I'm wondering if it may be useful to start utilizing specific [number]_tone_* tags, such as three_tone_*, four_tone_*, five_tone_*, etc. This can help to find posts by using the specific number of colors/tones that are present, in a similar fashion to the two_tone_* group of tags (two_tone_fur, two_tone_hair, etc). I decided to make this as a result of topic #27120.

I don’t know, sounds nice in theory but not practical when tagging. Imagine the number of n_tone_hair/tail/ears/feathers/wings/beak/mane/eyes/butt/markings/clothing/etc. tags.

Also I don’t think people would be remembering/searching with the number of colours on a character, as compared to the actual colours themselves.

Of course, the same thing could be said for the multi_* tags and their respective numbers (e.g., multi_horn & 4_horns), but those are distinct enough on a character to warrant such a tag. Not so much so for the number of colours on a character though.

It doesn't really add any more complexity. The two tone tag already exists, and n_tone tags will not be any more complex. I just need to start filling those tags, and I can do so rather quickly if I chose to do so.

I want to embrace this change and I want to hit the ground running to make it a reality.

e621 doesn't even support counting more than two people in an image (trio and quartet are aliased to group), I'm not sure it's ready for more than two tones.

strikerman said:
e621 doesn't even support counting more than two people in an image (trio and quartet are aliased to group), I'm not sure it's ready for more than two tones.

This. While I'm not necessarily opposed to more color tags, I also don't think it should be a priority over something like this. I feel like trio and quartet would be much more useful for tagging and searching than how many colors happens to be on a character's ears or tail. But I'm also of the opinion that there doesn't need to be a tag for every number possible, in regards to both colors and character count. If more tags are unaliased, then I'd propose just leaving it at 3-4, and leave the rest to multicolored_* and group.

strikerman said:
e621 doesn't even support counting more than two people in an image (trio and quartet are aliased to group), I'm not sure it's ready for more than two tones.

There is threesome and foursome, though of course they only apply to sexual situations.

I am in favour of this one, it would be helpful to specific things, like theses:

post #1191344 post #2325292 post #2124564

there is like, eight_tone_tail, seven or eight_tone_fur, seven_tone_pawpads, six or seven_tone_horn (because of cyan) and since body parts are present in like... 99% of posts, it would be pretty useful to expand numbers. I dont see any harm on having it while also maybe having them imply multicolored_*

We have like... 1 to 10 tags for eyes, arms, legs, fingers, penises and so on, and they even have aliases/implications, someone foresaw that for sure.

So if we going to have specific number for that my suggestion would be a max of ten_tone, just like body parts

laranja said:
We have like... 1 to 10 tags for eyes, arms, legs, fingers, penises and so on, and they even have aliases/implications, someone foresaw that for sure.

So if we going to have specific number for that my suggestion would be a max of ten_tone, just like body parts

I think most of those go up to 10 because they generally come in pairs. I'd be find with leaving it as it currently is, but if we do expand it I don't think we need to go over maybe five_tone_*. I'm sure the vast majority that go above that would likely fall under rainbow_*.

Also, if we do expand the tags, should we change them to #_tone_*, e.g. 2_tone_fur, to match the other count tags?

While I lean more in favor of utilizing the [number]_tone_* tags, I do see some of the cons to using them as well.

If these tags are used, they should definitely implicate their equivocal multicolored_* tags. Though I'm leaning more in favor of using the two_tone_* formatting over the 2_tone_* style, I would be fine with either format.

This seems rather fraught with problems. How often is someone going to know the specific number of colors a character has when searching for a post with them? Presumable three_tone_* wouldn't imply two_tone_*, so if you're off by just one, you won't find what you're looking for. There's also that it can be easy to tag wrong; notably, two_tone_fur is used with "distinct colors. Not to be used when only two shades of one color are used. (Example: don't use two_tone_fur if fur is just light-blue and dark-blue)". The latest two posts with it seem to be incorrectly tagged to me, with a character having two shades of brown/orange, and another character having two shades of blue (a lot of other posts seem to show the same mistake). It's not always clear when a color counts as a distinct color separate from the others. And how should white and black should be handled? Given that they aren't 'tones' per-se, but rather all colors (low saturation high luminosity) or no colors (no luminosity); should light blue and white be considered two distinct tones, or two shades of one color? How about if some people see a slight hue on what others see as pure white or black? Do we need to have a color picker tool at the ready to count it correctly? To say nothing about how objects can be influenced by lighting, making some people see more body colors where others see the base body color being slightly adjusted by external light.

watsit said:
There's also that it can be easy to tag wrong; notably, two_tone_fur is used with "distinct colors. Not to be used when only two shades of one color are used. (Example: don't use two_tone_fur if fur is just light-blue and dark-blue)".

ah crap

strikerman said:
ah crap

It seems even the site's own implications aren't entirely aligned with this meaning, either. For example, dipstick_tail (with the prominent example of eevee) implies multicolored_tail, which is "a tail with at least two different colors". But eevee's tail (and I'd wager most dipstick_tails) is clearly two shades of one color. So who knows what the distinction is supposed to be.

Eevee can work fine because the site does differentiate brown_fur and tan_fur, but that really just draws back to what you said about different levels of saturation and luminosity, given the close relationship between orange, brown, and tan.

• I think posts where it's ambiguous how many different colors are present should just be tagged with multicolored_*. It's better to keep those posts tagged less specifically instead of potentially mistagging them.
• Adjustments caused by lighting usually isn't a problem (at least from my experience) when determining the amount of colors or tones present.
• For posts with different hues/tones of the same color, I'm not quite sure. There was some talk about a multi_tone_* group of tags (though I think multitone_* or multitonal_* may be a better format), to help distinguish between 'multiple colors present' and 'multiple tones of the same color present'. I'm guessing this could be used in conjunction with the multicolored_* tags, but I do wonder what tagging complications could arise from that. Alternatively, the multicolored_* tags could be adjusted to apply to 'multiple colors' and 'multiple tones of the same color', but that would probably come with its own tagging complications too.

d.d.m. said:

• I think posts where it's ambiguous how many different colors are present should just be tagged with multicolored_*. It's better to keep those posts tagged less specifically instead of potentially mistagging them.

Sounds like an easy avenue for disagreements. I still don't see much usefulness of specific color counts for searching, especially if different people may have a different perception of the count.

d.d.m. said:

• Adjustments caused by lighting usually isn't a problem (at least from my experience) when determining the amount of colors or tones present.

In my experience, it's not terribly uncommon that there's ambiguity in whether a subtle color shading is a result of an external light source, or unique body coloring (at least without additional/external context to know how the character is normally depicted, which violates TWYS).

d.d.m. said:

• For posts with different hues/tones of the same color, I'm not quite sure. There was some talk about a multi_tone_* group of tags (though I think multitone_* or multitonal_* may be a better format), to help distinguish between 'multiple colors present' and 'multiple tones of the same color present'.

Separating multitone_ and multicolor_ makes it even more confusing to properly tag, let alone search, in my opinion. Hue/tone is the shift along the red->green->blue->red color space. Changing the saturation or brightness can change the perceived color without changing the hue/tone, but whether or not a change in hue/tone changes the color depends on your ability to distinguish the difference. Ultimately the problem is the dividing line; orange and brown, for example, are the same color , but our brains are trained to perceive it differently depending on the surroundings. Yet this site has different tags for them, so does it influence the color count? On the reverse side, this site doesn't distinguish cyan from blue for tagging, but they're distinctly different tones/colors, so does it count toward the tone and/or color count?

watsit said:
This seems rather fraught with problems. How often is someone going to know the specific number of colors a character has when searching for a post with them?

I'm fairly confused by your question. If you're looking for three distinct colors present on one character, search for that. It will be undertagged, but you'll get something. When it comes to tagging specific colors, try your best. I have to correct two tone images that are more than two tones. It happens, and it is not a big deal that it happens.

watsit said:
There's also that it can be easy to tag wrong; notably, two_tone_fur is used with "distinct colors. Not to be used when only two shades of one color are used. (Example: don't use two_tone_fur if fur is just light-blue and dark-blue)". The latest two posts with it seem to be incorrectly tagged to me, with a character having two shades of brown/orange, and another character having two shades of blue (a lot of other posts seem to show the same mistake).

The first example has two characters in it. One with blue, and white fur , and the other with yellow, tan, brown, and orange fur (four tones). The second example has a character with orange fur, and a very light yellow, but it's so close to white, it could be tagged as such. The two tone fur tag should apply in both cases.

watsit said:
It's not always clear when a color counts as a distinct color separate from the others. And how should white and black should be handled? Given that they aren't 'tones' per-se, but rather all colors (low saturation high luminosity) or no colors (no luminosity); should light blue and white be considered two distinct tones, or two shades of one color? How about if some people see a slight hue on what others see as pure white or black? Do we need to have a color picker tool at the ready to count it correctly? To say nothing about how objects can be influenced by lighting, making some people see more body colors where others see the base body color being slightly adjusted by external light.

The rest of your post seems like you are over-complicating things. These considerations affect the color tags as a whole, including existing ones. Most probably don't take into consideration light changes in hue and wouldn't be sure enough to tag as such anyways. This isn't a complication. This issue affects monotone, and two_tone just the same. Although if you would tag a color as yellow, instead of white, then you are claiming that it is distinct enough from white to count as a distinct color tone.

So if you consider light blue as blue and not white when you tag, this means that white is a separately distinct color if it is also present. Tan is not tagged brown. Brown is not tagged orange. Pink is not tagged red. These are distinct color tones as per our current system. We defined them to be distinct regardless of if they exist in the color spectrum.

These tags are intended mainly for convenience. It is better to have the colors separated into their own slots so people can have quick access to specific multicolored examples. Sure there will be some mistags and grey areas, but those affect every color tag. I'm sure there will be less "false-positive" results when searching for a specific color count than it is to find that specific color count using multicolored, which is the only option currently.

d.d.m. said:

• For posts with different hues/tones of the same color, I'm not quite sure. There was some talk about a multi_tone_* group of tags (though I think multitone_* or multitonal_* may be a better format), to help distinguish between 'multiple colors present' and 'multiple tones of the same color present'. I'm guessing this could be used in conjunction with the multicolored_* tags, but I do wonder what tagging complications could arise from that. Alternatively, the multicolored_* tags could be adjusted to apply to 'multiple colors' and 'multiple tones of the same color', but that would probably come with its own tagging complications too.

Are you suggesting you repurpose the tag groups I'm using for more than 2 colors for something else?

I think you are mistaking the word tone for shade. A tone is for distinctly different colors, and is currently not used to refer to hues or shades of the same color except for the ones we have defined to be distinct (grey, tan, brown, pink).

I think it may be too late to redefine what two_tone means.

thevileone said:
I'm fairly confused by your question. If you're looking for three distinct colors present on one character, search for that.

I more mean, how often is someone going to know the specific number of colors a character in a post has to search for it? For example, I know the characters in laranja's examples have multiple colors, but I couldn't begin to guess the actual number without referring back to the posts to count them out. In other words, if you need to search for a post, it's not likely you'll be able to recall the specific number of colors to search for, making the tags' usefulness questionable (assuming the post was even tagged with the proper count in the first place, making its reliability questionable too).

thevileone said:
The first example has two characters in it. One with black (or very dark blue if you wanted to be more accurate) fur, and the other with yellow, tan, brown, and orange fur. It would be a four tone character.

The left character has dark blue and light blue fur, the very example the wiki gives for what doesn't count. The right character has brown and light brown (tan), which is the same in principle (different shades of the same color), just with the difference that tan is distinctly tagged here. If that's the distinction being drawn, that it depends on the number of tagged colors rather than the number of visually distinct hues, it may be useful to clarify the wiki in that regard. I'd also suggest to rename them to n_color_fur, n_color_body, etc, since a color's tone is more associated with its hue, that is excluding saturation and value/brightness differences, whereas "color" is a more nebulous term that relies on more arbitrary divisions anyway.

Though either way, I don't see any yellow fur. If you're counting the gems, they actually look more green to me, but they aren't fur. That further complicates this actually, if the gems would be considered part of her body; three-tone fur, but four-tone body? Implications of fur/feathers/etc to body with a specific number of colors won't always be valid, further reducing the likelihood of proper tagging.

thevileone said:
The second example has a character with orange fur, and a very light yellow, but it's so close to white, it could be tagged as such. The two tone fur tag should apply in both cases.

I'm seeing brown and an almost-but-not-quite-white tan color. But even still, the penis and nipples are part of their body and are a fleshy pink, so even if it is two-tone yellow and white fur, the body as a whole has three tones -- yellow, white, and pink -- as their body includes more than just fur.

People confuse shades of yellow with shades of tan all of the time. I'm pretty sure about what I'm seeing after I took some extra time to look at it.

post #2392090

This depicts a two tone character and a four tone character. I thought the color was actually black on my monitor, but I confirmed with my friend that it stays blue, so it is a two tone character, but not all monitors will show this. Adding highlights to a dark blue color does not equate to a separate tone under the current definition.

The outer facial color of the second character is what tan should look like. The inner color containing the snout is yellow, not tan.

Actually I think there may be a bit of white around her ears, which would make that character five tones, not four.

post #2392115

This depicts a two tone character featuring orange (too light to be brown), and a yellowish white color. It could possibly be called a tan, but we call a lot of colors that are more yellow, and white, than brown tan, because it is cool to do so I guess. It's even tagged as white fur in the image by someone. They tagged the horns as white too, but those are light yellow. Too yellow to be in the white group.

As to if I feel that high contrast colors should be treated as distinct, I agree. If the artist intended to distinguish between multiple fur color layers by having two different variants on the same color, it is distinct enough IMO. It makes less sense to portray a more complexly colored character as if it was simply colored, because the colors haven't crossed into another color group.

Edit: Silly me, I didn't even notice her tail. Yea, it is three tone. Orange, white, and brown.

thevileone said:
post #2392090

This depicts a three tone character and a four tone character. The artist was trying to create the illusion of a dark blue character by adding a black base layer and then including bits of blue highlighting in it. So the black and blue are distinctly separated.

I don't see it. The darker colors on the first/left character are clearly blue to me, a deep blue but blue nonetheless. Some parts of it may technically reach black (maybe, I don't have my color-picker at the ready to verify any #000000 hex color pixels), but that would only be due to the lack of enough light, not any apparent difference in the fur itself. Or do you mean the panties or thongs they appear to be wearing?

thevileone said:
The outer facial color of the second character is what tan should look like. The inner color containing the snout is yellow, not tan.

You mean the bridge of the nose and above the eyes is tan, while the sides of the muzzle and cheeks are yellow? I'd have called them brown and tan, respectively. Either way, I'm still having trouble seeing orange; at best, I recognize the tiger-stripe pattern on her and my brain knows to associate tigers with orange fur, but the rest of the image is bright enough to make my brain interpret it as brown, leading to confusion.

thevileone said:
Actually I think there may be a bit of white around her ears, which would make that character five tones, not four.

You sure that's not the smoke from the blue character's cigarette?

thevileone said:
This depicts a two tone character featuring orange (too light to be brown), and a yellowish white color. It could possibly be called a tan, but we call a lot of tags that are more yellow, and white, than brown tan, because it is cool to do so I guess. It's even tagged as white fur in the image by someone. They tagged the horns as white too, but those are light yellow. too yellow to be in the white group.

Color perception is far from straight-forward. There are elements of culture and language that go into a person recognizing different colors, as well as general familiarity. Someone who works with colors a lot may be able to more precisely identify colors than laymen. It can also be influenced by (im)properly calibrated or misconfigured monitors, room lighting, and subtle forms of colorblindness. I take a pretty lenient stance on how others identify and tag colors vs what I would've tagged, but consequently I see that causing problems with tagging specific color counts, since different people see things differently to some degree.

I edited my post quite a bit since you made that. As you said brown is orange, so if you see brown, you must also be seeing orange. I do use a color picker to help me avoid some of the pitfalls of using a monitor that can lie to you. Shows me black instead of blue, but I thought it was dark blue at the start, so heh.

Facial colors as I see them. Even the color picker says that region is tan or near tan.

https://i.imgur.com/qHgdO2W.jpg

Color picker says dark salmon or something, but I know these to be along the border group of what is normally consider brown and what is normally considered orange. It has to be one or the other, and it is closer to orange IMO. It looks like the orange I see on tigers to me.

https://i.imgur.com/ieYQNaN.jpg

watsit said:
Some parts of it may technically reach black (maybe, I don't have my color-picker at the ready to verify any #000000 hex color pixels), but that would only be due to the lack of enough light, not any apparent difference in the fur itself.

Regardless of lighting or shadows or whatever, I tend to just tag colors by whatever it looks like. Even if the character's fur isn't strictly meant to be 'black', the fact that it looks black means that other people may very well search for black_fur in reference to the image. Tag what you see and all that, better to potentially have too many tags rather than too few tags IMO.

strikerman said:
Regardless of lighting or shadows or whatever, I tend to just tag colors by whatever it looks like. Even if the character's fur isn't strictly meant to be 'black', the fact that it looks black means that other people may very well search for black_fur in reference to the image. Tag what you see and all that, better to potentially have too many tags rather than too few tags IMO.

That's what I'm saying, though. It doesn't look black to me, it looks dark blue. Some specific pixels may technically be black, but I'd search for and call that "blue_fur", not "black_fur".

watsit said:
That's what I'm saying, though. It doesn't look black to me, it looks dark blue. Some specific pixels may technically be black, but I'd search for and call that "blue_fur", not "black_fur".

Honestly I would've tagged that specific image with black. It's a very dark blue, too dark for me to be comfortable (at least solely) with blue.

Honestly, to me the husky in the first picture looks black and white. Just black and white. The blue seems to be a lighting effect and nothing more. The character next to the husky looks brown, orange and tan. Brown as the primary color + rosette/stripe color, and tan as the countershading. The orange on the shoulder does genuinely appear to be orange, and not just a lighting effect, unlike the "blue" on the husky. I don't see white or yellow on that character.

The second picture looks like a dark tan and white to me. With brown hair and tail tuft, and pale yellow horns. Only mentioning the horns since they got brought up. Otherwise I'm only really meaning the fur/hair.

This is how I would have tagged both of these pictures. Not sure what that says about my monitor, or me as an artist if that's considered wrong, but that's what I see.

As a side note: Using a color picker is also not always reliable because what surrounds a color can make it appear different to the eye than what it actually is. For example, I have a friend who uses a brown for one of his character's jackets. Yet, because of all the reds and other warm colors he uses, the jacket winds up looking more green than brown. Eyes are weird.

god is this a " what color is the dress " moment