firstly thanks for your videos, your enthusiasm for your subject shines through and is really inspiring! about how colour affects how light it looks, i remember my lecturer at uni saying it was based on how many cones it activates in the retina. we have three types of cone cells, activated differently by ranges of wavelength, peaking their response at long, medium and shorter wavelengths respectively. yellow light activates the largest number of cones so it looks lightest to us. orange and green, then red and turquoise, and finally dark blue and violet activate progressively fewer cones so they look progressively darker. if you're motivated enough, would you be able to do a similar video offering advice to traditional art. like tips on how to understand how this works when painting with watercolour, acrylic, etc.?
What you're mentioning in your video is the result of different mathematical projection between color spaces. Imagining you have 3 spaces [RGB, HSV, Gray]. If you want to get an accurate point from one space to the other you need a function that will transit you to one space to another. Here you used 2 different function to transit from one space to another. One evens the 3 colors, the other applies a different function to the 3 colors to keep the same value. What you found is also part of the color toning techniques 3D renderer have to represent accurate/different dark values (Arri ALEXA for example)
When I started learning digital art, this gave my drawings a huge improvement. Remember guys when you want to go darker change the hue towards purple and when you want to go lighter change the hue towards yellow (for white lighting source).
Color is a well researched field of study. There are explinations for this phenomenon. Things like the "Abney effect" the "Munsell color system" and "Perceptual Uniformity" is what yall are looking for.
It is and it isn't. Qualitatively, a lot is known. But quantitatively, today's standards still often assume a color space that fits on a regular screen. Very high saturations combined with very high dynamic ranges are still not that well-fleshed out it seems. Lots of approaches exist, but none of them are entirely perceptively uniform across the entirety of our vision. Progress is being made though. And for the LDR/sRGB domain that this video focuses, things are working pretty well by now.
We get taught this in advanced art classes, but in more of a roundabout way. It's never explicitly said but it's definitely shown in some projects (such as in black and white still lives).
Even though I’m colorblind, I love watching these types of videos! They help me get a better understanding of how colors work and how other people see them
exactly why i'm watching this video! my friends that aren't colourblind already struggle a lot with colour, but for us it's a completely different beast
It's because of the cones in your eyes! =D I NEED to make a video about this, but essentially the 3 color receptors (red, green, blue) in your eyes aren't equally sensitive. Green cones are the most sensitive and blue cones are the least sensitive, so when your 2 most sensitive types of cones are both stimulated (green + red) you see yellow! =D The only color more stimulating to your eyes is white, which is ALL cones. That's also why blue is the darkest pure hue. Because its the least your eyes can be stimulated apart from NO stimulation, which is black/darkness.
Also: Black and white filters don't usually just take the value of the... value... Instead, they used what is called a "luminance" formula which determines how bright a certain color is based on the aforementioned color cones.
I'm a traditional artist, I think something like this is more intuitive when mixing paint. Gray is made with black and white, and white will always lighten. So it makes sense that a mid gray (meaning a lot of white was added) will be lighter than a pure blue pigment out of the tube. Similarly, it's common to lighten a color with yellow and no white, making it obvious that yellow is lighter than the other colors. I think it's beneficial for people to at least play around with mixing colors even if they're a digital artist.
Agree. Also need to remind people that HBS in computers is just a way to represent a color using 3 components. It really does not have to be physically accurate/representative
The thing is that color saturate or has higher chroma at different brightness levels, so a saturated green or yellow will have a brighter color value, red usually saturated at 50 per cent and blue saturates darker. That's why you can't get a dark saturated yellow for example, also in photography they use colored filter in enlargement machines to control contrast on developing film to paper and that why it is used red as color for the light on those rooms
Well, technically that's only true for opaque mediums. In mediums like watercolor you can have a yellow that's so transparent it has almost no effect on value, only on hue (for example Nickel Azo Yellow PY150) For digitital artists: transparent paints are basically always set to multiply. That's why your point with mixing grays is a good one though. With transparent mediums/color multiplication you can mix two colors of very high saturation AND value, like magenta and green, and get a perfect black with extremely low value. It's really fascinating when you think about it - but it makes sense when you remember paint mixing is subtractive. If a color is completely saturated, it means the rest of the spectrum is absorbed completely. Which results in the whole spectrum being absorbed when you mix two highly saturated complementary colors together and you get black. Gosh, I love color theory.
I actually saw this my self a few years ago. I had two water bottles that I would drink from at night, one was red and the other one was blue. All the lights were off, the only light in the room was the faint glow of the moon through the draperies. I remember being surprised that the blue bottle was a lighter gray even though the red one looked lighter during the day. :)
ooh! actually this is a thing with human eyes perceiving the color red. In dark light, our ability to sense red *at all* takes a nose-dive, so bright reds during light become dark greys during dark. Can't recall where I learned that, but it meshes with my experience.
@@Arkylie fr, when it’s dark, and the only light in the room is from the moon, I can see the color of my blue walls faintly, but my red bedsheets are just grey.
That's different thing, it's not how we see colors, mostly how physics works (illumination and absorption), It's fact that light from moon is bluish (mostly caused by atmosphere) and obviously blue substances will transmit blue light more likely than red light (that light will be absorbed) and vise versa It's like experiment with sodium that causes black fire (the bright light of sodium lamp is absorbed by vaporized sodium if you put wet tissue soaked in salt water into the fire) And the thing mentioned by @Arkylie and @Henry Zhang is purkinje effect And thing mentioned in video is another thing (calculating linear luminance) :)
No, adding saturation didn't decrease value. What you did was a combination of adding saturation and decreasing value. So you went horizontally across the picker? That just shows that on that picker, a contour of colours of equal value is lower on the left and higher on the right.
@@rosiefay7283 saturating a market with a product(putting a lot of product in a market) decreases the value of that product in that market, basic economics lol
What you're seeing is just what the conversion to grayscale is doing - it's setting the value of each pixel to the luminance (not the value) of the original color. Luminance is a perceptual quantity that is the same for two colors that appear to have the same brightness. Value is a mathematical quantity that's just max(R, G, B); it doesn't mean much in terms of human perception when compared between different hues and/or saturations.
@YumeLeaf CyanideWater I0 But cameras don't work the same way our eyes do. Just because our eyes perceive colors a certain way doesn't explain why blue has a lower value than other colors. Now I think the original comment's explanation makes a lot more sense because it relates directly to pixels. I'm willing to bet that if we could only see in black and white, blue light would still look darker than yellow light. So this phenomenon has nothing to do with our perception of color.
@YumeLeaf CyanideWater I0 Answering your first comment: That's exactly the point. Photoshop does NOT just take the value and throw away the rest when you switch to black and white. It instead uses a smart formula based on our perception of colors to create the grayscale pixels. So when he says "different colors have different values", it's not exactly right. Value doesn't care, it's an axis orthogonal to the hue. What he actually means is: "Photoshop will convert different hues in grayscale in different ways to more accurately represent our perception."
This comment needs more upvotes. While watching the video I also thought: There's not _the_ greyscale conversion. There are different ways of converting color to greyscale and the conversion he's doing in Photoshop is apparently not just picking the brightness value.
@@dannydewario1550 Actually they sort of do. The cones (colour sensing parts of eyes) are about 60% green, 29% red and 11% blue). In a 3 CCD camera the CCD output is weighted in the same 60/29/11 ratio! This is the same reason that in a single CCD/CMOS camera the bayer filter is 2parts green to 1part red and 1part blue.
A function I learned in a CS class that might be helpful to understand this phenomenon is how to calculate luma. When converting RGB to grayscale, there are a couple ways you can do it. A naive way is to just average the R, G, and B values, but this has the same problem as removing saturation did in this video. A way that's more accurate to our eyes is to calculate the "luma" of each pixel, where luma = 0.2126*R + 0.7152*G + 0.0722*B. (The reason this is more accurate is probably something to do with our eyes being more sensitive to green light than to red and to blue). This is almost certainly what photoshop is doing under the hood when you change colors to grayscale. It also reveals why the blues end up so dark, because there's a very low coefficient on the blue values of the pixels compared to red and green. Yellow, having both red and green, ends up the brightest after the luma conversion.
The luma calculation must be about more than just eye sensitivity. The human eye is not very sensitive to blue, which is reflected in the low coefficient for blue, but it's about equally sensitive to red and green, whereas the luma formula is very heavily biased towards green.
@@beeble2003 it's because the eye isn't actually sensitive to red or green. what you might instead say is that it's (very) sensitive to the absence of blue it follows that the resulting spectral distribution must be dominated by green light, since whatever range of frequencies you consider to be red is limited by and around the lower extreme of the visible spectrum itself; meanwhile whatever wavelengths you call green get to range more or less as far as you want in both directions from somewhere around the middle of the visible spectrum
@@SplendidKunoichi That doesn't make sense. Darkness and green like are both "the absence of blue" but both look very difficult. And "whatever wavelenghts I call green" are limited to being between red and blue, whereas red can trail off as far into the infra red as you want to call "visible".
@@beeble2003 the key here is that "visible" is not subjective, even if the color names are. "the absence of blue" is a fancy way of saying "yellow", but really it could mean any number of different colors, so long as you can find them on the rainbow; that is, it has to be visible. red light indeed does always "trail off" into invisibility no matter how you slice it. infrared light is invisible; infrared is not a color. wavelengths composing greenish light never do this because spectral green is bounded on either side by other colors (red and blue as you say) that are just as visible albeit not as bright, again no matter how you slice it. it's hard to describe without getting into the math, but visible or not there's a finite amount of light/radiant energy present in the spectral distribution, it just happens that this distribution peaks in the very middle of the frequency range detectable to our cone cells (ie. the G in ROY G BIV). outside of that, spectral red and violet are only seen as very dim side-effects of metal glowing red-hot and cheap blacklights that leak into purple. we can perceive the corresponding IR as warmth and UV as sunburn, but never as color.
@@SplendidKunoichi OK, that makes sense. Describing it as "not-blue" was just a completely unhelpful way of saying "yellow", which is a completely different thing.
I used this explanation for why colors are not a full 3D space in a presentation for uni and got praise for it. Thank you for not only improving my paintings but also my grades! :)
Perhaps you're missing more information about colour spaces. Colours can be perfectly represented in 3D space, but each axis will not necessarily mean something intuitive. That is why we have many 3D colour spaces. If you want one space where one axis represents a realistically perceived bright/dark measure, use HSL or LAB where L is luminance which is meant to be perceptually accurate. HSV, on the other hand, has V for value and it only makes sense mathematically as it is a convenient way to have a saturation axis S that represents deviation from pure grey linearly.
Keep in mind, all of this was done in sRGB space. While it's useful for storing colors, it isn't so great for picking or modifying them. Using a perceptual space, such as Okhsl (which has been added to Photoshop since this video was made), will alleviate some of these issues. Also, the triangle-type pickers work somewhat better because they are an approximation of a perceptual color space. If you take a slice of hue in a perceptual space, you will get a roughly triangle-shaped "flag" of color, with grayscale on one side, and the most saturated point at the tip. However, the shape you get will not be a perfect triangle, and its shape changes depending on hue.
I'm very interested by this, I'm having a hard time understanding how colors and values work, I don't really understand how the different spaces available work and why they're different, but I'm so eager to learn One thing I didn't get from your comment was in the first part, what did you mean by "while it's useful for storing colors"? And also you said the shape of a slice of a perceptual space will vary based on hue, do you know of any website where I could see what this shape should look like for different hues? Does it rely on the greyscale of each hue?
I'm going to wager the perceived "darker" values of red and blue hues has more to do with the sensitivity of our eyes to those color frequencies rather than their physical wavelengths directly. Since seeing "yellow" requires two of our three different receptors to be firing that excess of signal travelling along our nerves probably overloads our sense of 'value' and makes that hue appear "lighter" other hues. You'll see the same effect happen with cyan, which requires our blue and green receptors to both be sending signals at the same time as well.
since our eyes are more sensitive to red and green vs blue, it'd also make sense that yellow would be the brightest, as opposed to magenta and cyan that require mixing less sensitive blue receptors
I dont know... purple is two, or some says purple don't even exist, and it is darker. Red, green, and blue which are the rods, not even have same values.
Johannes Itten also talked about this topic in his book. He used the word "luminosity". For example, yellow has the most luminosity of all, so it tends to be lighter in values
There is multiple ways you can turn a picture in shades of grey: luminance, luma, lightness and value. the closest by design to a perceived luminosity independent of hue is the lightness from CIELab. Lightness from HSL is also independent of hue but yellows appear much darker.
I wonder if it's possible to have a painting that converted to grayscale shows a different image. I mean I know it's possible, just don't know how to find one :P
@@Starfloofle Steganography is much broader term, but in context of graphics it usually means using the least significant bits of each color component of each pixel to hide some message. The technique was used for example in spore to encode your creature code into the creature's photo. There is also some virtual console (i forgot the name) that uses cartridges in form of images with game code encoded using steganography. It's very interesting topic, but not exactly what I meant here ;)
Woah the comment section wise getting interesting over and over! I'd like to see too, maybe in future there's a mystery genre that would use this technique to hide some real meaning or event that secret in art medium haha! Gonna be interesting tho!
Oh my God, I had a problem with "Hue" blending mode a few weeks ago, and it had me pull my hair out. But this explains a lot. The description of the Hue mode says: It keeps the brightness and saturation, but replaces the hue. So I was wondering, why the brightness and saturation changes after I use this mode? I asked so many people and nobody had any idea why it happens. But now I realize, there is nothing wrong with the blending mode. It was our fundamentals that was wrong. Great video! Thank you so much.
Something that's happened to me is that blending modes like "Hue" or "color" don't let you paint with for example, a really dark yellow, or a really Bright red without it looking like something else. Hues have a "Range of values" in which they are readable. Color theory is hard
Ok here's a bit of explanation on a few things you seem to be confused of. The color picker: Your color picker is set to HSB/HSV (same thing), not to HSL (technically there is a difference between HSV and HSL - the L stands for luminance). If you pick two colors with different Value they should in theory stay different in Value when you desaturate them, and that's exactly what happens when you use the HSL-saturation slider shown at 4:50. The HSL-saturation slider does exactly that. It decreases the saturation while keeping the hue and value staying the same. Interestingly enough photoshop now calls it HSL (Lightness), but it's actually HSV as seen in Coral Painter. Oh it gets worse.. in Clip Studio Paint they call it luminocity (but with a V in brackets.. wth). Anyway. Let's say you're using a different way to desaturate the image. How does the desaturation work? One way would be to extract only one channel from the RGB channels and apply it to all the other channels. Cheap digital cameras use this method when applying a build in black and white filter. They overwrite the red and blue channel with the green channel, to get a greyscale image. But that's actually a bad idea. Why? because areas with strong colors like magenta that doesnt have any green in it at all will come out as black or really dark. Moving on what else could you do? You could convert the colors to HSL (luminance) where bright colors like yellow and cyan actually stay bright and visually darker colors like blue and violett stay darker and then strip away the saturation. That way you would at least get an accurate representation of the brightness of the colors in the greyscale image. Actually a good way would be to convert to LAB - wich stands for Luminance, A and B (but wth is A and B???? they are color channels. look it up, it's quite interesting) and decrease the contrast in the A and B channel, that way you only keep the luminance channel. But... you need to convert between RGB and LAB and stuff and that sucks and yeah.... Where am I getting with this? Color theory, especially digital, especially color science, especially with all the f*ing weird lingo is weird and it's complcated and the more you dive into it the more confused and frustrated you get... Let's maybe just... accept that it is what it is and move on before we loose all fun at doing what we're doing? Because I mean that's why you do it right?
@Charon TheReflector I mean I don't say that knowing a lot and maybe even striving to know every little thing about it is a bad thing, but when it comes to creating art it sometimes can get really overwhelming if you spend too much time with all the technical aspects. Understanding it is great and will definetly make you better but while you're trying to understand it you should never lose the fun and joy of doing it.
HSB and HSV names the same thing, but keep in mind that not all software use the same color plotting. Picking the HSB/HSV value from one software and input that directly to another software might not yield the same color, even if both are properly color managed and set to the same color space. Same for HSL I think, but since it's more niche I think it's a bit more unified across software. Also, it's not uncommon to adjust the greyscale conversion matrix to need. For accurate soft proofing result though, we have a few "standarized ones". But unfortunately there are multiple of those. Lab is a good one though as in math that would always be accurate, except... It has to be converted to RGB at some point eventually. And the conversion matrix is also subject to standard variance. And... We still can't see accurate color. That really sucks when think of it, but again Lab is more of a garage dragon to begin with anyway. And... When all of above is solved, we now have HDR and wide color gamut image to deal with. And things just goes so wild that I personally just gave up and decided to go SDR and sRGB for life. Anything else I can probably have by buying a bottle of pigments I guess.
The fact that hue also affects how light or dark a portion of the painting will appear is probably why we tend to gravitate towards using cooler colors for shading and shadows and warmer colors for anything that will be bright or vibrant
As far as i know the human eye perceives green as brighter colors, that's why when converted to greyscale they make green the brightest colors. This is a formula from my computer graphics course, to convert RGB to greyscale: Y = 0.299 R + 0.587 G + 0.114 B
our cone sensitivities overlap with each other, the green cone overlaps with both red and blue more than red does with blue and vice versa, and the brain reads the intensity of each of the cells combined. So green would be the most intense, however yellow is right at the overlap of green and red, making it the largest value out of all. (also blue our sensitivity is lower than other cones)
That's not really right. The names we give colors are determined by how often we see the color. When we see yellows, we almost always see bright, highly saturated yellows, while rarely seeing dark or less saturated yellows, so we only named the shade of yellow that is bright and highly saturated. Red and pink are named as different colors, since whenever we see bright reds, they tend to be very bright (skin, flowers etc. are usually very bright reds), and whenever we see darker reds, they tend to be much darker(blood, flowers, fruits, all usually have a dark red color). We often needed to distinguish dark and bright reds, so they were given different names. Meanwhile, we keep seeing all different brightnesses of blue and green, and we didn't have any use for giving the different brightnesses different names, so the names were used as blanket terms for all brightnesses. We have always been able to tell apart different colors and how bright they are. It was how often we saw different colors, and how much we needed to distinguish different shades that dictated what colors we decided to give names to. Those names stuck, and now the names are used for colors of wildly different brightnesses.
@@adrianks47 The formula you have shouldn't be named greyscale, it's miss leading. True greyscale is the average of RGB. Y = (R+G+B)/3. If it takes into consideration the human eye perceived color they should name it average human perceived greyscale.
@@HenrikMyrhaug This may have once been true but now the cardinal versions of colors are determine by math and their absolute max value hue vs each other. Red pixels at 100% show pure red light and same for green and blue. Show both red and green at 100% and you get an approximation of pure yellow.
The more experienced I get in drawing, the more I prefer to start a piece in grayscale then apply color only after values are established. My brain just processes wavelengths so unreliably.
I discovered this just last year but i didn't know that i could've make the same tests on all colors and connect it to value to make more accurate color choices. Now this is a 7 mins to better painting haha
5:35 Krita users can also use a soft-proofing. It's also under (ctrl+y) shortcut, but the grayscale colorspace has to be picked in the settings under color management tab (the default space is CMYK). And great video by the way :)
@@JosephDavies it could be quite hard to get something so deep in image rendering with python plug-in in Krita, but luckily she already has a quite advanced colorspace management :)
also, 5:00 Filter layers in Krita have a selection of colour modes. You can select the mode in the "types" dropdown. Including HSV and HSLuma You can also change the colour mode for the colour picker toolbar, in settings> colour picker. you can even pick a number for the delta, shift for the colour H/S/L values! This is why Krita (for me) is the best program, and it is open source, the community and user can always make it better!
GIMP has a "Convert colors to gray" option in its desaturation menu that allows you to check values without messing with saturation at all. You can see the difference between desaturation and converting to gray. The resulting images are vastly different in value.
6:49 I love it when science and art overlap! If I remember correctly, I think eyes with normal color vision tend to be more sensitive to green light (more “green” cones than “red” and “blue”). Since red and green cones overlap in the wavelengths they pick up, yellow light appears even brighter.
this can be fixed almost entirely by switching to the LCH colour space. It's essentially the same, but hue and saturation are properly decoupled from value. This is because it's mapped to human perception of colours, not the RGB gamut. I don't know if you can do that in Photoshop, but I tend to design on the web using chroma-js which supports this colourspace. AFAIK the reason this happens is due to the cones in your eyes having different sensitivities to wavelengths of light.
@@aaronbredon2948 in theory yes, but I've found that HCL matches human color perception better than at least the standard CSS HSL. Is there a version of HSL that better matches human perception? CIELAB maybe? I don't know.
@@memyself4852 the CSS color models aren't great. You really need to be using a professional color managed program. Photoshop should be (used to be) the gold standard for color management (if you could navigate the settings) The big issue is that you cannot maintain max lightness and max saturation at the same time, and the amount of loss varies by hue. That is why the LAB and other models where saturation is derived work better for maintaining lightness.
Hi, I am colorblind and I found this extremely interesting. Since I often have problems differentiating or identifying specific colors, I usually rely on the values (unconciously). For example, I sometimes can't tell if a color is green or brown, but I see that one is "darker" than the other (even if on the same saturation), and that helps me know which one is which. I do this very often and it is usually hard to explain to others... believe me, this video made me understand how I see and why I do what I do. This will be very helpful for my art and even my everyday! Thanks! So believe it or not, you unintentionally made the perfect video explaining how colorblind people may distinguish color.
Thanks for making this, so many people whine and wail about how hard it is to understand colour theory when all you need to do is turn on a Marco Bucci video..
Tip: actually in Procreate on iPad, the “Hue Saturation Brightness” slider box is effective in showing the value change in hues, it doesn’t even them all out like other programs :) (but the white fill on “color” mode is effective too)
I remember reading about how each hue has its own inherent value, yellow being the brightest and purple being the darkest. Johannes Itten actually made a brightness scale of each primary and secondary color
One of those things fine arts teachers can never teach us. I never knew this...I assumed it because for some reason all my pics turn out super dark and I can't control my lighting. Makes sense! Thank you!
I was always curious why some color pickers were a triangle and some were just a square (and the occasional weird circle). This makes so much more sense now!
You can use the Lab color space to accurately see the perceived brightness of your color choice! In the "swatch panels" you can see the value labeled as "L" decreasing as he picks a more saturated color.
i just see it as the “darker” colors absorb more light, the “brighter” colors reflect more light, so the darker colors will have a darker hue. kinda like color changing light bulbs. you can have the brightness set to 100%, but the white light will light up more of the room than the yellow light, yellow will light more than the green, green more than the blue, blue more than red, etc.
But what makes any of those colors "brighter" or "darker" than one another, especially when there is a brightness and darkness adjuster already built into the color pickers? That's the question the video is asking. Why are some colors brighter at the same values, and what physical properties does that relate to? My best guess is that it has to do with how the colors are created digitally using rgb, because those three colors appear generally darker while colors like yellow magenta and cyan which are made by combining colors are mostly brighter after grayscale is applied, presumably relating to how many subpixels are activated to create each color. For example the brightest yellow would be (255, 255, 0) while the brightest red would be (255, 0, 0), half as many subpixels being used. And in terms of pigment the effect would work similarly but in the opposite way, where cmy are the primary colors and are combined to make other colors. Since this form of color is subtractive, rgb would still be darkest as more light(specifically a greater amount of different wavelengths)would have to be removed to create those colors. Although the details are less obvious to me than with digital color because I don't know how the hues are created
@@Keijo_ You are on the right track. Our psychological response to a color is based on how much the color activates each of our three color "cones". Humans are naturally much more sensitive to green/lime/yellow than red or blue. We can distinguish hues more accurately there and we see them as brighter even if they are the same intensity. When a software application "greyscales" a picture, it uses a model of the human color response to determine the average cone activation, or the "luminance" of a color. A pure blue pixel (0, 0, 255) has maybe half the luminance of a pure green pixel (0, 255, 0). Blue lightbulbs don't illuminate a room as well as other colors for the same reason. Blue light can bounce around the room, but our eyes don't pick up on it well. It has nothing to do with whether they "absorb more light". They're not absorbing light, they're producing it.
It would be the firing of two simultaneous cones perhaps and our eyes bias toward green (camera sensors tend to have many more green sensors). Another is "yellow" is defined by its lightness. Try making 'dark yellow'.! You end up with greenish brownish mud it doesn't exist any more than 'dark pink' due to pink (not magenta!) Being a 'light red'. Its just an issue with colour space probably. To make pink i.e a lighter shade you add white which in turn desaturates. On the terms 'value' does not figure this and only represents the max firing of a hue, whereas in other models its termed 'lightness' which by definition factors these 'tints' (add white) in and is coupled with 'chroma' not saturation. Chroma differs by being invariable for lightness... Yellow is not the same as 'light cyan' of equal chroma in lightness in such a colour model.
incredibly well-done video! Wow. As a fellow video-maker and editor, I winced at how much time must have gone into chopping up such a snappy, to-the-point, visually CLEAR video. Excellent.
In RGB, on digital screens, it is vital to understand that there's a huge difference between the objective brightness between RGB and CMY the reason is simple: for red we have one light active, just the R in RGB, yet for yellow we also add green G, so digital yellow, magenta and cyan are about twice as bright as red, green and blue respecticely. add to that the specifics of how human cones perceive color and it becomes a bit messy, but that inherent distinction between the hues on digital screens is important
If you want it simple warm colors have lighter values and cold colors have darker values, if you divide it that way it makes it easier to find the right value you want (though it's not alwase like that but starting with these basics is what helped me more to understand)
The brightest colors will be the ones between red and green, closer to green, which is where yellow sits. It’s the brightest because it contains both red and green. Blue is the darkest, so the darkest colors sit between blue and red, but purple is brighter than blue because it contains red, which is more luminous than blue. In short: Brightest - between Red and Green Mid - between Green and Blue Darkest - between Blue and Red Desaturating adds more colors, so it equals in brighter colors.
Thank you for this explanation. i was told not to just turn down saturation before but when i asked why, they couldnt explain. this makes the reasoning easy to explain and follow.
in traditional oil painting atelier classes we learn that a "neutral" green has the most "neutral" local value. while a bright "warm" yellow has the lightest local value.
this is why when i kept not being happy with my drawing i just turned greyscale on and grabbed a random color pallete and worked with the values, and i love the way it came out when i turned color back on
THIS!!! I've been struggling with this particular issue lately and I thought it was just me being dumb and confused. Thank you for clearing the fog a bit =)
I can’t tell you how much this will help me! I’m a 7th grader and I love art. I’ve been doing it since 2nd grade and I’ve improved A LOT! But I do have to say I’m not good with color theory, I just recently learned about it and it still hurts my Brian but watching this video helped me understand a lot. It will help me improve my drawings in the future and help me get more comfortable with drawing in color instead of just black and white 😅. I’m mostly self taught. The only thing that I didn’t pick up by myself is anatomy. And I still suck at it. I would ask my art teacher for some guidance and help but I don’t have one. I’m an online student and the program I’m in only allows 4 classes to be worked in! While I do have to admit it is great only have to worry about 4 classes it honestly sucks because art isn’t one of those classes. Like I said, I love art. It’s been my dream to become a professional artist since I was 7! Not being able to have access to someone that knows a lot about something I’m really passionate about really does suck. But i only have half a year left of online school and I’m going to be going back to real school next year! But until then I’ll stick to watching videos like this to help me understand how certain things in art work instead of watching tutorials on HOW to do something. Videos like this really help with understanding art, instead of just copying mark by mark of some famous artist. Trust me, not something you would want to do. Biggest regret in life. It lead me nowhere in my art and left me bummed out. I soon began to realize those aren’t the kind of videos you would want to be watching if you are really passionate in your work and want to learn something new! While those videos can be useful for some, it just wasn’t right for me you know? I wanted to feel like I was actually learning something useful that I could use in drawings/ paintings. Those videos just weren’t what I was looking for. Not saying they aren’t useful. Or that you shouldn’t watch them, they just weren’t right for ME. You know? 😅 Sorry for my little ramble 😅. Again, this will help me tons!!
i remember watching a video a while back on how to draw "hyperpop" and it showed values of each color completely saturated. i found this video just recently and it's nice to see more things about it
If you create a spectrum starts from blue ends with turquoise you will find that it is gradually getting lighter. This is why a lot of artists change their hues a little towards blue/purple to get a more dynamic feel in their piece.
In Krita, you can open up a new view alongside the one you're currently working on, and have two different windows of the same drawing, one in grayscale and one in full color, and they both update in real time.
Are you reading my mind???? Just thought of this recently!!! When I go darker and slightly more saturated the colors end up muddy, so afterwards I go slightly lighter and more saturated, much better results! Oh just saw the thumbnail didn't even watch yet lol
The book "Color and Light" by James Gurney talks about this phenomenon called "peak chroma value". So essentially whenever you have a hue, that hue's strongest chroma will be located on a spot in a value range from light to dark. Yellow's strongest chroma is always a light value, red's strongest chroma is a mid, and blue's strongest is a dark. If you paint many gradations of chroma and value of each of those colors you'll create a graph that visually shows where the chromas concentrate on the value scale.
I KNEW IT! I didn't know how to articulate this, but I've suspected this for some time now... >_>' I've run into this dilemma pretty consistently and had no clue how to adjust for it. Thank you for illuminating this weirdly specific phenomenon.
1:14 "you know like value is determined by this vertical placement and then you plug in hue and saturation after" nah bro I just randomly click on points of the square until the color that comes out look cool
Wow and i was thinking it’s just in rgb images When converting rgb image to greyscale , the color channels contribution will be like this 30% red 59% green and 11% blue and averaging the the value by 3 Thanks for the art and science lesson 🙏
this is something i could never find the right words to describe when i was doing my art research final on advanced color! thank you for the satisfaction
wow! what an incredible video!!! i've been studying colours more then never this year and this video came like a cherry on top, but if the cherry whas the size of a watermellon
2:46 great now even eyedropper tool is deceptive and those digital artists who paint initially in grayscale and then add color in after are gonna really need this info
thats gonna be a bit difficult tho. You'd need to use a different hue arrangement in the first place, based on evolutionary bias(green is the most sensitive color to our eyes, so its the least grey, red being second and blue being third), then you need to find an arrangement where blue, red and green together are the brightest grey while blue alone is the darkest. Essentially, its impossible to do on a 2d scale, and its going to be fairly though to create a 3d scale thats easy enough to work with. pnghut.com/png/CvDbv7RDGC/hsl-and-hsv-color-space-barvni-model-picker-james-clerk-maxwell-transparent-png is basically the most accurate, easy to use 3d shape that fills picking color to full satisfaction, but its not exactly easy to work with unless you create a VR room.
@Kyaru Momochi Depends which hue you are talking about. If you are talking about computer hue, then there are accurate color pickers. Accurate doesn't mean corresponding exactly to reality or peception.
@Kyaru Momochi Something is not flawed just because it doesn't contain every single color a human eye can see. Sometimes I don't need to know what a picture looks like printed out, but only what it looks like on a computer screen, and in those cases color pickers are not flawed, they show exactly what you what, what it will look like on a screen. Sometimes I just need to see all the colors the computer can produce, not all the colors I can see. A color picker that does that is not flawed, its doing exactly what I want.
The value assigned to colors when you switch to B&W were determined experimentally by showing people colors and asking them how bright they think they are, but you could also think of them as being entirely arbitrary.
In terms of art, framing human color perception as arbitrary feels completely useless. Afaik the grayscale conversion algorithm shown here isn't based on self-reporting but physical evidence from the distribution of cones in human eyes. This video is very misleading though, should've just explained how we perceive color..
@@terryriley6410 Maybe he doesn't understand how we perceive color. No one does, that's why it's called color "theory" and not color "law." And how can the video be misleading when he didn't make any claims about how we perceive color? He just said the software is misleading. Which it is... Everything else was just him speculating and asking why things are the way they are.
@@TheYeetedMeatbut the proof may not always be true, there are a lot of examples in science when the theory turned out to be erroneous, although it seemed logical
@@lkoyumil and did I say otherwise? It’s just that most people misinterpret theory as hypothesis and then use that to try and claim that alternative things should be taught despite being obviously false.
Just a quick head up, YOU CAN DO SOMETHING SIMILAR TO THE PROOF STUFF IN KRITA. Go to image>change color space>soft proofing>choose the color space to whatever color space is convenient (I just use grayscale). Or if you want it to work across all your files go to settings>color management>soft proffing>choose the color space to any gray space. Short cut will be Ctrl+Y There are other ways to do it that use LUT management but you will be better off just reading the manual to know what it is.
The different algorithms for convert to grayscale by either “convert to bw” or “decrease saturation” is not explained, yet everything centers on it. Also I miss the obvious gamma discussion when comparing how we perceive the linear color space when it’s modified.
@@arkarts2987 On the contrary, I think Andrea's comment is rather on point. The Value computations used by Photoshop et al. attempt to mimic human perception of color, so the "something strange you should know about color" is more about human perception than digital painting, though the disagreement between Photoshops' formulas hints at this perceptual thing in an interesting way. Marco's statement that "this stuff happens with real acrylic pigments too" is slightly silly. _Of course_ the blue acrylic would have less Value than the red acrylic because he's using _the same_ algorithm to compute Value. It's silly (but only slightly) because he brings up this example to demonstrate the inherent nature of Value, as if he has applied different procedures to discover the same Value difference. But in both cases the procedure is the same, the only difference being that the color data in the image is extracted from the photo instead of generated by the software. So, like how the disagreement between Photoshop's algorithms hints at the non-intrinsic and very human-centric nature of Value, Andrea's comment hints at the weirdness of this part of the video: Marco presents the acrylic demonstration *as if* it shows something new. If anything it just shows that his camera accurately represents color (for our purposes) and that his computer can talk to it.
@@pwhqngl0evzeg7z37 I understand that completely and you explained it very well. Although there isn’t really any other way of testing it unless you got people who are completely colorblind and surveyed them about their observations for the colors, but that form of color blindness isn’t very common and probably wouldn’t further prove anything. So maybe he could have just skipped that part of the video completely because it’s probably unlikely anyone would bring up “well what about physical pigments?” And for the few people who do bring it up this exact conversation would probably arise. So it’s hard to criticize his method of proving his point because there aren’t many other ways of testing it but at the same time the method is still slightly flawed, that’s my conclusion.
@@arkarts2987 Yeah I agree. What he's trying to prove is kinda impossible to prove, except by appealing to an authority, in this case (I assume) the neuro/bio/psychological consultants for the creation of the Value formula, and trusting that they did it right. The only thing that I can think of as another way of proving this would be to make a mechanical replica of the eye, but the accuracy of this is essentially no better than the formula. Interesting thoughts nonetheless.
This phenomenon is explained in short but amazing detail in the book "Artists' Master Series: Color and Light" by Pickard. It has to do with the rods and cones in our eyes that our brains use to perceive color. After explaining _why_ colors have inherent values, it gave patterns and tips for how to remember and make use of it: colors are lighter toward green (though green is not lightest, yellow is, and blue is darkest), and you can get lighter colors in your artwork not just by adding white/desaturating (which will also lower the chroma or intensity of the color), but by adding a different hue which is inherently bright (which will allow you to control the color intensity you want in your overall painting composition).
I've just found this video. Maybe someone has mentioned the following already: The RGB values could help to understand what's happening here. The RGB values are what really makes up your digital color. The HSL sliders are ways to maneuver within this colorspace, like cars with different properties driving on roads, so every tool (car) can lead to different results. But the RGB values are what the streets are made of, so to speak. Each RGB channel is its own brightness / intensity. And the easy-to-understand part in comparison to HSL is, that you can add each of them. It's pretty linear / straight forward. 255 is max brightness for each channel 255+255+255 RGB is the max, pure white. 0+0+0 is black. The RGB values also explain your gray vs violet example. The gray is 128+128+128 which is more than the violet with 77+0+255 in total "intensity", so it's the brighter color. We do need the HSL (or similar systems) to navigate within a 3D color space intuitively, because no one can just perfectly assume RGB values. ("hey give me that slightly desaturated dark violet", " oh you want an 118-76-119, I knew it straight away" not going to happen) The way the wavelengths of light add up is not that intuitive to work with. That red light and green light add up to yellow and are in total brighter than the separate colors. 255+0+0 (red light) added to 0+255+0 (green light) will end up as 255+255+0. The RGB system and colorspaces are a really precise way to describe the characteristic of the addition of light wavelengths and also works well in a digital code. That's why it also applies to the photo of your acrylic pigments. The lights wavelengths add up in the real world, and the RGB system describes it fairly accurately. All cameras save their images in RGB systems, professional cameras with more differentiations between the darkest value and brightest value. 0 - 255 are 256 possible values, which is 2^8 referring to 8bit. Most professional cameras are nowadays in the 14 bit range, so they have 16384 different values between dark and bright. But for easier working they are converted to 8 bit in the process of developing the digital image on a pc. The RGB System is everywhere where light is emitted. Your Monitor is working in the same way and adding up values of single red, green and blue diodes to create a pixel. The HSL and colorwheels are needed to give the RGB colorspaces a user-friendly and intuitive way of using it. They are not using the addition of wavelengths to describe a color. They let you use the mentioned tools of Hue, Satuarition and Value (Brightness) to describe a color intuitively and they try to match it / translate it to a color from the RGB color space. This process can always have it's minor faults or inconsistencies. Standard RGB (sRGB) is the most common RGB colorspace, but there are many more. For example: Some are bigger and allow for extremer saturations for professional use. But they all follow the same principle. A different system is needed to describe colors for printing. When Light is not emitted, but only reflected. Here you need CMYK (cyan, magenta, yellow and key (black). It works the other way around. It's not adding up, it's subtracting. You start with a blank paper, it is white, because it can reflect every RGB color from the emitted light it was hit by. If you add Cyan color on the paper, the red part of the incoming RGB light is absorbed and only green and blue will be reflected, blue and green light wavelengths will add up again to cyan. So you are removing wavelengths here every time you add a color. We will go full circle when you understand that the primary colors in printing (cyan, yellow, magenta) are the secondary colors in RGB (red+green=yellow, green+blue=cyan, blue+red=magenta) and vice versa. (Add Cyan and yellow on a sheet and green will be seen, because the green wavelength is the only one not absorbed)
Not that simple. It's not how it works for humans. (0,0,255) is darker than (255,0,0) It's shown clearly at the end of the video. That's because the blue component adds less to the overall (perceived) brightness than the red component. So for the RGB device (monitor) that's emitting light, (0,0,255) and (255,0,0) may have the same energy, but for the human eye blue looks darker. That's what the video is about. You also claim "it's pretty linear", again that's true from the perspective of the light emitting device, not the human eye. HSV is just a different way to navigate the RGB space, but it's still RGB space, designed to describe how machine represent colors not how the human eye percieves them. For the same reason it's wrong to think that (255,0,0) is as bright as (0,0,255), moving for constant values of V leads to changes in percieved brightness, even if, moving along V, only brightness changes. The point of the video is that while V changes only brightness, that doesn't mean that it's the only axis that changes brightness. Both H and S change it too. For a color space aimed at representing colors the way the human eyes sees them, check out L*a*b*.
theres an interesting concept called luma, which is a formula that weighs every rgb color channel by a percentage to get a single value that describes perceived brigthness. commonly its calculated as luma = 0,21*R + 0.72*G + 0.07*B, which returns a value between 0-255 since the factors add up to 1. low numbers suggest darker colors and higher numbers brighter. very useful if you want to determine if text color should be black or white to be readable on a screen with variable background colors
@@leckererlurch237 That's still an approximation, it's based on an average contribution of each channel to the overall brightness. It is good and useful, and way simpler than perform colospace transformation, which is not necessarily linear.
So I oddly learnt about this in one of my computer science modules (possibly computer vision). Its got to do with the way we perceive colours. We have 3 receptors to see, red green and blue. Green is the most sensitive, red sort of in the middle and blue the least sensitive. The way we convert from RGB to greyscale is based on the way we see the different colours and if you looks up the formula used to convert to greyscale its Gray = 0.299R + 0.587G + 0.114B. Im just looking at the formula for converting RGB to HSV and it doesn't really make sense to me... Id have guessed that it would be the same as the conversion but turns out that would be the lightness value and there are different conversion standards... the above is SDTV standard (aka recommendation 601). Adobe used Rec. 240 and it is Y = 0.212R + 0.701G + 0.086B. HDTV used Rec. 709 which is Y = 0.2126R + 0.7152G + 0.0722B. UHD and HDR used Rec.2020 which is Y = 0.2627R + 0.678G + 0.0593B. Interesting! To get the value of HSV it its the intensity of the colour thats has the highest intensity so which is bizarre to me because that implies that if as long as you only change the values of colour channels that aren't the highest intensity the value won't change? (255, 0, 0) will have the same value as (255, 200, 200)? Back to to the brightness if you have pure blue (0, 0, 255) and pure green at 20% (0,51,0) they would have the same lightness which according to my quick test pinta (I don't have photoshop) yes they do as they grayscale for both are (29,29,29). For what its worth to get red to the same value would be red at 38% so (97,0,0) All very interesting stuff and my mind is officially blown. I wonder if there are any paintings where there are secret messages hidden that can be revealed merely by putting it into greyscale. I know that there is a way of adding secret message to images by altering the RGB value by 1 and then when the intended receiver uses a program to compare the original to the one they were sent they can get the binary from it and decode the message, pretty genius actually. Called steganography and technically hiding a message within a picture that can be seen by making grey scale would also be a form of steganography
Rec.709 and Rec.2020 are using a different standard of "red", "green" and "blue". Rec.709 primaries documented the capacity of CRT monitors in 1990s. Mainstream computer photos and videos also use this same set of RGB primaries. Meanwhile, Rec.2020 specified a set of purer "red", "green" and "blue" than what we had in the past, using a single wavelength light. (piano vs. tuning fork) The recent iPhones are taking photos and videos in P3 color space. They used the same "blue" of mainstream but a purer "red" and "green", albeit not as pure as Rec.2020. Another factor is gamma. sRGB, what mainstream computer photos standardized on, is slightly different from Rec.709 here. But that doesn't matter. What's matter is that "20% brightness" doesn't mean a 255*20% RGB value. You need to put the 20% brightness into the gamma curve. A (128, 128, 128) pixel is emitting significantly less than 50% light of a (255, 255, 255) pixel. (We feel 50% because human perception is not linear.) Uncalibrated LCD monitors and TVs may or may not fit the curve exactly as specified but they still roughly follow suit. GIMP converts (0, 0, 255) into (70, 70, 70) for grayscale. Your (29, 29, 29) looks strange.
@@billyswong From my little bit of game dev experience I've learnt, to properly simulate perceived audio volume, you'd need to logarithmically increase the value. Could the same be true for perceived brightness?
@@keppycs Overall, yes, else we would perceive objects changing colour depending on the amount of sunshine. But the situation is in fact more complicated as part of the logarithmic / exponential perception is achieved by compensation in our mind, not by the eyes themselves. Thus a series of optical illusion and tricks is possible by crafting special images.
![Seungmin "그렇게, 천천히, 우리(As we are)" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/kAzmhLHePqU/mqdefault.jpg)
Hey all! You can download the value/saturation chart here: rb.gy/fn7c0v
Thank you so much...
You are my best, favourite art teacher, ever... I learned so much from you... Thank you so much...
Love you...
❤️❤️❤️
can you post a non-psd version?
firstly thanks for your videos, your enthusiasm for your subject shines through and is really inspiring!
about how colour affects how light it looks, i remember my lecturer at uni saying it was based on how many cones it activates in the retina.
we have three types of cone cells, activated differently by ranges of wavelength, peaking their response at long, medium and shorter wavelengths respectively. yellow light activates the largest number of cones so it looks lightest to us. orange and green, then red and turquoise, and finally dark blue and violet activate progressively fewer cones so they look progressively darker.
if you're motivated enough, would you be able to do a similar video offering advice to traditional art. like tips on how to understand how this works when painting with watercolour, acrylic, etc.?
You used orange in yellow chart?
What you're mentioning in your video is the result of different mathematical projection between color spaces. Imagining you have 3 spaces [RGB, HSV, Gray]. If you want to get an accurate point from one space to the other you need a function that will transit you to one space to another. Here you used 2 different function to transit from one space to another. One evens the 3 colors, the other applies a different function to the 3 colors to keep the same value. What you found is also part of the color toning techniques 3D renderer have to represent accurate/different dark values (Arri ALEXA for example)
Makes me want to see a color painting that, when converted to greyscale, is just a solid grey block.
same
now i wanna do that lol
Just a straight up "Fuck you" to people who don't have rods
@@cupoftea3499 isn’t it cones that do color?
@@Obbsnja sorry, I got them mixed up, yeah, people who don't have cones.
This both educated me and made me more confused at the same time
the more you know, the more you know you dont know
@@Matty002 hahaha true
Hah same
@@Matty002 René Descartes
this happens when someone who is trying to educating you also says that they neither know
When I started learning digital art, this gave my drawings a huge improvement. Remember guys when you want to go darker change the hue towards purple and when you want to go lighter change the hue towards yellow (for white lighting source).
This just added 3 years to my life span, why did nobody tell me this
Obrigado
My brain said note this comment ❤❤❤
look at this chad giving away life changing tips for free
this is making my brain hurt cuz im stupid and dont know if im always following this (usually am, but am i ALWAYS following it?)
Color is a well researched field of study. There are explinations for this phenomenon. Things like the "Abney effect" the "Munsell color system" and "Perceptual Uniformity" is what yall are looking for.
It is and it isn't.
Qualitatively, a lot is known. But quantitatively, today's standards still often assume a color space that fits on a regular screen. Very high saturations combined with very high dynamic ranges are still not that well-fleshed out it seems.
Lots of approaches exist, but none of them are entirely perceptively uniform across the entirety of our vision.
Progress is being made though. And for the LDR/sRGB domain that this video focuses, things are working pretty well by now.
@@Kram1032 Never said we handle any of it correctly. Just that we have a decent idea about perception and colour.
Thank you!
Thank you sir
Pin this!
Me: Okay... cool
my brain: *starts melting cuz everything I believed was a lie *
Same
Really takes "visible confusion" to a whole new level, doesn't it?
We get taught this in advanced art classes, but in more of a roundabout way. It's never explicitly said but it's definitely shown in some projects (such as in black and white still lives).
Thank you. I though it was only me.
what did you believe in?
Even though I’m colorblind, I love watching these types of videos! They help me get a better understanding of how colors work and how other people see them
Good for you! ❤️ Btw if you don't mind me asking, what type/category of colour blind are you? (sorry if it sounds rude I only speak Malay)
Me too! I love watching these kind of videos but i am colorblind. It is really interesting to see other people's view of these kind of things
exactly why i'm watching this video! my friends that aren't colourblind already struggle a lot with colour, but for us it's a completely different beast
imagine being colorblind, at least have the self-respect to be blind ☠
@@berjoxhn5142 your message is invaluable
It's because of the cones in your eyes! =D
I NEED to make a video about this, but essentially the 3 color receptors (red, green, blue) in your eyes aren't equally sensitive. Green cones are the most sensitive and blue cones are the least sensitive, so when your 2 most sensitive types of cones are both stimulated (green + red) you see yellow! =D The only color more stimulating to your eyes is white, which is ALL cones.
That's also why blue is the darkest pure hue. Because its the least your eyes can be stimulated apart from NO stimulation, which is black/darkness.
Not the case for the cones in mine... * *Cries in deuteranopia* *
This is great, thanks! Assuming that's why you can use blue to 'rest' your eyes?
Also: Black and white filters don't usually just take the value of the... value... Instead, they used what is called a "luminance" formula which determines how bright a certain color is based on the aforementioned color cones.
omg please do make a video reftgyhuj
Thank you! I was thinking this was probably behind it, but it's really good to feel validated XD
I'm a traditional artist, I think something like this is more intuitive when mixing paint. Gray is made with black and white, and white will always lighten. So it makes sense that a mid gray (meaning a lot of white was added) will be lighter than a pure blue pigment out of the tube. Similarly, it's common to lighten a color with yellow and no white, making it obvious that yellow is lighter than the other colors. I think it's beneficial for people to at least play around with mixing colors even if they're a digital artist.
Agree. Also need to remind people that HBS in computers is just a way to represent a color using 3 components. It really does not have to be physically accurate/representative
Good points, both of you!
@@AlejandraRafa thank you!
The thing is that color saturate or has higher chroma at different brightness levels, so a saturated green or yellow will have a brighter color value, red usually saturated at 50 per cent and blue saturates darker. That's why you can't get a dark saturated yellow for example, also in photography they use colored filter in enlargement machines to control contrast on developing film to paper and that why it is used red as color for the light on those rooms
Well, technically that's only true for opaque mediums. In mediums like watercolor you can have a yellow that's so transparent it has almost no effect on value, only on hue (for example Nickel Azo Yellow PY150) For digitital artists: transparent paints are basically always set to multiply. That's why your point with mixing grays is a good one though. With transparent mediums/color multiplication you can mix two colors of very high saturation AND value, like magenta and green, and get a perfect black with extremely low value. It's really fascinating when you think about it - but it makes sense when you remember paint mixing is subtractive. If a color is completely saturated, it means the rest of the spectrum is absorbed completely. Which results in the whole spectrum being absorbed when you mix two highly saturated complementary colors together and you get black. Gosh, I love color theory.
I actually saw this my self a few years ago. I had two water bottles that I would drink from at night, one was red and the other one was blue. All the lights were off, the only light in the room was the faint glow of the moon through the draperies. I remember being surprised that the blue bottle was a lighter gray even though the red one looked lighter during the day.
:)
ooh! actually this is a thing with human eyes perceiving the color red. In dark light, our ability to sense red *at all* takes a nose-dive, so bright reds during light become dark greys during dark.
Can't recall where I learned that, but it meshes with my experience.
@@Arkylie fr, when it’s dark, and the only light in the room is from the moon, I can see the color of my blue walls faintly, but my red bedsheets are just grey.
for future people finding this thread: this is called the purkinje effect
That's different thing, it's not how we see colors, mostly how physics works (illumination and absorption), It's fact that light from moon is bluish (mostly caused by atmosphere) and obviously blue substances will transmit blue light more likely than red light (that light will be absorbed) and vise versa
It's like experiment with sodium that causes black fire (the bright light of sodium lamp is absorbed by vaporized sodium if you put wet tissue soaked in salt water into the fire)
And the thing mentioned by @Arkylie and @Henry Zhang is purkinje effect
And thing mentioned in video is another thing (calculating linear luminance)
:)
1:58 "adding saturation decreased value!"
Well yeah, that's just basic supply and demand
No, adding saturation didn't decrease value. What you did was a combination of adding saturation and decreasing value. So you went horizontally across the picker? That just shows that on that picker, a contour of colours of equal value is lower on the left and higher on the right.
@@rosiefay7283 I don't know how to tell you this, but I am not the creator of this video, I'm just a random commenter making a joke
@@rosiefay7283 saturating a market with a product(putting a lot of product in a market) decreases the value of that product in that market, basic economics lol
@@z-beeblebrox lol
@@danzjz3923 You're gay.
What you're seeing is just what the conversion to grayscale is doing - it's setting the value of each pixel to the luminance (not the value) of the original color. Luminance is a perceptual quantity that is the same for two colors that appear to have the same brightness. Value is a mathematical quantity that's just max(R, G, B); it doesn't mean much in terms of human perception when compared between different hues and/or saturations.
^This should be top comment
you seem to be the only one here who explains the situation accurately
@YumeLeaf CyanideWater I0 But cameras don't work the same way our eyes do. Just because our eyes perceive colors a certain way doesn't explain why blue has a lower value than other colors. Now I think the original comment's explanation makes a lot more sense because it relates directly to pixels. I'm willing to bet that if we could only see in black and white, blue light would still look darker than yellow light. So this phenomenon has nothing to do with our perception of color.
@YumeLeaf CyanideWater I0 Answering your first comment: That's exactly the point. Photoshop does NOT just take the value and throw away the rest when you switch to black and white. It instead uses a smart formula based on our perception of colors to create the grayscale pixels.
So when he says "different colors have different values", it's not exactly right. Value doesn't care, it's an axis orthogonal to the hue.
What he actually means is: "Photoshop will convert different hues in grayscale in different ways to more accurately represent our perception."
This comment needs more upvotes. While watching the video I also thought: There's not _the_ greyscale conversion. There are different ways of converting color to greyscale and the conversion he's doing in Photoshop is apparently not just picking the brightness value.
@@dannydewario1550 Actually they sort of do. The cones (colour sensing parts of eyes) are about 60% green, 29% red and 11% blue). In a 3 CCD camera the CCD output is weighted in the same 60/29/11 ratio! This is the same reason that in a single CCD/CMOS camera the bayer filter is 2parts green to 1part red and 1part blue.
A function I learned in a CS class that might be helpful to understand this phenomenon is how to calculate luma. When converting RGB to grayscale, there are a couple ways you can do it. A naive way is to just average the R, G, and B values, but this has the same problem as removing saturation did in this video. A way that's more accurate to our eyes is to calculate the "luma" of each pixel, where luma = 0.2126*R + 0.7152*G + 0.0722*B. (The reason this is more accurate is probably something to do with our eyes being more sensitive to green light than to red and to blue). This is almost certainly what photoshop is doing under the hood when you change colors to grayscale. It also reveals why the blues end up so dark, because there's a very low coefficient on the blue values of the pixels compared to red and green. Yellow, having both red and green, ends up the brightest after the luma conversion.
The luma calculation must be about more than just eye sensitivity. The human eye is not very sensitive to blue, which is reflected in the low coefficient for blue, but it's about equally sensitive to red and green, whereas the luma formula is very heavily biased towards green.
@@beeble2003 it's because the eye isn't actually sensitive to red or green. what you might instead say is that it's (very) sensitive to the absence of blue
it follows that the resulting spectral distribution must be dominated by green light, since whatever range of frequencies you consider to be red is limited by and around the lower extreme of the visible spectrum itself; meanwhile whatever wavelengths you call green get to range more or less as far as you want in both directions from somewhere around the middle of the visible spectrum
@@SplendidKunoichi That doesn't make sense. Darkness and green like are both "the absence of blue" but both look very difficult. And "whatever wavelenghts I call green" are limited to being between red and blue, whereas red can trail off as far into the infra red as you want to call "visible".
@@beeble2003 the key here is that "visible" is not subjective, even if the color names are.
"the absence of blue" is a fancy way of saying "yellow", but really it could mean any number of different colors, so long as you can find them on the rainbow; that is, it has to be visible.
red light indeed does always "trail off" into invisibility no matter how you slice it. infrared light is invisible; infrared is not a color. wavelengths composing greenish light never do this because spectral green is bounded on either side by other colors (red and blue as you say) that are just as visible albeit not as bright, again no matter how you slice it.
it's hard to describe without getting into the math, but visible or not there's a finite amount of light/radiant energy present in the spectral distribution, it just happens that this distribution peaks in the very middle of the frequency range detectable to our cone cells (ie. the G in ROY G BIV). outside of that, spectral red and violet are only seen as very dim side-effects of metal glowing red-hot and cheap blacklights that leak into purple. we can perceive the corresponding IR as warmth and UV as sunburn, but never as color.
@@SplendidKunoichi OK, that makes sense. Describing it as "not-blue" was just a completely unhelpful way of saying "yellow", which is a completely different thing.
"this kinda blue my mind"
"yellow there, i red about this somewhere and it blue my mind, orange they amazing? it makes you pink about it"
@@SreenikethanI Gray there! I had green it before, so I'll just cyan move on
@@SreenikethanI makes me remember that one guy in flapjack
I love you for this comment 😂
@@DylanRomanov Say your name but remove the r 😳
I used this explanation for why colors are not a full 3D space in a presentation for uni and got praise for it. Thank you for not only improving my paintings but also my grades! :)
would love to read or listen to your explanation if its archived/transcribed anywhere!
@@MrValsung me too
@@MrValsung me tooo
Could you share your findings?
Perhaps you're missing more information about colour spaces. Colours can be perfectly represented in 3D space, but each axis will not necessarily mean something intuitive. That is why we have many 3D colour spaces. If you want one space where one axis represents a realistically perceived bright/dark measure, use HSL or LAB where L is luminance which is meant to be perceptually accurate. HSV, on the other hand, has V for value and it only makes sense mathematically as it is a convenient way to have a saturation axis S that represents deviation from pure grey linearly.
Keep in mind, all of this was done in sRGB space. While it's useful for storing colors, it isn't so great for picking or modifying them. Using a perceptual space, such as Okhsl (which has been added to Photoshop since this video was made), will alleviate some of these issues. Also, the triangle-type pickers work somewhat better because they are an approximation of a perceptual color space. If you take a slice of hue in a perceptual space, you will get a roughly triangle-shaped "flag" of color, with grayscale on one side, and the most saturated point at the tip. However, the shape you get will not be a perfect triangle, and its shape changes depending on hue.
As a color nerd, yes, you are completely right.
I'm very interested by this, I'm having a hard time understanding how colors and values work, I don't really understand how the different spaces available work and why they're different, but I'm so eager to learn
One thing I didn't get from your comment was in the first part, what did you mean by "while it's useful for storing colors"? And also you said the shape of a slice of a perceptual space will vary based on hue, do you know of any website where I could see what this shape should look like for different hues? Does it rely on the greyscale of each hue?
I'm going to wager the perceived "darker" values of red and blue hues has more to do with the sensitivity of our eyes to those color frequencies rather than their physical wavelengths directly. Since seeing "yellow" requires two of our three different receptors to be firing that excess of signal travelling along our nerves probably overloads our sense of 'value' and makes that hue appear "lighter" other hues. You'll see the same effect happen with cyan, which requires our blue and green receptors to both be sending signals at the same time as well.
Great observation. This is quite accurate.
since our eyes are more sensitive to red and green vs blue, it'd also make sense that yellow would be the brightest, as opposed to magenta and cyan that require mixing less sensitive blue receptors
In addition, there are several ways to convert to greyscale, i suppose some handle perceptual issues and others focus more on different metrics.
I dont know... purple is two, or some says purple don't even exist, and it is darker. Red, green, and blue which are the rods, not even have same values.
or similarly that the abundance of these wavelengths in sunlight influences brightness of those pigments in natural lighting.
Johannes Itten also talked about this topic in his book. He used the word "luminosity". For example, yellow has the most luminosity of all, so it tends to be lighter in values
Yup, it has to do with the physiology of the eye.... our eyes are most adjusted for the yellow-green wavelengths
Johannes Itten is a crackhead, don’t believe anything he says
sure it is not Luma instead?
There is multiple ways you can turn a picture in shades of grey: luminance, luma, lightness and value.
the closest by design to a perceived luminosity independent of hue is the lightness from CIELab.
Lightness from HSL is also independent of hue but yellows appear much darker.
This explains so much about why I've always struggled with colour in my digital art vs traditional art. Thank you!
Unsaturated color: "You simply have less value"
Lmao
i feel like such an art nerd for laughing at this
" 39 buried, 0 found. " - Saturated color
Red's mother loves him less than Blue's mother loves her son, apparently. And we don't even TALK about Yellow's home life.
lol true 😂
I'd love to see a painting where when converted to grayscale everything is one colour, if that's even possible.
I wonder if it's possible to have a painting that converted to grayscale shows a different image. I mean I know it's possible, just don't know how to find one :P
@@Starfloofle Steganography is much broader term, but in context of graphics it usually means using the least significant bits of each color component of each pixel to hide some message. The technique was used for example in spore to encode your creature code into the creature's photo. There is also some virtual console (i forgot the name) that uses cartridges in form of images with game code encoded using steganography.
It's very interesting topic, but not exactly what I meant here ;)
Woah the comment section wise getting interesting over and over! I'd like to see too, maybe in future there's a mystery genre that would use this technique to hide some real meaning or event that secret in art medium haha! Gonna be interesting tho!
yes it is and ive done it i have 1 color of blue and black so everything is black when converted but its a beautiful bird when not.
Aizawa this you
You actually go straight to the point, no intro or backstory bs. Just what we need to learn about
Me, an artist who doesn't paint : ah yes I understand everything and will use this in all of my art /s
Why the /s?
/s means that you’re saying it in a sarcastic tone!
@@vynbit crap “serious” starts with an “s” too this must cause so much confusion. I do know that serious is /srs tho but aaaaa
@@sexfuk2119 when people are serious, they usually don't put anything at the end of their sentence
@@_Asperothh I'm a neon bitch lol. I use eye bleeding color everytime I can
Fantastic vid, as always dude!
Hoi
defollynitely!
This helps a lot when choosing what color to use when shading and lighting your piece.
People: Value, Lightness, Brightness, Luminosity.
This guy: Value, Value, Value, Value.
Exactly! I was so frustrated watching this video. He muddles together different concepts so much it’s infuriating
So that's why I'm confused the more I watch this video
literally all the same thing in the context of painting.
@@eldrickzero4885 But they are not the same in the context of computer color, yet he calls them all the same thing, which is confusing
@@prismarinestars7471 he talks about painting tho
FINALLY someone talked about it, i've noticed this stuff as well and it confused the hell out me whenever i sat down to paint
I appreciate how you demonstrated these on more than just Photoshop
This is big! Will help a lot with thumbnail color composition in the future!
Hello my son is a big fan of yours
@@gtvucy1361 How about you keep this to yourself
what the?!?!?!? mysti on this comment section? you were the last person I'd except! love your content, thx.
Hi!!!
@@gurudattachaware2589 Your a great parent. You just corrected this guy, anyways have a nice day and God bless you.
Oh my God, I had a problem with "Hue" blending mode a few weeks ago, and it had me pull my hair out. But this explains a lot.
The description of the Hue mode says: It keeps the brightness and saturation, but replaces the hue. So I was wondering, why the brightness and saturation changes after I use this mode?
I asked so many people and nobody had any idea why it happens. But now I realize, there is nothing wrong with the blending mode. It was our fundamentals that was wrong.
Great video! Thank you so much.
Something that's happened to me is that blending modes like "Hue" or "color" don't let you paint with for example, a really dark yellow, or a really Bright red without it looking like something else. Hues have a "Range of values" in which they are readable.
Color theory is hard
@@miguelcarmona3036 thats why we need to master it :p
That 'Proof Setup -> Custom -> Dot Gain 20%' (CTRL+Y to toggle) method is awesome. Thank you for everything!
Ok here's a bit of explanation on a few things you seem to be confused of.
The color picker: Your color picker is set to HSB/HSV (same thing), not to HSL (technically there is a difference between HSV and HSL - the L stands for luminance).
If you pick two colors with different Value they should in theory stay different in Value when you desaturate them, and that's exactly what happens when you use the HSL-saturation slider shown at 4:50. The HSL-saturation slider does exactly that. It decreases the saturation while keeping the hue and value staying the same.
Interestingly enough photoshop now calls it HSL (Lightness), but it's actually HSV as seen in Coral Painter. Oh it gets worse.. in Clip Studio Paint they call it luminocity (but with a V in brackets.. wth).
Anyway. Let's say you're using a different way to desaturate the image. How does the desaturation work? One way would be to extract only one channel from the RGB channels and apply it to all the other channels. Cheap digital cameras use this method when applying a build in black and white filter. They overwrite the red and blue channel with the green channel, to get a greyscale image. But that's actually a bad idea. Why? because areas with strong colors like magenta that doesnt have any green in it at all will come out as black or really dark.
Moving on what else could you do? You could convert the colors to HSL (luminance) where bright colors like yellow and cyan actually stay bright and visually darker colors like blue and violett stay darker and then strip away the saturation. That way you would at least get an accurate representation of the brightness of the colors in the greyscale image.
Actually a good way would be to convert to LAB - wich stands for Luminance, A and B (but wth is A and B???? they are color channels. look it up, it's quite interesting) and decrease the contrast in the A and B channel, that way you only keep the luminance channel. But... you need to convert between RGB and LAB and stuff and that sucks and yeah....
Where am I getting with this? Color theory, especially digital, especially color science, especially with all the f*ing weird lingo is weird and it's complcated and the more you dive into it the more confused and frustrated you get... Let's maybe just... accept that it is what it is and move on before we loose all fun at doing what we're doing? Because I mean that's why you do it right?
This should be the top comment
Mind ~= blown.
Yo what that's sick you write that😂
@Charon TheReflector I mean I don't say that knowing a lot and maybe even striving to know every little thing about it is a bad thing, but when it comes to creating art it sometimes can get really overwhelming if you spend too much time with all the technical aspects. Understanding it is great and will definetly make you better but while you're trying to understand it you should never lose the fun and joy of doing it.
HSB and HSV names the same thing, but keep in mind that not all software use the same color plotting. Picking the HSB/HSV value from one software and input that directly to another software might not yield the same color, even if both are properly color managed and set to the same color space. Same for HSL I think, but since it's more niche I think it's a bit more unified across software.
Also, it's not uncommon to adjust the greyscale conversion matrix to need. For accurate soft proofing result though, we have a few "standarized ones". But unfortunately there are multiple of those. Lab is a good one though as in math that would always be accurate, except... It has to be converted to RGB at some point eventually. And the conversion matrix is also subject to standard variance. And... We still can't see accurate color. That really sucks when think of it, but again Lab is more of a garage dragon to begin with anyway.
And... When all of above is solved, we now have HDR and wide color gamut image to deal with. And things just goes so wild that I personally just gave up and decided to go SDR and sRGB for life. Anything else I can probably have by buying a bottle of pigments I guess.
The fact that hue also affects how light or dark a portion of the painting will appear is probably why we tend to gravitate towards using cooler colors for shading and shadows and warmer colors for anything that will be bright or vibrant
omg. youre probably right
No, its not that. Shadows has less color because no light and in nature shadows are usually blueish because of the blue sky.
@@fragiledate I think it is also that we often use the sun as a light source, which makes the highlighted areas look warmer
No. Red green and blue are darker than yellow cyan magenta and white because combined colors add more light.
@@fragiledate literally came to comment that! Omg that's such a good point!
life changing marco bucci .. i watched thi svideo 15 times and toke notes. now im seeing values everywhere.. so awesome.. thank god
In a nutshell, human eye perceives some colors as darker than others because of how cone cells are arranged in it
As far as i know the human eye perceives green as brighter colors, that's why when converted to greyscale they make green the brightest colors.
This is a formula from my computer graphics course, to convert RGB to greyscale: Y = 0.299 R + 0.587 G + 0.114 B
our cone sensitivities overlap with each other, the green cone overlaps with both red and blue more than red does with blue and vice versa, and the brain reads the intensity of each of the cells combined. So green would be the most intense, however yellow is right at the overlap of green and red, making it the largest value out of all. (also blue our sensitivity is lower than other cones)
That's not really right. The names we give colors are determined by how often we see the color.
When we see yellows, we almost always see bright, highly saturated yellows, while rarely seeing dark or less saturated yellows, so we only named the shade of yellow that is bright and highly saturated.
Red and pink are named as different colors, since whenever we see bright reds, they tend to be very bright (skin, flowers etc. are usually very bright reds), and whenever we see darker reds, they tend to be much darker(blood, flowers, fruits, all usually have a dark red color). We often needed to distinguish dark and bright reds, so they were given different names.
Meanwhile, we keep seeing all different brightnesses of blue and green, and we didn't have any use for giving the different brightnesses different names, so the names were used as blanket terms for all brightnesses.
We have always been able to tell apart different colors and how bright they are. It was how often we saw different colors, and how much we needed to distinguish different shades that dictated what colors we decided to give names to. Those names stuck, and now the names are used for colors of wildly different brightnesses.
@@adrianks47 The formula you have shouldn't be named greyscale, it's miss leading. True greyscale is the average of RGB. Y = (R+G+B)/3. If it takes into consideration the human eye perceived color they should name it average human perceived greyscale.
@@HenrikMyrhaug This may have once been true but now the cardinal versions of colors are determine by math and their absolute max value hue vs each other. Red pixels at 100% show pure red light and same for green and blue. Show both red and green at 100% and you get an approximation of pure yellow.
The more experienced I get in drawing, the more I prefer to start a piece in grayscale then apply color only after values are established. My brain just processes wavelengths so unreliably.
do you do it in pencil then first?
@@pizzapuke19 I use photoshop but yeah I usually do a sketch then fill in the values.
just paint how you want and slowly incorporate new findings about color and value into your art. doesnt have to be perfect
You don’t know how long I have been looking for information like this. Thank you so much for sharing your knowledge.
I discovered this just last year but i didn't know that i could've make the same tests on all colors and connect it to value to make more accurate color choices.
Now this is a 7 mins to better painting haha
haha - I literally asked myself "How can I make a '10 Minutes To Better Painting' episode ... but shorter and with less work?"
5:35 Krita users can also use a soft-proofing. It's also under (ctrl+y) shortcut, but the grayscale colorspace has to be picked in the settings under color management tab (the default space is CMYK).
And great video by the way :)
Thanks! I was thinking I might have to write a plugin if it didn't already have the feature somewhere. This is very useful!
@@JosephDavies it could be quite hard to get something so deep in image rendering with python plug-in in Krita, but luckily she already has a quite advanced colorspace management :)
@@wojtryb Krita Master App.
What profile do you set it to? Default gives me the same effect as lowering the saturation (which is wrong according to the video)
also,
5:00 Filter layers in Krita have a selection of colour modes.
You can select the mode in the "types" dropdown.
Including HSV and HSLuma
You can also change the colour mode for the colour picker toolbar, in settings> colour picker. you can even pick a number for the delta, shift for the colour H/S/L values!
This is why Krita (for me) is the best program, and it is open source, the community and user can always make it better!
GIMP has a "Convert colors to gray" option in its desaturation menu that allows you to check values without messing with saturation at all. You can see the difference between desaturation and converting to gray. The resulting images are vastly different in value.
6:49 I love it when science and art overlap! If I remember correctly, I think eyes with normal color vision tend to be more sensitive to green light (more “green” cones than “red” and “blue”). Since red and green cones overlap in the wavelengths they pick up, yellow light appears even brighter.
this can be fixed almost entirely by switching to the LCH colour space. It's essentially the same, but hue and saturation are properly decoupled from value. This is because it's mapped to human perception of colours, not the RGB gamut. I don't know if you can do that in Photoshop, but I tend to design on the web using chroma-js which supports this colourspace.
AFAIK the reason this happens is due to the cones in your eyes having different sensitivities to wavelengths of light.
Any of the color models with an 'L' (lightness) component work for this - HSL, LAB, LCH, etc.
@@aaronbredon2948 in theory yes, but I've found that HCL matches human color perception better than at least the standard CSS HSL. Is there a version of HSL that better matches human perception? CIELAB maybe? I don't know.
@@memyself4852 the CSS color models aren't great. You really need to be using a professional color managed program.
Photoshop should be (used to be) the gold standard for color management (if you could navigate the settings)
The big issue is that you cannot maintain max lightness and max saturation at the same time, and the amount of loss varies by hue.
That is why the LAB and other models where saturation is derived work better for maintaining lightness.
Hi, I am colorblind and I found this extremely interesting. Since I often have problems differentiating or identifying specific colors, I usually rely on the values (unconciously).
For example, I sometimes can't tell if a color is green or brown, but I see that one is "darker" than the other (even if on the same saturation), and that helps me know which one is which.
I do this very often and it is usually hard to explain to others... believe me, this video made me understand how I see and why I do what I do. This will be very helpful for my art and even my everyday! Thanks!
So believe it or not, you unintentionally made the perfect video explaining how colorblind people may distinguish color.
This explains why i was never satisfied with the contrast in my paintings. Thank you so much!
HOLY GUACAMOLE, you blew the lid off this color conspiracy. The part at 6:10 just unlocked a whole world for me. Thanks Marco
Thanks for making this, so many people whine and wail about how hard it is to understand colour theory when all you need to do is turn on a Marco Bucci video..
The "bright" blue being dark in value was a revelation. It clicked something inside. Thank you!
Tip: actually in Procreate on iPad, the “Hue Saturation Brightness” slider box is effective in showing the value change in hues, it doesn’t even them all out like other programs :) (but the white fill on “color” mode is effective too)
Procreate is awesome-
Ohh that’s why I knew all of this and was confused why it’s even a question-
dang i can only wish to have procreate
@@AeroCrafts it’s only 10 bucks one time purchase so
@@arcadeii you’ll need to buy a £700+ tablet tho :(
"Different Hues, Saturate to different Values"
Really gotta remember this, love how you put that into words.
I noticed something related while doing speedpaints/studies: in real life, darker tones tends to be more saturated and vice versa.
I remember reading about how each hue has its own inherent value, yellow being the brightest and purple being the darkest. Johannes Itten actually made a brightness scale of each primary and secondary color
you literally just blew my mind . i’m pretty sure i’m never going to colour anything ever again
One of those things fine arts teachers can never teach us. I never knew this...I assumed it because for some reason all my pics turn out super dark and I can't control my lighting.
Makes sense!
Thank you!
4:49 that absolutely blew my mind. I’ve been always using reduced saturation to make images grayscale. Oh my god, thank you so much for this vid
I was always curious why some color pickers were a triangle and some were just a square (and the occasional weird circle). This makes so much more sense now!
You can use the Lab color space to accurately see the perceived brightness of your color choice! In the "swatch panels" you can see the value labeled as "L" decreasing as he picks a more saturated color.
i just see it as the “darker” colors absorb more light, the “brighter” colors reflect more light, so the darker colors will have a darker hue. kinda like color changing light bulbs. you can have the brightness set to 100%, but the white light will light up more of the room than the yellow light, yellow will light more than the green, green more than the blue, blue more than red, etc.
But what makes any of those colors "brighter" or "darker" than one another, especially when there is a brightness and darkness adjuster already built into the color pickers? That's the question the video is asking. Why are some colors brighter at the same values, and what physical properties does that relate to? My best guess is that it has to do with how the colors are created digitally using rgb, because those three colors appear generally darker while colors like yellow magenta and cyan which are made by combining colors are mostly brighter after grayscale is applied, presumably relating to how many subpixels are activated to create each color. For example the brightest yellow would be (255, 255, 0) while the brightest red would be (255, 0, 0), half as many subpixels being used. And in terms of pigment the effect would work similarly but in the opposite way, where cmy are the primary colors and are combined to make other colors. Since this form of color is subtractive, rgb would still be darkest as more light(specifically a greater amount of different wavelengths)would have to be removed to create those colors. Although the details are less obvious to me than with digital color because I don't know how the hues are created
@@Keijo_ You are on the right track. Our psychological response to a color is based on how much the color activates each of our three color "cones". Humans are naturally much more sensitive to green/lime/yellow than red or blue. We can distinguish hues more accurately there and we see them as brighter even if they are the same intensity. When a software application "greyscales" a picture, it uses a model of the human color response to determine the average cone activation, or the "luminance" of a color. A pure blue pixel (0, 0, 255) has maybe half the luminance of a pure green pixel (0, 255, 0).
Blue lightbulbs don't illuminate a room as well as other colors for the same reason. Blue light can bounce around the room, but our eyes don't pick up on it well. It has nothing to do with whether they "absorb more light". They're not absorbing light, they're producing it.
It would be the firing of two simultaneous cones perhaps and our eyes bias toward green (camera sensors tend to have many more green sensors).
Another is "yellow" is defined by its lightness. Try making 'dark yellow'.! You end up with greenish brownish mud it doesn't exist any more than 'dark pink' due to pink (not magenta!) Being a 'light red'.
Its just an issue with colour space probably. To make pink i.e a lighter shade you add white which in turn desaturates. On the terms 'value' does not figure this and only represents the max firing of a hue, whereas in other models its termed 'lightness' which by definition factors these 'tints' (add white) in and is coupled with 'chroma' not saturation. Chroma differs by being invariable for lightness... Yellow is not the same as 'light cyan' of equal chroma in lightness in such a colour model.
Its going to depend on whether we are talking about additive or subtractive colour. Be careful not to conflate the two.
Blue more then red? Or red more then blue?
incredibly well-done video! Wow. As a fellow video-maker and editor, I winced at how much time must have gone into chopping up such a snappy, to-the-point, visually CLEAR video. Excellent.
In RGB, on digital screens, it is vital to understand that there's a huge difference between the objective brightness between RGB and CMY
the reason is simple: for red we have one light active, just the R in RGB, yet for yellow we also add green G, so digital yellow, magenta and cyan are about twice as bright as red, green and blue respecticely.
add to that the specifics of how human cones perceive color and it becomes a bit messy, but that inherent distinction between the hues on digital screens is important
6:35 whoa that's smart way to deal with colour value, that can be an example to us
Oh wth your sketch suddenly made all the sense about color values!
This was an amazing example! Thanks!
Just when I thought I was starting to get the hang of colors T_T
Same!! 😂😂😂
Same
same lol
If you want it simple warm colors have lighter values and cold colors have darker values, if you divide it that way it makes it easier to find the right value you want (though it's not alwase like that but starting with these basics is what helped me more to understand)
Isn't green a cold color and magenta a warm color? Green is lighter then magenta
The brightest colors will be the ones between red and green, closer to green, which is where yellow sits. It’s the brightest because it contains both red and green. Blue is the darkest, so the darkest colors sit between blue and red, but purple is brighter than blue because it contains red, which is more luminous than blue.
In short:
Brightest - between Red and Green
Mid - between Green and Blue
Darkest - between Blue and Red
Desaturating adds more colors, so it equals in brighter colors.
Please make a painting with many different hues off colour, that all use the same value! :D That will be mind blowing
This has been boggling my mind. THANK YOU!
The art on your pfp is very pretty
@@Deedeedoodad Thank you! 🥰 It still needs cleaning up 😖
Thank you for this explanation. i was told not to just turn down saturation before but when i asked why, they couldnt explain. this makes the reasoning easy to explain and follow.
in traditional oil painting atelier classes we learn that a "neutral" green has the most "neutral" local value. while a bright "warm" yellow has the lightest local value.
39 SECONDS EARLY IM LIVING THE DREAM BABY
There you go -> 🥇🏆
this is why when i kept not being happy with my drawing i just turned greyscale on and grabbed a random color pallete and worked with the values, and i love the way it came out when i turned color back on
Manga artists: What is color? All we know are black and white
*Laught in full color cover art* No.
Grey is feeling left out
It's probably why manga cover art looks like an oblivion screenshot
THIS!!! I've been struggling with this particular issue lately and I thought it was just me being dumb and confused. Thank you for clearing the fog a bit =)
Now I understand why in many drawings shadow's hue tends to Blue/Violet, they fairly are the most "black" colours 🌟
I can’t tell you how much this will help me!
I’m a 7th grader and I love art. I’ve been doing it since 2nd grade and I’ve improved A LOT!
But I do have to say I’m not good with color theory, I just recently learned about it and it still hurts my Brian but watching this video helped me understand a lot. It will help me improve my drawings in the future and help me get more comfortable with drawing in color instead of just black and white 😅.
I’m mostly self taught. The only thing that I didn’t pick up by myself is anatomy. And I still suck at it. I would ask my art teacher for some guidance and help but I don’t have one. I’m an online student and the program I’m in only allows 4 classes to be worked in!
While I do have to admit it is great only have to worry about 4 classes it honestly sucks because art isn’t one of those classes. Like I said, I love art. It’s been my dream to become a professional artist since I was 7! Not being able to have access to someone that knows a lot about something I’m really passionate about really does suck. But i only have half a year left of online school and I’m going to be going back to real school next year! But until then I’ll stick to watching videos like this to help me understand how certain things in art work instead of watching tutorials on HOW to do something.
Videos like this really help with understanding art, instead of just copying mark by mark of some famous artist. Trust me, not something you would want to do. Biggest regret in life. It lead me nowhere in my art and left me bummed out. I soon began to realize those aren’t the kind of videos you would want to be watching if you are really passionate in your work and want to learn something new! While those videos can be useful for some, it just wasn’t right for me you know? I wanted to feel like I was actually learning something useful that I could use in drawings/ paintings. Those videos just weren’t what I was looking for. Not saying they aren’t useful. Or that you shouldn’t watch them, they just weren’t right for ME. You know? 😅
Sorry for my little ramble 😅. Again, this will help me tons!!
it's been one year, how are you doing now /genuine question
same
I"m curious too! Is there a place where you've been sharing your art?
i remember watching a video a while back on how to draw "hyperpop" and it showed values of each color completely saturated.
i found this video just recently and it's nice to see more things about it
If you create a spectrum starts from blue ends with turquoise you will find that it is gradually getting lighter. This is why a lot of artists change their hues a little towards blue/purple to get a more dynamic feel in their piece.
In Krita, you can open up a new view alongside the one you're currently working on, and have two different windows of the same drawing, one in grayscale and one in full color, and they both update in real time.
THIS IS EXACTLY WHAT I NEEDED BRO OMFG
'This is not just digital'
*greyscales reality*
'See?'
Are you reading my mind???? Just thought of this recently!!! When I go darker and slightly more saturated the colors end up muddy, so afterwards I go slightly lighter and more saturated, much better results! Oh just saw the thumbnail didn't even watch yet lol
Same! I was explaining this to my friend when we took a road trip yesterday!
@@mfrobles93 haha... very funny. If you're serious, you're lucky. :(
The book "Color and Light" by James Gurney talks about this phenomenon called "peak chroma value". So essentially whenever you have a hue, that hue's strongest chroma will be located on a spot in a value range from light to dark. Yellow's strongest chroma is always a light value, red's strongest chroma is a mid, and blue's strongest is a dark. If you paint many gradations of chroma and value of each of those colors you'll create a graph that visually shows where the chromas concentrate on the value scale.
I KNEW IT! I didn't know how to articulate this, but I've suspected this for some time now... >_>' I've run into this dilemma pretty consistently and had no clue how to adjust for it.
Thank you for illuminating this weirdly specific phenomenon.
1:14 "you know like value is determined by this vertical placement and then you plug in hue and saturation after" nah bro I just randomly click on points of the square until the color that comes out look cool
it's one the only video about colour value that made me understand it, it actually blew my mind too ! incredible
Wow and i was thinking it’s just in rgb images
When converting rgb image to greyscale , the color channels contribution will be like this
30% red 59% green and 11% blue
and averaging the the value by 3
Thanks for the art and science lesson 🙏
this is something i could never find the right words to describe when i was doing my art research final on advanced color! thank you for the satisfaction
Dude good luck on your paper.
wow! what an incredible video!!! i've been studying colours more then never this year and this video came like a cherry on top, but if the cherry whas the size of a watermellon
"it's not just digital"
5 seconds later
"when i switch this frame to greyscale"
aight that's it i'm heading out
filthy heretech
He turned the image greyscale with his own mind.
Take a picture of it and convert that to greyscale using a filter
i mean what other way would there be to do that? Hire a colorblind that sees everything in grayscale?
@@raidev_ grayscale is a digital concept in the first place
2:46 great now even eyedropper tool is deceptive
and those digital artists who paint initially in grayscale and then add color in after are gonna really need this info
when colors weren't hard already, that example with purple vs yellow blew my mind too
You’re gonna be the first person to make a fully accurate color picker, aren’t you?
thats gonna be a bit difficult tho. You'd need to use a different hue arrangement in the first place, based on evolutionary bias(green is the most sensitive color to our eyes, so its the least grey, red being second and blue being third), then you need to find an arrangement where blue, red and green together are the brightest grey while blue alone is the darkest.
Essentially, its impossible to do on a 2d scale, and its going to be fairly though to create a 3d scale thats easy enough to work with. pnghut.com/png/CvDbv7RDGC/hsl-and-hsv-color-space-barvni-model-picker-james-clerk-maxwell-transparent-png is basically the most accurate, easy to use 3d shape that fills picking color to full satisfaction, but its not exactly easy to work with unless you create a VR room.
There's nothing inaccurate about the way color pickers are right now, it's just people make assumptions about them that just aren't true.
@Kyaru Momochi Depends which hue you are talking about. If you are talking about computer hue, then there are accurate color pickers. Accurate doesn't mean corresponding exactly to reality or peception.
@Kyaru Momochi Something is not flawed just because it doesn't contain every single color a human eye can see. Sometimes I don't need to know what a picture looks like printed out, but only what it looks like on a computer screen, and in those cases color pickers are not flawed, they show exactly what you what, what it will look like on a screen.
Sometimes I just need to see all the colors the computer can produce, not all the colors I can see. A color picker that does that is not flawed, its doing exactly what I want.
Wow I caused a scientific debate
The value assigned to colors when you switch to B&W were determined experimentally by showing people colors and asking them how bright they think they are, but you could also think of them as being entirely arbitrary.
In terms of art, framing human color perception as arbitrary feels completely useless. Afaik the grayscale conversion algorithm shown here isn't based on self-reporting but physical evidence from the distribution of cones in human eyes. This video is very misleading though, should've just explained how we perceive color..
@@terryriley6410 Maybe he doesn't understand how we perceive color. No one does, that's why it's called color "theory" and not color "law." And how can the video be misleading when he didn't make any claims about how we perceive color? He just said the software is misleading. Which it is...
Everything else was just him speculating and asking why things are the way they are.
@@Kirito_2016 theory means it is rigorously proven.
@@TheYeetedMeatbut the proof may not always be true, there are a lot of examples in science when the theory turned out to be erroneous, although it seemed logical
@@lkoyumil and did I say otherwise? It’s just that most people misinterpret theory as hypothesis and then use that to try and claim that alternative things should be taught despite being obviously false.
Just a quick head up, YOU CAN DO SOMETHING SIMILAR TO THE PROOF STUFF IN KRITA. Go to image>change color space>soft proofing>choose the color space to whatever color space is convenient (I just use grayscale). Or if you want it to work across all your files go to settings>color management>soft proffing>choose the color space to any gray space.
Short cut will be Ctrl+Y
There are other ways to do it that use LUT management but you will be better off just reading the manual to know what it is.
The different algorithms for convert to grayscale by either “convert to bw” or “decrease saturation” is not explained, yet everything centers on it. Also I miss the obvious gamma discussion when comparing how we perceive the linear color space when it’s modified.
Funnily enough, I learnt this when studying scuba diving.
Not only value, is brightness.
Wow! I was actually struggling with this exact thing the other day and thought it was me. Thanks for the explanation!
"This stuff happens with these *real* acrylic pigments too, it's not just digital!"
Proceeds at showing a digital B/W pic of the colors
That drove me insane!!
I think you may have missed his point.
@@arkarts2987 On the contrary, I think Andrea's comment is rather on point. The Value computations used by Photoshop et al. attempt to mimic human perception of color, so the "something strange you should know about color" is more about human perception than digital painting, though the disagreement between Photoshops' formulas hints at this perceptual thing in an interesting way.
Marco's statement that "this stuff happens with real acrylic pigments too" is slightly silly. _Of course_ the blue acrylic would have less Value than the red acrylic because he's using _the same_ algorithm to compute Value. It's silly (but only slightly) because he brings up this example to demonstrate the inherent nature of Value, as if he has applied different procedures to discover the same Value difference. But in both cases the procedure is the same, the only difference being that the color data in the image is extracted from the photo instead of generated by the software.
So, like how the disagreement between Photoshop's algorithms hints at the non-intrinsic and very human-centric nature of Value, Andrea's comment hints at the weirdness of this part of the video: Marco presents the acrylic demonstration *as if* it shows something new. If anything it just shows that his camera accurately represents color (for our purposes) and that his computer can talk to it.
@@pwhqngl0evzeg7z37 I understand that completely and you explained it very well. Although there isn’t really any other way of testing it unless you got people who are completely colorblind and surveyed them about their observations for the colors, but that form of color blindness isn’t very common and probably wouldn’t further prove anything. So maybe he could have just skipped that part of the video completely because it’s probably unlikely anyone would bring up “well what about physical pigments?” And for the few people who do bring it up this exact conversation would probably arise. So it’s hard to criticize his method of proving his point because there aren’t many other ways of testing it but at the same time the method is still slightly flawed, that’s my conclusion.
@@arkarts2987 Yeah I agree. What he's trying to prove is kinda impossible to prove, except by appealing to an authority, in this case (I assume) the neuro/bio/psychological consultants for the creation of the Value formula, and trusting that they did it right. The only thing that I can think of as another way of proving this would be to make a mechanical replica of the eye, but the accuracy of this is essentially no better than the formula. Interesting thoughts nonetheless.
This phenomenon is explained in short but amazing detail in the book "Artists' Master Series: Color and Light" by Pickard. It has to do with the rods and cones in our eyes that our brains use to perceive color.
After explaining _why_ colors have inherent values, it gave patterns and tips for how to remember and make use of it: colors are lighter toward green (though green is not lightest, yellow is, and blue is darkest), and you can get lighter colors in your artwork not just by adding white/desaturating (which will also lower the chroma or intensity of the color), but by adding a different hue which is inherently bright (which will allow you to control the color intensity you want in your overall painting composition).
I've just found this video. Maybe someone has mentioned the following already: The RGB values could help to understand what's happening here. The RGB values are what really makes up your digital color. The HSL sliders are ways to maneuver within this colorspace, like cars with different properties driving on roads, so every tool (car) can lead to different results. But the RGB values are what the streets are made of, so to speak. Each RGB channel is its own brightness / intensity. And the easy-to-understand part in comparison to HSL is, that you can add each of them. It's pretty linear / straight forward. 255 is max brightness for each channel 255+255+255 RGB is the max, pure white. 0+0+0 is black. The RGB values also explain your gray vs violet example. The gray is 128+128+128 which is more than the violet with 77+0+255 in total "intensity", so it's the brighter color. We do need the HSL (or similar systems) to navigate within a 3D color space intuitively, because no one can just perfectly assume RGB values. ("hey give me that slightly desaturated dark violet", " oh you want an 118-76-119, I knew it straight away" not going to happen) The way the wavelengths of light add up is not that intuitive to work with. That red light and green light add up to yellow and are in total brighter than the separate colors. 255+0+0 (red light) added to 0+255+0 (green light) will end up as 255+255+0. The RGB system and colorspaces are a really precise way to describe the characteristic of the addition of light wavelengths and also works well in a digital code. That's why it also applies to the photo of your acrylic pigments. The lights wavelengths add up in the real world, and the RGB system describes it fairly accurately. All cameras save their images in RGB systems, professional cameras with more differentiations between the darkest value and brightest value. 0 - 255 are 256 possible values, which is 2^8 referring to 8bit. Most professional cameras are nowadays in the 14 bit range, so they have 16384 different values between dark and bright. But for easier working they are converted to 8 bit in the process of developing the digital image on a pc. The RGB System is everywhere where light is emitted. Your Monitor is working in the same way and adding up values of single red, green and blue diodes to create a pixel. The HSL and colorwheels are needed to give the RGB colorspaces a user-friendly and intuitive way of using it.
They are not using the addition of wavelengths to describe a color. They let you use the mentioned tools of Hue, Satuarition and Value (Brightness) to describe a color intuitively and they try to match it / translate it to a color from the RGB color space. This process can always have it's minor faults or inconsistencies. Standard RGB (sRGB) is the most common RGB colorspace, but there are many more. For example: Some are bigger and allow for extremer saturations for professional use. But they all follow the same principle. A different system is needed to describe colors for printing. When Light is not emitted, but only reflected. Here you need CMYK (cyan, magenta, yellow and key (black). It works the other way around. It's not adding up, it's subtracting. You start with a blank paper, it is white, because it can reflect every RGB color from the emitted light it was hit by. If you add Cyan color on the paper, the red part of the incoming RGB light is absorbed and only green and blue will be reflected, blue and green light wavelengths will add up again to cyan. So you are removing wavelengths here every time you add a color. We will go full circle when you understand that the primary colors in printing (cyan, yellow, magenta) are the secondary colors in RGB (red+green=yellow, green+blue=cyan, blue+red=magenta) and vice versa. (Add Cyan and yellow on a sheet and green will be seen, because the green wavelength is the only one not absorbed)
Not that simple. It's not how it works for humans. (0,0,255) is darker than (255,0,0) It's shown clearly at the end of the video. That's because the blue component adds less to the overall (perceived) brightness than the red component.
So for the RGB device (monitor) that's emitting light, (0,0,255) and (255,0,0) may have the same energy, but for the human eye blue looks darker. That's what the video is about.
You also claim "it's pretty linear", again that's true from the perspective of the light emitting device, not the human eye.
HSV is just a different way to navigate the RGB space, but it's still RGB space, designed to describe how machine represent colors not how the human eye percieves them. For the same reason it's wrong to think that (255,0,0) is as bright as (0,0,255), moving for constant values of V leads to changes in percieved brightness, even if, moving along V, only brightness changes.
The point of the video is that while V changes only brightness, that doesn't mean that it's the only axis that changes brightness. Both H and S change it too.
For a color space aimed at representing colors the way the human eyes sees them, check out L*a*b*.
theres an interesting concept called luma, which is a formula that weighs every rgb color channel by a percentage to get a single value that describes perceived brigthness. commonly its calculated as luma = 0,21*R + 0.72*G + 0.07*B, which returns a value between 0-255 since the factors add up to 1. low numbers suggest darker colors and higher numbers brighter. very useful if you want to determine if text color should be black or white to be readable on a screen with variable background colors
@@leckererlurch237 That's still an approximation, it's based on an average contribution of each channel to the overall brightness. It is good and useful, and way simpler than perform colospace transformation, which is not necessarily linear.
So I oddly learnt about this in one of my computer science modules (possibly computer vision). Its got to do with the way we perceive colours. We have 3 receptors to see, red green and blue. Green is the most sensitive, red sort of in the middle and blue the least sensitive. The way we convert from RGB to greyscale is based on the way we see the different colours and if you looks up the formula used to convert to greyscale its Gray = 0.299R + 0.587G + 0.114B.
Im just looking at the formula for converting RGB to HSV and it doesn't really make sense to me... Id have guessed that it would be the same as the conversion but turns out that would be the lightness value and there are different conversion standards... the above is SDTV standard (aka recommendation 601). Adobe used Rec. 240 and it is Y = 0.212R + 0.701G + 0.086B. HDTV used Rec. 709 which is Y = 0.2126R + 0.7152G + 0.0722B. UHD and HDR used Rec.2020 which is Y = 0.2627R + 0.678G + 0.0593B.
Interesting!
To get the value of HSV it its the intensity of the colour thats has the highest intensity so which is bizarre to me because that implies that if as long as you only change the values of colour channels that aren't the highest intensity the value won't change? (255, 0, 0) will have the same value as (255, 200, 200)?
Back to to the brightness if you have pure blue (0, 0, 255) and pure green at 20% (0,51,0) they would have the same lightness which according to my quick test pinta (I don't have photoshop) yes they do as they grayscale for both are (29,29,29). For what its worth to get red to the same value would be red at 38% so (97,0,0)
All very interesting stuff and my mind is officially blown. I wonder if there are any paintings where there are secret messages hidden that can be revealed merely by putting it into greyscale. I know that there is a way of adding secret message to images by altering the RGB value by 1 and then when the intended receiver uses a program to compare the original to the one they were sent they can get the binary from it and decode the message, pretty genius actually. Called steganography and technically hiding a message within a picture that can be seen by making grey scale would also be a form of steganography
Rec.709 and Rec.2020 are using a different standard of "red", "green" and "blue". Rec.709 primaries documented the capacity of CRT monitors in 1990s. Mainstream computer photos and videos also use this same set of RGB primaries. Meanwhile, Rec.2020 specified a set of purer "red", "green" and "blue" than what we had in the past, using a single wavelength light. (piano vs. tuning fork)
The recent iPhones are taking photos and videos in P3 color space. They used the same "blue" of mainstream but a purer "red" and "green", albeit not as pure as Rec.2020.
Another factor is gamma. sRGB, what mainstream computer photos standardized on, is slightly different from Rec.709 here. But that doesn't matter. What's matter is that "20% brightness" doesn't mean a 255*20% RGB value. You need to put the 20% brightness into the gamma curve. A (128, 128, 128) pixel is emitting significantly less than 50% light of a (255, 255, 255) pixel. (We feel 50% because human perception is not linear.) Uncalibrated LCD monitors and TVs may or may not fit the curve exactly as specified but they still roughly follow suit. GIMP converts (0, 0, 255) into (70, 70, 70) for grayscale. Your (29, 29, 29) looks strange.
@@billyswong From my little bit of game dev experience I've learnt, to properly simulate perceived audio volume, you'd need to logarithmically increase the value. Could the same be true for perceived brightness?
@@keppycs Overall, yes, else we would perceive objects changing colour depending on the amount of sunshine.
But the situation is in fact more complicated as part of the logarithmic / exponential perception is achieved by compensation in our mind, not by the eyes themselves. Thus a series of optical illusion and tricks is possible by crafting special images.