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.
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.
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.
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.
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).
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?
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...
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.
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.
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.
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.
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.
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.
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.
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
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.
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.
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).
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.
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.
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.
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.
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, 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.
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.
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.
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.
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
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.
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!
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.
https://youtube.com/@baumgartnerrestoration
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.
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.
https://www.vangoghstudio.com/what-were-the-original-colors-...
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?
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.
Its absolutely lossy and you'd have to know a lot about each piece to know how big the loss is.
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.
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).
hundreds of years of oxidation will make everything brown.
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?
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...
I would like to know that formula also. this can be an interesting tool to reveal true colors of paintings as the painters intended...
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.
> 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.
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
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.
So we can use AI or even deterministic algorithms to recover the original palette.
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.
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.
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.
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.
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.
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
Why no contemporary/modern painters? First looked for Basquiat then Lichtenstein , nothing
Most contemporary artists works are still under copyrght laws.
The colours they used aren’t
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
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.
Hey ouli, your hello email bounces.
See also: https://amandahinton.com/blog/creating-a-color-palette-from-...
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.
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.
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).
Quite possible Claude/Anthropic likes your colors
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.
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.
Color extractor from pictures:
https://git.2f30.org/colors/log.html
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.
I fixed it. Thank you for helping to correct such an obvious error!
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.
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.
Great suggestion, thank you! I'll try reaching out to some art schools to see if they'd find the site useful.
Weird how similar many of those are to Commodore 64 palette.
You are absolutely right, I did not realize how similar some of these palettes are to Commodore 64 palette
see this palette for example: https://paletteinspiration.com/fauvism-palettes/fauvism-19-p...
or other Fauvism palettes: https://paletteinspiration.com/fauvism-palettes/
nice
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.
Adding OKLCH color codes will be a great addition. Thank you for your suggestion!
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.
As you suggested OKLCH colour option is added now to the site. Thanks again.
It's wonderful. Thank you for building this.
Thank you for your kind words!
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.
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.
this is interesting, we should wire this to frontend design system library that automatically helps user use these palette.
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.
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.
Thank you. Glad you find it interesting.
As a gruvbox enjoyer, I approve.
Very nice. My only gripe is the automatic page switching on scroll, never encountered that before and I absolutely hate it
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.
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
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.
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!
I fix it. hope the new version provides better user experience
app version?
coming soon