bring emotions, mood to the movie shots to tell the story in more visually captivating fashion.
Theodores 4 days ago [-]
I am currently looking for colour palettes and this website is of interest to me.
Small snag, some UTF8 things are going on with some colour names, I am sure you know and have cursed accordingly.
I like OKLCH colours and the ability to mix them in interesting ways using CSS things.
This means I don't do hex codes for colours in CSS. I can translate though, however, soon some people will demand OKLCH, so you might as well add it in, trying to get it natural with the picker.
I appreciate the masters but I wonder how this would work using other sources, for example, Sunday newspaper supplements from the last century, and their glossy adverts, which were to a higher standard than what we get today.
ouli 4 days ago [-]
Adding OKLCH color codes will be a great addition. Thank you for your suggestion!
I am aware that Advertisement palettes mostly based on Alphonse Mucha work since I could not include more recent ad illustrations for copywrite reasons.
ouli 4 days ago [-]
As you suggested OKLCH colour option is added now to the site. Thanks again.
Theodores 3 days ago [-]
Mighty impressed! You have done well by sharing your project 'warts and all' on HN. Thanks for listening, just that UX matter of 'loading next palette and you are good to go!
ouli 3 days ago [-]
Thank you for kind words! The loading next palette feature is fixed also. You can check it whenever you have free time.
Theodores 2 days ago [-]
For my current pet project there is an abundance of sensible palettes, with the pet project being a home baking app for my uncle, who keeps meticulous records of so many things. The idea is that there is a 'baker's log' with recipes that can be amended until perfection is achieved.
I digress, back to HARSH criticism!!!
Playfair, Serif, in small sizes, I am not a fan. I can see why, and it works for H1 and H2, but, for those of us with elderly eyes, please can we have a sans-serif such as Inter, which does work on smaller sizes? Maybe A/B test that one, with the screen zoom level captured. My hunch is that I am not alone in having elderly eyesight and that people will be more engaged if they have an easier time of reading.
Next bit of harsh criticism - HTML standards. There does seem to be a use case for 'figure' and 'figcaption'. These work like the 2003 era 'div' and nobody cares about HTML, but why bother with HTML if everything is just divs nested in spans and divs?
In truth, 'div' is the element of last resort and not needed since we have joyful content sectioning elements such as 'section', 'article', 'aside' and much else. We also have a layout engine in CSS, as in CSS grid, and, in my experience, the 'div' is not helping. Code is just better when it is not a sea of divs, each with a gazillion class tags.
We have arrived at a situation where people just use 'div' because that was how grandpa wrote HTML when the iPhone came along and table layouts suddenly had to be 'responsive', as in bloated. We built whole departments around writing code badly with 'div' elements, with whole teams of micromanagers. On pet projects there is no need for any of this, plus you can do things like scoped CSS so there is no Firefox support, but also no need to torment oneself with CSS compiler things that create bloat and 'add to only' stylesheets.
But hey, none of this matters if everything works, which it does!
I also think you need some type of funnel to get people to the website in such a way that they come back. This might not matter to you, but, imagine you find this site, do you bookmark it, share a link with oneself, or what? There needs to be some type of hook, and bookmarks don't cut it these days. People want to Google Search for what they know.
One hook might be an upload feature, so you can upload many things. I like de Stihl artists from a century ago and some of them used paint that is still sold today. I also like old posters for British Raleigh bicycles and train posters from the period. The train posters were by commercial artists that made adverts, not art. Yet you still have some brilliant colours going on in their work.
Yes there are services for taking any image and making a palette from it, however, they haven't got the art. They can't do a 'goes with' feature, to show art that is in the user's colour scheme, from what they upload.
I am impressed by the colour naming, this is actually quite hard to write by oneself since you need to find something in a 3D search space. It seems to me that you have the hard stuff mostly solved but tinkering with the UI for stickiness is there to be done, ad nauseum.
Please don't take my HARSH criticisms unfairly, HN is the place for people to niggle, because some people like to cut the heads off others to feel taller, however, I am a genuine punter, and I have a bit more work to do before I get into the colour scheme, so I kind of need to be able to read things easily, remember where your website is and give it a proper go.
What might be useful is a palette export option in :root { CSS variables }, for example:
The vars could be named colours, where possible, so '--warm-caramel' and so on.
I still get the autoload of the next palette, mayb what you need instead is a good bit of doom scrolling, a.k.a. infinite scroll. That might be more engaging. I say this because my recipe app also needs infinite scroll for browsing purposes, so the user (my uncle) does not have to drill down to the recipes.
Daub 3 days ago [-]
I have a history of research in computational aesthetics with an emphasis on old master paintings. A few notes:
1. You have used the RGB hue wheel not the RYB hue wheel. The later defines the complementary pairs red/green, blue/orange, yellow/purple which are more useful for an artist when evaluating hue. To those who would say that the difference between RTGB and RYB complementaries is subjective, I would point out that the red/green, blue/orange, yellow/purple pairs have been known by artists since before Newton first expressed hue on a wheel.
2. I think it is fair to assume an artist has a consistent color pallet, but I'm not sure that a genre would.
3. User S0und has implied that paintings have changed too much from their point of creation for your work to be valid. For sure it is true that paintings have oxidized, accumulated multiple layers of dust-attracting varnish and been damaged by UV exposure. However, the relative difference between the colors they have used remains almost the same, and this is where the value of your work lays.
4. The key problem of visualizing the color of a painting is its dimensionality. A standard histogram does a super fine job at expressing the lightness value of an image. A radial hue histogram does the same for the hue/saturation. But there is no 2D graphical visualization that does both. Personally, I would go further and visualize separately the hue, saturation and lightness signatures of the artists. For most pre-modern painters, the lightness values closely mirrored their saturation values. Artists like Gericault and most of the Impressionists learned to separate them.
Great points!
1. I'll try to add RYB option for color harmony wheel
2. You are absolutely correct, there are no consistent palettes across genres, but I thought it will be interesting to see multitude of palettes artists use for the same genre
3. User S0und's observation about oxidized or decayed colors is very interesting, but even with muted, brownish colors these artworks still look amazing.
4. Any suggestion how can I add such color visualization across body of work of a given artist or a style?
Daub 2 days ago [-]
> Any suggestion how can I add such color visualization across body of work of a given artist or a style?
I think that the best visualization of the colors of a painting is by using the painting itself.
In class I demonstrate lightness by first desaturating the work (or copy pasting the L in Lab). I then do a controlled posterize on the image - basically a stepped curve in Photoshop. I try to isolate the dark, middle and light. These are relative values that can often manifest as lumps in the histogram. Painters tend to be very deliberate in the way they organize them. This page explains what I am getting at:
In my experience such posterizing is best done manually but AI might be able to do it.
Hue at saturation are more difficult for the simple reason that they are difficult to disengage from lightness.
Like lightness, saturation is generally organized according to low, middle and high. For most of art history, the saturation of a painting would closely follow its lightness. It was Gericault who separated them. Check out the saturation vs lightness of his Lobster painting for an example of this.
Hue is a beast. Sure, most paintings done before the impressionists are pretty unsaturated. But even Rembrandt would be careful to use a red brown against a green brown. Check out the Rembrandt image on this page to see this in action.
I think that a radial histogram is the best way to visualize hue. This would show not only the hues but there relationship to each other on the RYB wheel and also there quantity. There should be a cut-off point for hue that is visible. In our work we established a cut off - all hues with very low saturation were ignored.
Hope this helps.
ouli 3 days ago [-]
I love the website about color pigment history - it explains a lot about some of the color choices artists made in different time periods. Thank you for sharing the link.
saltypixel 2 days ago [-]
As a designer and creative director, I use Adobe Color and Pinterest for client branding/re-brand inspiration, or just straight to the pantone books. This is interesting to me as another tool. To expand the color variations, maybe get movements like Pop in there...maybe not 'masters', but would expand beyond the yellows and browns
Sorry for the inconvenience. The email works now. Regarding the article - I use similar ideas to extract colors form artwork images, only difference is I added color prevalence scale for each color and limited it to 10 colors per palette.
khazhoux 4 days ago [-]
ouli, I feel like the algorithm fails to capture the essential colors and their roles. For example, for Mondrian I expect to find a palette with white, black, red, blue, yellow. Instead, the palette is a big swatch of white, a big off-white, and then little strips of color, with some equally-weighted grays too.
For Monet, many of the paintings have an important color highlight (e.g., the orange sun), which isn't captured in any of the palettes.
Needs more tweaking.
code51 4 days ago [-]
Why do I only see Claude in this UI?
It seems Claude is picking up the very same color scheme for many webpage building requests.
ouli 4 days ago [-]
It is puzzling for me why you see these colors as Claude UI colors, since I carefully choose these colors as my main brand colors across all other art related sites (eg https://ouliart.com).
NetOpWibby 4 days ago [-]
Quite possible Claude/Anthropic likes your colors
ouli 4 days ago [-]
[flagged]
NetOpWibby 4 days ago [-]
:) Thank you!
trox 3 days ago [-]
As a traditional painter, I would love to see the individual pigments used, instead of dominant colors.
I know this is not straightforward, but maybe doable to some degree based on the range of hue/value/saturation and context?
realitysballs 4 days ago [-]
Why no contemporary/modern painters? First looked for Basquiat then Lichtenstein , nothing
ouli 4 days ago [-]
Most contemporary artists works are still under copyrght laws.
mock-possum 4 days ago [-]
The colours they used aren’t
Daub 2 days ago [-]
There are two copyrights at work: that of the artwork itself, and that of the photo thereof.
ouli 3 days ago [-]
I still need to access and download such copyrighted images in order to derive palette colors and showcase sample images for those palettes.
79a6ed87 4 days ago [-]
Hi, I like it! 2 things I noticed:
1) Aren't "Modernism" and "Modernismo" the same thing? I'm a Spanish speaker and from my POV they are
2) Selected "Naïve Art" style and it's broken (images not loading). Probably something to do with the diaeresis
ouli 4 days ago [-]
Both issues are fixed now. Yes "Modernism" and "Modernismo" are the same, only some Latin American artists have art style as Modernismo instead just Modernism.
afolkest 4 days ago [-]
Love it. I'll be using this on a weekly basis in my art practice.
Let me know if you ever create an API endpoint.
ouli 4 days ago [-]
Thank you for kind words! That was my exact intention to share empirically proven color palettes for artists and designers like you. Adding API endpoint is a great idea. I'll let you know when it will be ready.
This is great! Love the idea. You should send to some art programs, sure they would get a kick out of it. Also gives another use case outside of just digital design.
ouli 4 days ago [-]
Great suggestion, thank you! I'll try reaching out to some art schools to see if they'd find the site useful.
vardump 4 days ago [-]
Weird how similar many of those are to Commodore 64 palette.
ouli 4 days ago [-]
You are absolutely right, I did not realize how similar some of these palettes are to Commodore 64 palette
feedback -- painting and screen luminescence are fundamentally different. Secondly, the chase after "most famous" seems to show the motivation here.
AI generated arts, IMHO, is an economic F-U to the people who spend real years of their lives to train in arts, and to the people who spend more time after that to teach.
beezle 4 days ago [-]
The letter directory is based upon artist first name... seems odd especially as most are going to know the last name and only maybe the first.
ouli 4 days ago [-]
I fixed it. Thank you for helping to correct such an obvious error!
ouli 4 days ago [-]
Great observation. Thank you for pointing to an obvious error. I will fix it as soon as possible. it is not acceptable to order artists by their first name.
yubblegum 4 days ago [-]
It's wonderful. Thank you for building this.
ouli 4 days ago [-]
Thank you for your kind words!
xydac 4 days ago [-]
this is interesting, we should wire this to frontend design system library that automatically helps user use these palette.
nmstoker 4 days ago [-]
Yes, exactly this. It falls far short of the potential if it just shows the colours alone and not how they might appear if applied to sites, charts, illustrations or whatever you might want them for.
ouli 4 days ago [-]
I am planning to add a section where people can re-color their portraits, landscape images or even interior rooms using carefully curated palettes based on master painters palettes. Applying to websites, illustrations or charts also can be extremely useful.
ouli 4 days ago [-]
Thank you. Glad you find it interesting.
frangonf 4 days ago [-]
As a gruvbox enjoyer, I approve.
austinjp 4 days ago [-]
So... ummm... the website and all the submitter's comments here seem very Claude generated, no?
Spam filters are going to have to get a lot more sophisticated. "Slop" filters, even.
oybng 4 days ago [-]
Very nice. My only gripe is the automatic page switching on scroll, never encountered that before and I absolutely hate it
ouli 4 days ago [-]
Thank you for the kind words and insightful feedback. My intention with page-switching on scroll was to offer more color palettes without requiring extra clicks. I had some reservations about it too, but couldn't find a better way to provide a continuous feed of similar palettes. I'll work on improving that feature.
oybng 4 days ago [-]
It's very convenient, I wish I could offer a worthy suggestion. The trouble in my case is that it's very sensitive and the palettes are barely in view before the page refreshes, they don't reach the center of the screen. Thanks for sharing
CSSer 4 days ago [-]
The problem, from a UX standpoint, is that you need a visual affordance for the behavior. That is, you must indicate that it's about to happen and give the user the opportunity to abort. Alternatively, a continuous gallery could suffice.
ouli 4 days ago [-]
Adding visual clues for automatic scrolling is something I really need to rethink in order to make this feature work as intended. Thank you for the hint!
ouli 4 days ago [-]
I fix it. hope the new version provides better user experience
https://youtube.com/@baumgartnerrestoration
> Starting in the Renaissance, artists made sculpture and architecture that exalted form over color, in homage to what they thought Greek and Roman art had looked like. In the eighteenth century, Johann Winckelmann, the German scholar who is often called the father of art history, contended that “the whiter the body is, the more beautiful it is,” and that “color contributes to beauty, but it is not beauty.” When the ancient Roman cities of Pompeii and Herculaneum were first excavated, in the mid-eighteenth century, Winckelmann saw some of their artifacts in Naples, and noticed color on them. But he found a way around that discomfiting observation, claiming that a statue of Artemis with red hair, red sandals, and a red quiver strap must have been not Greek but Etruscan—the product of an earlier civilization that was considered less sophisticated. He later concluded, however, that the Artemis probably was Greek. (It is now thought to be a Roman copy of a Greek original.) Østergaard and Brinkmann believe that Winckelmann’s thinking was evolving, and that he might eventually have embraced polychromy, had he not died in 1768, at the age of fifty, after being stabbed by a fellow-traveller at an inn in Trieste
https://www.newyorker.com/magazine/2018/10/29/the-myth-of-wh... cited by https://bsky.app/profile/ellipticalnight.bsky.social/post/3m...
Man, what a line. What a horror, this projection of opinion! From the "Father of Art History"! To rob the world so! I feel this way all the time, that anti-sentiment, that the pure marble world just stately and so is art and perfection, over the colors of the universe & it's possibility!
> "the whiter the body is, the more beautiful it is"
This should make your blood run cold, imo. A world locked in amber view of reality, static, sedate. Whew.
For some additional context; many old pigments were not stable at all.
https://www.vangoghstudio.com/what-were-the-original-colors-...
Or do you need to infer it based on location, budget, time, climate etc?
As far as I'm aware there is no way to know for sure what colours originally looked like, especially if the information is limited. There are so many variables, we can only guess.
There has been some research on what causes degradation on paper/pigment but as far as I know much of it ends up as a mystery, a fact of time...
Generally, a painting is best viewed and photographed in as close a lighting environment as it was painted. I have seen many paintings 'blasted' by unnaturally bright gallery lights. There is a reason why a gallery lighting designer is a real job. However, in my experience effective gallery lighting designers are as rare as rocking horse shit.
It is true that the paint the artists applied many years ago will often bear little resemblance to that which we now see. This is less true of earth browns and very true of paintings done at the beginning of the pigment revolution, when wonderful colors were produced which were later discovered to be very 'fugitive'. However, the relative relationship between these paints remains more or less intact, and IMHO this is the most important factor in aesthetic evaluation.
Another factor is how a photo flattens such differences as rough vs smooth (and their consequent light reflection properties). In a Titian painting, huge areas are untouched rough red oxide primer on rough canvas, vs the slight gloss of oil paint. Importantly, old masters would often apply their lights as (smooth) thick paint and their dark as thin glazes (or scumbles) above a thin primer (red or green or yellow ochre or whatever). The frisson between these layers gave the darks their depth that they would otherwise have lacked. This is mastery of dynamic range at its finest. Googles art project photos comes close to capturing this. For an example, check out any portrait by Durer in Google art project.
But seriously. I am not against ai slops in general, people explore what’s possible to make. But I don’t really appreciate them at Show HN - there should be a new thread for ai-made or -assisted projects, or there should be a disclosure in project description.
And the pigments fade. And even worse, they fade at various rates and some are almost completely gone.
Basically
$average = new Imagick( $file );
$average->quantizeImage( $numColors, Imagick::COLORSPACE_RGB, 0, false, false ); //Reduce the amount of colors to 10
$average->uniqueImageColors(); //Only save one pixel of each color
I've thought about making this for a long time to help me with painting, but in that case I think to be useful you need a bit more ways to see the data -- mostly, the thing that is the most important is value. So to get something useful out of it you need a distribution of the hues conditioned value.
And for design, the problem is a bit different. You may have a good looking palette, but 'inverting it' for dark mode is not trivial, and neither are gradients, getting intermediate colors, or getting a shifted hue.
It's called inspiration so it's fulfilling its promise, I'm just curious what are your thoughts on these since you obviously thought a lot about it.
>As a creative, one needs to develop their own <creative thing/technique> for what they are trying to create. So what is the value in looking at how other creative things were made?
That's not the how of the process at all, that's the end result. The "how" for traditional art is completely different; it is taking pigments as a base rather than light, and the algebra of composition is not the same either.
Small snag, some UTF8 things are going on with some colour names, I am sure you know and have cursed accordingly.
I like OKLCH colours and the ability to mix them in interesting ways using CSS things. This means I don't do hex codes for colours in CSS. I can translate though, however, soon some people will demand OKLCH, so you might as well add it in, trying to get it natural with the picker.
I appreciate the masters but I wonder how this would work using other sources, for example, Sunday newspaper supplements from the last century, and their glossy adverts, which were to a higher standard than what we get today.
There are 2 art style pages namely Advertising and Posters styles: https://paletteinspiration.com/advertisement-palettes/ https://paletteinspiration.com/poster-palettes/
I am aware that Advertisement palettes mostly based on Alphonse Mucha work since I could not include more recent ad illustrations for copywrite reasons.
I digress, back to HARSH criticism!!!
Playfair, Serif, in small sizes, I am not a fan. I can see why, and it works for H1 and H2, but, for those of us with elderly eyes, please can we have a sans-serif such as Inter, which does work on smaller sizes? Maybe A/B test that one, with the screen zoom level captured. My hunch is that I am not alone in having elderly eyesight and that people will be more engaged if they have an easier time of reading.
Next bit of harsh criticism - HTML standards. There does seem to be a use case for 'figure' and 'figcaption'. These work like the 2003 era 'div' and nobody cares about HTML, but why bother with HTML if everything is just divs nested in spans and divs?
In truth, 'div' is the element of last resort and not needed since we have joyful content sectioning elements such as 'section', 'article', 'aside' and much else. We also have a layout engine in CSS, as in CSS grid, and, in my experience, the 'div' is not helping. Code is just better when it is not a sea of divs, each with a gazillion class tags.
We have arrived at a situation where people just use 'div' because that was how grandpa wrote HTML when the iPhone came along and table layouts suddenly had to be 'responsive', as in bloated. We built whole departments around writing code badly with 'div' elements, with whole teams of micromanagers. On pet projects there is no need for any of this, plus you can do things like scoped CSS so there is no Firefox support, but also no need to torment oneself with CSS compiler things that create bloat and 'add to only' stylesheets.
But hey, none of this matters if everything works, which it does!
I also think you need some type of funnel to get people to the website in such a way that they come back. This might not matter to you, but, imagine you find this site, do you bookmark it, share a link with oneself, or what? There needs to be some type of hook, and bookmarks don't cut it these days. People want to Google Search for what they know.
One hook might be an upload feature, so you can upload many things. I like de Stihl artists from a century ago and some of them used paint that is still sold today. I also like old posters for British Raleigh bicycles and train posters from the period. The train posters were by commercial artists that made adverts, not art. Yet you still have some brilliant colours going on in their work.
Yes there are services for taking any image and making a palette from it, however, they haven't got the art. They can't do a 'goes with' feature, to show art that is in the user's colour scheme, from what they upload.
I am impressed by the colour naming, this is actually quite hard to write by oneself since you need to find something in a 3D search space. It seems to me that you have the hard stuff mostly solved but tinkering with the UI for stickiness is there to be done, ad nauseum.
Please don't take my HARSH criticisms unfairly, HN is the place for people to niggle, because some people like to cut the heads off others to feel taller, however, I am a genuine punter, and I have a bit more work to do before I get into the colour scheme, so I kind of need to be able to read things easily, remember where your website is and give it a proper go.
What might be useful is a palette export option in :root { CSS variables }, for example:
The vars could be named colours, where possible, so '--warm-caramel' and so on.I still get the autoload of the next palette, mayb what you need instead is a good bit of doom scrolling, a.k.a. infinite scroll. That might be more engaging. I say this because my recipe app also needs infinite scroll for browsing purposes, so the user (my uncle) does not have to drill down to the recipes.
1. You have used the RGB hue wheel not the RYB hue wheel. The later defines the complementary pairs red/green, blue/orange, yellow/purple which are more useful for an artist when evaluating hue. To those who would say that the difference between RTGB and RYB complementaries is subjective, I would point out that the red/green, blue/orange, yellow/purple pairs have been known by artists since before Newton first expressed hue on a wheel.
2. I think it is fair to assume an artist has a consistent color pallet, but I'm not sure that a genre would.
3. User S0und has implied that paintings have changed too much from their point of creation for your work to be valid. For sure it is true that paintings have oxidized, accumulated multiple layers of dust-attracting varnish and been damaged by UV exposure. However, the relative difference between the colors they have used remains almost the same, and this is where the value of your work lays.
4. The key problem of visualizing the color of a painting is its dimensionality. A standard histogram does a super fine job at expressing the lightness value of an image. A radial hue histogram does the same for the hue/saturation. But there is no 2D graphical visualization that does both. Personally, I would go further and visualize separately the hue, saturation and lightness signatures of the artists. For most pre-modern painters, the lightness values closely mirrored their saturation values. Artists like Gericault and most of the Impressionists learned to separate them.
BTW... do you know this website?: https://www.webexhibits.org/pigments/intro/history.html
I think that the best visualization of the colors of a painting is by using the painting itself.
In class I demonstrate lightness by first desaturating the work (or copy pasting the L in Lab). I then do a controlled posterize on the image - basically a stepped curve in Photoshop. I try to isolate the dark, middle and light. These are relative values that can often manifest as lumps in the histogram. Painters tend to be very deliberate in the way they organize them. This page explains what I am getting at:
https://rmit.instructure.com/courses/87565/pages/colour-ligh...
In my experience such posterizing is best done manually but AI might be able to do it.
Hue at saturation are more difficult for the simple reason that they are difficult to disengage from lightness.
Like lightness, saturation is generally organized according to low, middle and high. For most of art history, the saturation of a painting would closely follow its lightness. It was Gericault who separated them. Check out the saturation vs lightness of his Lobster painting for an example of this.
Hue is a beast. Sure, most paintings done before the impressionists are pretty unsaturated. But even Rembrandt would be careful to use a red brown against a green brown. Check out the Rembrandt image on this page to see this in action.
https://rmit.instructure.com/courses/87565/pages/colour-hue-...
I think that a radial histogram is the best way to visualize hue. This would show not only the hues but there relationship to each other on the RYB wheel and also there quantity. There should be a cut-off point for hue that is visible. In our work we established a cut off - all hues with very low saturation were ignored.
Hope this helps.
See also: https://amandahinton.com/blog/creating-a-color-palette-from-...
For Monet, many of the paintings have an important color highlight (e.g., the orange sun), which isn't captured in any of the palettes.
Needs more tweaking.
1) Aren't "Modernism" and "Modernismo" the same thing? I'm a Spanish speaker and from my POV they are
2) Selected "Naïve Art" style and it's broken (images not loading). Probably something to do with the diaeresis
Let me know if you ever create an API endpoint.
https://git.2f30.org/colors/log.html
see this palette for example: https://paletteinspiration.com/fauvism-palettes/fauvism-19-p...
or other Fauvism palettes: https://paletteinspiration.com/fauvism-palettes/
AI generated arts, IMHO, is an economic F-U to the people who spend real years of their lives to train in arts, and to the people who spend more time after that to teach.
Spam filters are going to have to get a lot more sophisticated. "Slop" filters, even.