All Things Digitigrade Legs


AkuTenshiiZero

 

Posted

Quote:
Originally Posted by SpittingTrashcan View Post
As to how a humanoid bipedal digitigrade would move, well. District 9 did a pretty good job of it, given that the "prawns" were CGI-rendered on top of human actors. In terms of crouching, digitigrade ankles tend to have a greater forward swing and a lesser rearward swing than human ankles. So where humans kneel by tucking their feet under their buttocks, a digitigrade would instead fold their feet forward to lie under their shins. This also has the advantage of allowing you to pop straight up into standing.
So, what you're saying is I should watch District 9 again? Possibly with subtitles this time? Good idea, actually.

So, what you're saying is that while humans would crouch as we do, a humanoid bipedal digitigrade would crouch like how a cat sits? Let me see if I get this straight: This would involve putting the entire foot, toe to heal, on the ground, then folding the shin and thigh over it and folded outwards with the body standing BETWEEN them? Hmm... That might actually look weird on a humanoid frame, but it does make sense.

Quote:
In fact, a lot of the variations humans have on sitting are essentially workarounds to the fact that we can't just comfortably fold up our feet directly under our legs, because our thighs and shins are freakishly long compared to our metatarsals, and our hips are twisted in order to put our thighs vertical. The entire human body shows evidence of having been pulled from a more bendy arrangement into a vertical alignment, because of course this is precisely what happened.
So, what you're saying is that when a human's butt is on the ground, there's no good position for the legs that could put weight on them? Huh... That actually makes sense. I remember trying to think of a good way to sit on the ground when I was a kid, and I know how horribly unstable people are with any stance like the above-describe, not unless you have your back to a wall, and even then it's not just comfortable. That's actually pretty cool to think about, by the way

Interestingly, while cats tend to not need surfaces to lean on, that never stops them leaning. As an owner of cats for years, I've noticed that, especially when they lie down, they prefer to do it in a corner, resting their back on one surface and their head on the perpendicular surface. I had a cat who wouldn't sleep unless he was in his corner of the sofa

Quote:
So the question may not be how a digitigrade would sit, crouch, and so on, because it simply wouldn't do those things. The hard question is how a digitigrade would stand straight - because, generally speaking, it wouldn't. Digitigrades tend to fall into two categories: quadrupeds, and creatures with a horizontal posture, either bending forward their thighs to place their shins and metatarsals underneath their center of mass, or possessing a hefty tail to counterbalance their torsos. Basically, it'd be hard to make a bipedal, erect-posture digitigrade not look somewhat bizarre.
Hence why I wanted to try and avoid real-life examples. As you said, most digitigrades are either quadrupeds or birds, neither of which have a skeletal structure directly applicable to a humanoid skeleton. Even a feline standing up isn't directly applicable as they stand on their entire feet. Canines standing up don't count, either, as they are visibly struggling for balance. I guess the walking dog that was born without front paws could count as an exception, but their skeletons really aren't built to stand upright.

I'm not sure we want to apply digitigrade animal habits directly into a humanoid bipedal digitigrade, as I'm not convinced a direct transition would look good. Quadrupeds, especially, are designed with a pelvis hat lets their legs ride up the side of their hips, with knees going up as far as their rib cages sometimes. Doing this on a humanoid frame would be really freaky, and not in the good sense, since human legs are designed to stand side by side next to each other, not on both sides of the body. Twisting them up into this frame would REALLY stretch the butt in a way that makes me hurt just thinking about it.

Actually, the primary reason I listed the HOMM5 Succubus is to demonstrate that humanoid bipedal digitigrades CAN look good and natural, even if they violate the laws of nature. I actually think her stance is about perfect on that pic, but I'd need to reexamine the actual in-game model more closely. Stupid game not letting me rotate critters on their info screen!

*edit*
It occurs to me that we're missing an important character of this type - Mass Effect's Tali-Zora, and MY GOD it's hard to find a decent picture of her online that isn't porn. Guess I wasn't the only one who got that impression, but I'm starting to think I should turn my Google search content protection back on...

Anyway, Tali is a good example of a character with a non-standard leg composition who still looks works and acts humanoid. I'm not sure if she counts as digitigrade, however, as she looks more like she just has really bowed-out bones on a regular humanoid leg. It looks like possibly her femurs are bowed forward while her tibia is bowed backwards. It gives the LOOK of digitigrade legs while not actually BEING digitigrade legs, at least I don't think they are. One could actually try and pretend there's a small, short bone between a straight thigh and a straight foot, a very minor shin, as it were, but I'm actually starting to doubt that's realistic.

That said, I also found a picture I could only describe as What? doing the same search which actually does look like it has either straight-up digitigrade legs, or at the very least very screwed-up knee joints. I'm not sure this is a good way to draw out the legs, myself, but I wanted to put it out there.


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

By the way, I'm finally home I managed to grab a few good shots of the Heroes of Might and Magic 5 Succubus as she appears in-game from both the character info screen and what I could catch from combat. And, yes, that is a glowing fiery triangle only present on the Succubus Mistress upgrade (actually pictured here). I'm not making this stuff up, people. This is a famous, mainstream game!

Anyway, I wanted to get better screenshots, but the camera controls suck too hard, so I have everything that I could offer. The big pic is from the creature info screen and the two little pictures are from combat. Note that, while she's standing up, the Succubus' legs are almost straight, with just a slight bend in the knees, but with the feet and hooves held almost vertically down. She has a kind of "seductress" pose and she wobbles suggestively, which is why her hands are constantly on her thighs and her hips are sort of always pushed to the side and her hip is constantly slanted.

However, every time you mouse-over the Succubus, she sort of crouches down and does this "Come and get me!" sort of alternating two-hand beckon, which is what the second smaller pic is from. Personally, I think that kind of leg stance is about as close to what I had in mind as I'm likely to get from other people's work. Either that or her legs in her idle stance. That's sort of what I'd like to see for digitigrade legs.

What do you think?


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

Well, "famous mainstream" game is debatable, I'm afraid. I'm pretty sure HoM&M:III was a bigger, more mainstream hit. They're both certainly good games, but HoM&M:V was marred horribly by its coding. Suppose that's to be expected when you have a group of coders who are fluent in one language, writing code in a language which uses syntax from a different language.

As for the presumed raciness of the content in that game, you have to keep in mind which country the game was created, and the cultural norms there. I personally find it quite refreshing, and the flak it garnered was undue. In fact, I wish more studios (especially mainstream ones) would consider subdued content like that. The female heroes for the Haven race, for instance, were quite conservative in their garm, as would be appropriate.

Aaaaaaaand, as for the actual topic of discussion here: Better digitigrade legs would be nice, though personally not a huge priority for my gaming experience. I'd sooner like a little more variety in the mundane "costume" pieces prior to fixing up one of the most exotic pieces.


 

Posted

Quote:
Originally Posted by Samuel_Tow View Post
Heh, yeah, I didn't even think of that one But yes, a digitigrade is a creature which has evolved feet designed to walk on their toes, not simply any animal who just chooses to walk on its toes some of the time. Human feet are designed to stand on an equilibrium between the toes and the heel, and high heels are no different. You CAN walk in them just on your toes, but that requires some serious calves. Most of the time, people walking in high heels are indeed walking on the actual high heel
I almost always walk on my toes when I'm in bare feet, and I'm almost always in bare feet when I'm at home.

Quote:
Originally Posted by Samuel_Tow View Post
And, yes, that is a glowing fiery triangle only present on the Succubus Mistress upgrade (actually pictured here). I'm not making this stuff up, people. This is a famous, mainstream game!
Is this the demonic equivalent of "Your vagina is haunted!"?


http://www.fimfiction.net/story/36641/My-Little-Exalt

 

Posted

Quote:
Originally Posted by VileTerror View Post
Well, "famous mainstream" game is debatable, I'm afraid. I'm pretty sure HoM&M:III was a bigger, more mainstream hit. They're both certainly good games, but HoM&M:V was marred horribly by its coding. Suppose that's to be expected when you have a group of coders who are fluent in one language, writing code in a language which uses syntax from a different language.
Well, obviously it's not a smash hit, not after Heroes of Might and Magic 4 basically murdered the series, but it's still a mainstream game that's famous enough to get its own review on GameTrailers. For reference, ours DOESN'T! More to point, I mean to say it's a professionally created created title that's famous enough for a lot of people to at least know it exists. Hell, I'd even count Advent Rising as a mainstream game, and YOU tell me if you've ever heard about it

I mean, it's not Requiem Hurts, after all.

Quote:
Aaaaaaaand, as for the actual topic of discussion here: Better digitigrade legs would be nice, though personally not a huge priority for my gaming experience. I'd sooner like a little more variety in the mundane "costume" pieces prior to fixing up one of the most exotic pieces.
See, this kind of goes back to a dangerous argument - The piece is exotic, i.e. not many people use it, so it receives few additions and doesn't really look very good, causing fewer people to use it. I, myself, have only two characters with digitigrade legs (out of thirty, last I counted), but considering there are a grand total of three lower leg options... OK, make that three out of 30, I have a robot... Considering there are three lower leg options for Monster legs, is that really much of a surprise? It's like asking why I don't have more trenchcoat-wearing characters when the entire game has ONE TRENCHCOAT!

I don't necessarily disagree with your assessment, of course. The Monster legs are indeed a more exotic option. However, consider that virtually nothing has been done about them since I7 when Robotoc Claws were added, bringing the total leg types to a staggering three. We don't even get the option to use Tights with Skin designs on these legs, forcing me to use my most hated type of garment - the below-the-knee pants - on my digitigrade alien girl, whereas I'd have gone with something more interesting like the Hacker bottom or even one of the two versions of shorts.

I fully appreciate leaving the stranger stuff alone while you focus on the more mundane stuff, but they've been leaving the stranger stuff alone for four years now. CoV gave us five monster heads and one extra soon thereafter. Five years later, we have 5 + 1 monster heads. CoV gave us Monster heads, and it took a while before all models and all torsos could actually use them. CoV gave us Monster Legs, and we have the exact same options we got then, +1. I honestly feel it's high time SOMETHING happened to them, and with hints being laid down for the Mutant booster pack, I have some hope for "more stuff" in this regard.

But that still won't fix the awkward legs. At this point, I'm not sure the awkward legs are even fixable.

Interestingly, a casual Google image search today threw up something it didn't give me before, at least not that I was able to see. It's a simple drawing of digitigrade vs. plantigrade legs that neatly encapsulates what I've been talking about. I could spend all day talking about how I'd envision digitigrade legs looking, but instead I'll point to this image and say "pretty much like that."

Not that that'll stop me from explaining it, of course Look at that pic and notice what I've been talking about - the thigh is bent forward, bringing the knee ahead of the centre of mass. The shin is strongly backwards, putting the heel as far behind the centre of mass as the knee is in front of it. Then the foot is angled forward, placing the toes (rather, the ball of the foot) right under the centre of gravity. It's exactly what I wanted to draw and failed.


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

1. The current monstrous lower legs could be easily improved with little effort by modifying the models to make them more vertical. The ankle joint is not used anyway so it would be a simple model change that would make them look significantly more stable.

2. Do not go overboard with the "Z" shape. The more the bones are off from vertical the more stress is placed on joints and tendons. At some point it just looks painful.

3. If previous boosters are anything to go by, at this point the art assets for Mutant booster are pretty much done and locked down. But there will still be a bug or two that will need fixing

4. I am not sure that many animations would need to be changed. The game has, what, 2 run cycles per model type (normal and ninja) ? And most of martial arts, but thats just one powerset.

5. If new, more monstrous (or tech-y) legs are added the movement should (would?) be similar to the russian rocket boots

6. Heh @ the tags.


I do not suffer from altitis, I enjoy every character of it.

 

Posted

sam, perhaps for am more mainstream game, remember morrowind? that sold very well and the argoninans and kajiit were both digitigrade. that was how i first heard the term, when oblivion changed them, much to many fan's dismay. the bad news of course is that part of why they were changed is because in morrowind they did cause a lot of extra animations needed, and they generally looked worse.


 

Posted

Quote:
Originally Posted by rian_frostdrake View Post
sam, perhaps for am more mainstream game, remember morrowind? that sold very well and the argoninans and kajiit were both digitigrade. that was how i first heard the term, when oblivion changed them, much to many fan's dismay. the bad news of course is that part of why they were changed is because in morrowind they did cause a lot of extra animations needed, and they generally looked worse.
Unfortunately, I never played any of the Elder Scrolls games. Not really my kind of thing. Looking for the Kajiit, though... Ugh. OK, I give up. I can't find a decent full-size pic of one of them that isn't erotic or a nude patch. Would you happen to have a decent screenshot handy so we're on the same page? Because this actually does sound like an interesting concept.

And, yeah. I guess animations will always be the bottom line, especially if you have to reuse plantigrade animations with them. In fact, a random image a random search threw up reminded me that Champions Online uses the same legs we have here, right along with everything that's wrong with them here. In fact, theirs look even stranger, because while ours are limited to hooves and claws and other legs that end on a toe, theirs actually end on a real shoe-shaped foot, which actually makes them look like someone steel-piped you in the shins. As far as animations goes, there is indeed something to worry about.

In fact, this brings to mind the earlier conversation about how a digitigrade would crouch, and the acceptance that it would not be like THAT. What we see on the last pic is basically human crouching with the heels touching the butt, since it actually misses the point of having digitigrade legs. At the same time, though, it doesn't actually look half bad. I don't recall crouching with digitigrade legs looking anything short of repulsive here.


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

Quote:
Originally Posted by Charcoal_EU View Post
1. The current monstrous lower legs could be easily improved with little effort by modifying the models to make them more vertical. The ankle joint is not used anyway so it would be a simple model change that would make them look significantly more stable.
Modifying in what way, though? That's the big question. Do you feel there's anyway to modify the legs such that they could still use the old knee joint without throwing animations too far off? I guess shifting the knee slightly forward and the bulk of the thigh forward from that might work, but I'm not sure it's change enough to be worth the tweak.

Other than that, yeah, I'd be up for a complete redesign, but I fear the animations problem.

Quote:
2. Do not go overboard with the "Z" shape. The more the bones are off from vertical the more stress is placed on joints and tendons. At some point it just looks painful.
While I agree that legs more reminiscent of an accordion bag is not a good design , I still feel we should be averse of making digitigrade legs TOO straight. Human tend to stand upright with their knees locked and their legs straight. At that point, how many joints they have along the straight leg is unimportant, because it's a straight leg. I still feel that whatever we do, we need to retain enough of the bent leg look to still feel like digitigrade legs. The succubus I posted is probably as extended as I'd like to see these legs. More than that just misses the point.

Quote:
3. If previous boosters are anything to go by, at this point the art assets for Mutant booster are pretty much done and locked down. But there will still be a bug or two that will need fixing
To be honest, JUST more Monster leg options (and better Monster heads, come on now!) would be enough for me without a specific redesign of the rig. At the same time, I have my doubts that this is what the pack will actually contain. When it comes to non-human parts, I've almost always been disappointed with the options we get, and I don't know why that is. Either my thinking is far too weird or the art team are far too conventional and not inspired to make weirder inhuman items, but the Monster bits we have feel like an artist was given the task, responded with "Dude! I do NOT want to draw that!" and was forced to anyway.

It would be easy to convince me otherwise with an awesome Mutant pack, but something tells me they'll all be human-frame parts almost entirely. Tails, monster legs, monster heads, monster head details, monster gloves and so forth... I'll get excited when, or rather IF, I see them expanded.

Quote:
4. I am not sure that many animations would need to be changed. The game has, what, 2 run cycles per model type (normal and ninja) ? And most of martial arts, but thats just one powerset.
The problem is that, as far as I can tell, each animation involves the WHOLE body, so if you put in a new set of legs that doesn't work with the old rig, you'll basically end up having to retweak EVERY animation out there. That's a bigger endeavour than even Shield Defence, if it's at all technically possible. I'm sure BABs would fly across the Atlantic just to punch me if he were made to go through every animation AGAIN

That said, this is definitely something I'd like to see. Let's put it like this - if there really IS a CoH 2, this is one of the things I want to see integrated into it from the get-go. I'm still unconvinced that a technical solution can't exist to bridge the gap here, though.

Quote:
5. If new, more monstrous (or tech-y) legs are added the movement should (would?) be similar to the russian rocket boots
I'm not sure I can agree with that. The man in that video is still walking around as normal, he's just hopping on every step. To avoid flipping himself over, he's adopting a kind of wider, swinging walk, which is something I can mimic with my own natural legs. Again, I'm not a fan of modelling animations after humans wearing digitigrade stilts, but I'd be interesting to see one of those running full-tilt. That has got to be awkward with our own biology's adaptation and limitations.

I'd also be averse of copying other games, especially giant robot shooters like Mechwarrior, or indeed the various walking robots in existence, as those tend to walk on a constant balance model, which living creatures don't actually do. Humans, especially, lack static balance almost the whole time when walking or running. What you'd see in a game is a digitigrade character placing a leg forward and THEN shifting weight to it, essentially balancing on one leg. What I envision an actual humanoid bipedal digitigrade doing is shifting weight forward and actually impacting on the feet... Which is actually not something I'm convinced backward-bent legs are good at.

Quote:
6. Heh @ the tags.
And I didn't write either! There's always one like that whenever I bring up a non-human subject, it seems. Comes with the territory, I guess.


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

ok sam, here is a shot i got from a modding site, the textures were updated versus vanilla, but it gives youa shot of how they handled it, and the meshes are the same.

also, yeah, i was playing co yesterday, trying to do a new "full dragon" costume for rian, and yeah, the digi legs look spindly in beast stance, and since he is a martial arts framework character...they look really oddly with the kicks, they kind of do here as well. the ideal would be something that looks like a raptor, but a raptor would run really oddly in context of how our heroes move, im not sure they could move sideways very well, given how the hips seem oriented toward very fast forward motion than sideways.

hey wow-ers, how do draeni do it? they seem to have longer shins.


 

Posted

Quote:
Originally Posted by rian_frostdrake View Post
ok sam, here is a shot i got from a modding site, the textures were updated versus vanilla, but it gives youa shot of how they handled it, and the meshes are the same.
That's, um... Oh, my... I don't think I like how these look

I think my biggest problem with the lizardmen in that pic is that they don't really look like they're digitigrades. They look like people standing on their toes and slouching. See, this is exactly the look I want to AVOID by trying to sell the foot part as much longer and much more vertical. Ideally, a decent digitigrade shouldn't look like or even remind of a person standing on their toes. If people can manage to confuse a cat's heel joint with its knee joint (and I know people who do), then that's what we should be looking at.

Looking at that picture, I'm reminded of something else that bugs me with the digitigrade feet we currently have. Let's have another look at this pic and notice the very bottom of the foot and what its joint does. The "toe" joing in these legs is actually curved down and back, which actually reminds me of the aliens from Independence Day, who actually walk on the knuckles of their feet... It's almost as if we are walking on what is LITERALLY the very tips of our toes, rather than the balls of our feet, because the ball of the foot is where the foot bones end and the toe bones begin.

Looking at the legs of practically every digitigrade I'm aware of, they all have one thing in common - in normal, standing-up stance, the toes are bent UP from the foot as though the animal were walking on its tip-toes (which it actually is). In pawed animals, the paws don't actually curve back even during walking, and at most the toes sit in-line with the feet at rest. Hoofed animals and walking birds do occasionally bend their toes "the other way," but they almost exclusively do that when sitting down, in a position where the heel touches the ground. They don't do that standing up or walking around.

So for our digitigrade legs to end in such an awkward angle for the toes, especially on hoofed feet, is just... Really, really odd. It feels like something that didn't have to be that way, actually.

Quote:
also, yeah, i was playing co yesterday, trying to do a new "full dragon" costume for rian, and yeah, the digi legs look spindly in beast stance, and since he is a martial arts framework character...they look really oddly with the kicks, they kind of do here as well. the ideal would be something that looks like a raptor, but a raptor would run really oddly in context of how our heroes move, im not sure they could move sideways very well, given how the hips seem oriented toward very fast forward motion than sideways.
Raptors and, to the best of my knowledge, a bunch of other dinosaurs, actually had the body shape of birds with their spines horizontal to the ground and perpendicular to the legs. As such, their stances and movement habits really don't apply to an upright humanoid. As well, Raptors have a part of their abdomen positioned below their hips and "between" their legs. Family jewels notwithstanding, the human abdomen is generally above the pelvis and the legs are structured to be next to each other with no room for vital organs. This very different pelvis simply makes their motions incompatible.

That said, I will join you in exclaiming how bad the Champions digitigrade legs are This was one of the huge shocks I got when I started that game, positive as I was that it would be better than this one in all respects. Instead, the broken-shin legs over there made my eyes turn around in my head. Ugh... OK, I will admit - ours are kind of embarrassingly cheap and really not very well animated... But the ones over there are just BAD. Seriously - you do NOT put a human foot on an inhuman leg. Even I find that too weird.

Quote:
hey wow-ers, how do draeni do it? they seem to have longer shins.
Well, my WoW-playing friend deactivated his account after getting robbed twice, and he never played a Drenai while he was there, but a Google search reveals that they don't look too bad. However, you have to remember that in WoW, each race and each gender is basically its own model with its own animations. They get away with having relatively fewer things to animate, with every other magic being the same hand wave and so many of the attacks sharing animations with different effects. I'm not really sure that's applicable here, considering how everything has to fit practically every situation for it to count.


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

Quote:
Originally Posted by Samuel_Tow View Post
Modifying in what way, though? That's the big question. Do you feel there's anyway to modify the legs such that they could still use the old knee joint without throwing animations too far off? I guess shifting the knee slightly forward and the bulk of the thigh forward from that might work, but I'm not sure it's change enough to be worth the tweak.

Other than that, yeah, I'd be up for a complete redesign, but I fear the animations problem.
Perhaps I was not clear enough. I meant changing the lower leg part to be closer to vertical both above and below the ankle joint. The reason it looks unstable now is because the ankle goes too far to the back with nothing to balance it. And since the knee is locked in place due to animations the one remaining option is to push the ankle forward.

Quote:
Originally Posted by Samuel_Tow View Post
The problem is that, as far as I can tell, each animation involves the WHOLE body, so if you put in a new set of legs that doesn't work with the old rig, you'll basically end up having to retweak EVERY animation out there. That's a bigger endeavour than even Shield Defence, if it's at all technically possible. I'm sure BABs would fly across the Atlantic just to punch me if he were made to go through every animation AGAIN

That said, this is definitely something I'd like to see. Let's put it like this - if there really IS a CoH 2, this is one of the things I want to see integrated into it from the get-go. I'm still unconvinced that a technical solution can't exist to bridge the gap here, though.
I am not so sure about the whole body thing. For example, wings seem to animate independently of (most? all?) powers going off. And I find it hard to believe that every animation in the game has "with wings" and "without wings" versions. Especially since wings were released at a time when development resources were a lot more constrained than now. So maybe it is possible for most upper-body animations to simply assume that the legs are there and use a single, shared fight-stance animation for the lower body. Would probably still need a tech change but it might be worth looking into.


Quote:
Originally Posted by Samuel_Tow View Post
I'm not sure I can agree with that. The man in that video is still walking around as normal, he's just hopping on every step. To avoid flipping himself over, he's adopting a kind of wider, swinging walk, which is something I can mimic with my own natural legs. Again, I'm not a fan of modelling animations after humans wearing digitigrade stilts, but I'd be interesting to see one of those running full-tilt. That has got to be awkward with our own biology's adaptation and limitations.

I'd also be averse of copying other games, especially giant robot shooters like Mechwarrior, or indeed the various walking robots in existence, as those tend to walk on a constant balance model, which living creatures don't actually do. Humans, especially, lack static balance almost the whole time when walking or running. What you'd see in a game is a digitigrade character placing a leg forward and THEN shifting weight to it, essentially balancing on one leg. What I envision an actual humanoid bipedal digitigrade doing is shifting weight forward and actually impacting on the feet... Which is actually not something I'm convinced backward-bent legs are good at.
It is the hopping that I had in mind. With no locking joints the muscles and tendons need to be stronger just to stand. And once these dynamic supports are in place, may as well use them while moving too. Hopping, elastic movement can be pretty efficient if one has the body structure to support it - works quite well for kangaroos for example.


I do not suffer from altitis, I enjoy every character of it.

 

Posted

Quote:
Originally Posted by Samuel_Tow View Post
But then it occurred to me that I don't really know what the corresponding positions to real-leg stances would be, which brings us back to the question: What do you believe a bipedal humanoid digitigrade would look like?
I think the example to follow would be bird legs.

When digitigrade mammals get on their hind legs, it's not a very stable position for them. They can usually only stay in this stance for short distances and they seem only able to hop a bit. Digitigrade legs don't really seem very efficient or flexible enough for proper bipedal motion.

If I look at mammal digitigrade legs, the motion vectors seem all wrong for bipedalism. Such legs are great for quadrapedal motion allowing for huge body stretching leaps for great running speed. But on humanoid body, the balance seems all wrong unless counterbalanced somehow by bending the spine and hips over and extending a tail like birds, or better still, kangaroos

But this is a science/fantasy game, who says you have to be constrained by real biological anatomy requirements?

Actually just ignore this post. You folks have already covered everything I could possibly say about this in great detail. Never mind.


"Civilization advances by extending the number of important operations which we can perform without thinking of them."

 

Posted

Quote:
Originally Posted by Charcoal_EU View Post
Perhaps I was not clear enough. I meant changing the lower leg part to be closer to vertical both above and below the ankle joint. The reason it looks unstable now is because the ankle goes too far to the back with nothing to balance it. And since the knee is locked in place due to animations the one remaining option is to push the ankle forward.
Hmm... That might work, but remember - pushing the ankle forward also has the side effect of shortening both the shin and the foot. Given how oversized the thigh is compared to them even how, this has the potential to look like an angled shin bone rather than a digitigrade leg, and that would be... Really bad.

Honestly, when I account for the technical restrictions, I'd say they did about as much as the system allowed them to. It's still BAD, but I can see why making Monster legs better would be a serious investment. I'd still want to see it, but I understand the problems.

Quote:
I am not so sure about the whole body thing. For example, wings seem to animate independently of (most? all?) powers going off. And I find it hard to believe that every animation in the game has "with wings" and "without wings" versions. Especially since wings were released at a time when development resources were a lot more constrained than now. So maybe it is possible for most upper-body animations to simply assume that the legs are there and use a single, shared fight-stance animation for the lower body. Would probably still need a tech change but it might be worth looking into.
Wings are are an entirely separate costume detail that just happens to have a semi-autonomous animation. Ming motion actually takes its cues from the same sequences that the character does, but does not actually take them from the character, itself. I have very serious doubt that the character itself is nearly as modular.

Additionally, wings aren't accounted for in power animations, but that's largely because wings are not used in power animations. There is no animation that assumes you have wings, like some kind of wind-driven wing gust attack the classic anime wing slice. Legs, however, very much ARE required for practically every animation done on the ground, even if they have no reason beyond propping up the torso. You could have a case for legs being extraneous in the air, but that's the one time when digitigrade legs don't actually matter even visually

Every animation used on the ground has its own stance. They all start from the same one, but the leg motion in each power is specific to that power. As such, you can't just replace the legs with a different skeleton without actually animating those different legs for all the old stances and animations.

You MIGHT be able to get away with doing a kind of mathematical transformation from the regular leg to a funky digitigrade leg, and I actually think I know enough about digitigrade legs to at least try and think about this now. We'll see.

Quote:
It is the hopping that I had in mind. With no locking joints the muscles and tendons need to be stronger just to stand. And once these dynamic supports are in place, may as well use them while moving too. Hopping, elastic movement can be pretty efficient if one has the body structure to support it - works quite well for kangaroos for example.
That's more than a little ambitious for an in-game animation, though, as it's a brand new run. On the other hand, so is Ninja Run, so it's actually not as ambitious as swapping legs That said, I'm not sure I'm a fan of it. It actually looks awkward and silly to me, somewhat reminiscent of a guy on a pogo stick, which is basically what it is. Animals that move in hops tend to do so by hopping with both legs, and those that walk and run with alternating steps don't tend to hop much.

Hell, even astronauts on the moon found out it was easier to move around in two-legged hops

You do have a point, though, in that a humanoid bipedal digitigrade would probably bot walk quite the same as a human would.


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

Quote:
Originally Posted by rian_frostdrake View Post
hey wow-ers, how do draeni do it? they seem to have longer shins.
Honestly, it's difficult to pay attention to sciency stuff like this with all the hip-waggling going on there.

Maybe I'll see if I can get my screenvideocapturing crap working later and get some good detail on their walking/running animations.
Until I do that: In the interest of seeing how everyone does digitigrade-legged animations, the City Vau-er, WoW Armory has a small selection of animations in the character-viewing pane there. And don't make fun of the "I am levelling up" fashion failure

As far as real-life digitigrades being all standing on two legs goes.. Standing Cat is impressive and standing, but probably could not do much in the way of jogging or crimefighting. What do you mean I'm just looking for excuses to throw links around?!


 

Posted

Quote:
Originally Posted by Chompie View Post
Honestly, it's difficult to pay attention to sciency stuff like this with all the hip-waggling going on there.
Remember, it took some serious science to get that hip-waggling just right. Just like it takes embarrassingly complex science to make the breasts of Japanese fighting game female fighters bounces

Quote:
Maybe I'll see if I can get my screenvideocapturing crap working later and get some good detail on their walking/running animations.
Until I do that: In the interest of seeing how everyone does digitigrade-legged animations, the City Vau-er, WoW Armory has a small selection of animations in the character-viewing pane there. And don't make fun of the "I am levelling up" fashion failure
Hmm... That was actually quite informative. I don't if Blizzard tried reusing animations from the other races or just started from scratch (they have money to burn, so it could be either), but those are pretty close to what I'd like to see. In fact, it showed me a few things of note:

1. Drenai feet are almost always very close to vertical no matter what they do, meaning most of their locomotion comes from their knees and their ankles, but relatively little from their toes. Speaking honestly, that actually looks quite natural.

2. They crouch the other way. When humans crouch, they bring the toes back and curve them up, sitting on their heels. When Drenai crouch, they bring their toes forward and fold them back. They can't actually sit on their heels since their shins are much shorter than their thighs, which looks surprisingly natural.

3. Both the knees AND the heels are animated. No doubt that's what gives Drenai their much more natural feel above our own static legs. Even simple animations like bobbing up and down doing the "chicken" emote use both knees and heels.

4. When standing up straight, their legs are almost straight, in turn, much like the Succubus I listed. The knees are still forward of centre and the heels are still back of centre, but by very little. Remembering what my Wing Tsun teacher said, holding a joint at around 120 degrees or so is actually pretty string for the muscles, so that's why it might look more natural than a more crouched pose.

5. WoW has better emotes I'm not sure they have MORE emotes, but between fingers and what must have been a higher monetary cost... They just look really good.

Quote:
As far as real-life digitigrades being all standing on two legs goes.. Standing Cat is impressive and standing, but probably could not do much in the way of jogging or crimefighting. What do you mean I'm just looking for excuses to throw links around?!
Personally, I prefer this version of Standing Cat


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

Quote:
Originally Posted by Samuel_Tow View Post
This is kind of where this whole thread came about. Since it's foolish to suggest adding a brand new lower legs skeleton to be animated completely over again, I got to thinking if it isn't possible to come up with a mathematical model for transforming regular-leg positions into digitigrade leg positions completely automatically. But then it occurred to me that I don't really know what the corresponding positions to real-leg stances would be, which brings us back to the question: What do you believe a bipedal humanoid digitigrade would look like?
Unfortunately, a digitigrade gait, particularly when running and leaping, is sufficiently different from a plantigrade gait that it would take completely redoing the lower-body skeleton and the associated animation. If you watch the Muybridge race horse photographs collected into an animation:



you can see the amount of flexion at both the tarsal and phalangeal joints, where the current animations just have a fixed ankle joint, keeping all the plantigrade leg animations but replacing the calf with the bent calf+tarsal segments of the digitigrade leg. Unfortunately, the number of large bipedal digitigrades is limited, and are all flightless birds, who have a greatly shortened femur compared to the classical satyr/faun proportions, so their gait is not as readily usable as a reference, and I don't think the devs are interested in putting in the amount of work that was done for this Pan animation just to make a small set of leg options look right.


"But in our enthusiasm, we could not resist a radical overhaul of the system, in which all of its major weaknesses have been exposed, analyzed, and replaced with new weaknesses."
-- Bruce Leverett, Register Allocation in Optimizing Compilers

 

Posted

Quote:
Originally Posted by srmalloy View Post
Unfortunately, a digitigrade gait, particularly when running and leaping, is sufficiently different from a plantigrade gait that it would take completely redoing the lower-body skeleton and the associated animation. If you watch the Muybridge race horse photographs collected into an animation, you can see the amount of flexion at both the tarsal and phalangeal joints, where the current animations just have a fixed ankle joint, keeping all the plantigrade leg animations but replacing the calf with the bent calf+tarsal segments of the digitigrade leg. Unfortunately, the number of large bipedal digitigrades is limited, and are all flightless birds, who have a greatly shortened femur compared to the classical satyr/faun proportions, so their gait is not as readily usable as a reference, and I don't think the devs are interested in putting in the amount of work that was done for this Pan animation just to make a small set of leg options look right.
*edited to snip the animation*

You know... On the one hand, I agree with you completely, but on the other hand, I keep wondering if there isn't a systemic solution to this. Practically, there most probably isn't, but at least theoretically, there might actually be. consider the following rig for one leg:

You have a hip joint with three axes of rotation.
You have a knee joint with one axis of rotation.
You have an ankle joint with, err... Is that one or two axes of rotation?
You have "a toe joint" which I'd actually give a full two axes of rotation.

Now, swapping in a digitigrade leg, you can actually re-map the joints as follows:

Digitigrade hip maps to plantigrade hip.
Digitigrade knee maps to nothing at all, but is calculated.
Digitigrade ankle maps to nothing at all, but is calculated.
Digitigrade toes map to plantigrade ankle.
Nothing maps to plantigrade knee, but digitigrade knee and ankle are calculated from it.
Nothing maps to plantigrade toes. Joint is not used.

This actually gives us a fairly straightforward mathematical problem, defined as such:

Define a transformation from plantigrade hip, knee and ankle orientation to digitigrade hip, knee, ankle and toe orientation, ensuring that the angle between digitigrade thigh and shin is equal to the angle of digitigrade shin and foot at all times. Map thigh and foot orientation directly.

This should actually be geometrically solvable, stated like this, at least if give the give the plantigrade thigh joint a normal vector (basically, a direction which tells is which way the broad side of the bone is facing). With this, it's fairly easy to define a plane foe the digitigrade leg to be drawn in, and with defined thigh and toe joint for it and the rule of always keeping the angles on the leg equal, it's mechanically simple to recalculate a new leg. You can then calculate the angle of the digitigrade toes off the foot by taking the angle of the plantigrade angle off its shin and adapting the calculations.

The best part about this is that we actually eliminate all the subjective variables. By defining a plane to draw the digitigrade leg into and fixing its "start and end" joints, there really is relatively little judgement to be had into the calculation, just a lot of analytical geometry and possibly vector mathematics. It also transforms what is ostensibly a 3D problem into a 2D solution with an added definition of the 2D space (plane). The reason this is possible is that the knee only has one axis of rotation, so thigh and shin will ALWAYS be in the same plane.

I can actually formulate this mathematically if anyone REALLY cares, as well as probably type out the exact equations. Somehow, I highly doubt people would be lining up to read that, though

Now, of course, whether this would actually LOOK GOOD is a legitimate question, but I suspect it would look about as passable as a fixed ankle joint on a digitigrade foot.

*edit*
Yano... When I got down to do some math on this, I ran into a little problem, and it's not the math. It's more a problem of biology. As a human, I can crouch down by bacisally sitting on my heels with my toes extended as far up as they will go, and I'm pretty much down to the ground because my feet are so short. A digitigrade can't do this. If they sit on their heels, as it were, they are quite a ways off the ground, pretty much the height of their feet, which on a digitigrade is significant. For such an animal to actually get low to the ground, it must do what Standing Cat did and basically put the whole foot flat on the ground. However, when you do this, when you have your creature do this, its toes are no longer under its butt, they're far, FAR forward, kind of like a human squatting, only much farther forward because of the much longer foot.

The reason this is a problem is that I'm trying to map plantigrade ankle to digitigrade toes, but we can sit on our ankles, whereas digitigrades physically cannot sit on their toes. I played around with the IK tool in Flash, and when I put in the proper joint restrictions, this is what I found out. Digitigrade legs are physically incapable of doing some of the things plantigrade legs are, at least if we want to remap animations joint-for-joint. It's mathematically calculable if you disregard joint rotation limits, but with them... Not so much.

Now, there IS a way to get the same poses to work for both plantigrade and digitigrade legs, but this involves some truly bizarre biology, and in a thread that includes things like succubuses, drenai and lizardmen, that's saying something. Basically, if the creature's shin bone were TWICE the length of its thigh and leg bones, then simple geometry would put the toes right under the butt with the legs folded in, but... Does anyone actually think such a travesty would look good? Because I have my doubts.

*edit*
Well... You be the judge. I can't say it looks as bad as I thought it would, but at the same time, I can't honestly say it looks GOOD. And if THAT'S what it takes to have interchangeable animations, then no, I don't think a direct copy is PHYSICALLY possible. Forget technical issues. The bone structure of such a character would simply not allow it.

Incidentally, would anyone have a good answer to the following two questions:

1: Are human thighs and shins the same length?
2: For a given thigh and shin length (call it 50-50), what length digitigrade legs would you use?

The first one is just me being curious, but the second one is THE showstopper at the moment. I tried equal length bones at two-thirds the length (so around 30-30-30 thigh-shin-foot), producing the problem in the first edit. But maybe it's my lack of imagination causing it, so I'd like to know what people imagine. For reference, I'd eyeball the length of limbs in City of Heroes to around 5:3:3 thigh:shin:foot, which looks really awkward to me.


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

OK, in case anyone cares, here's a bit of what turned out to be high school math. I'm not going to bother giving proofs and explanations, instead choosing to just dump the results. I'm going off the following picture:



And let me spare a few words to describe it. Line colours are as follows:

Black - original plantigrade leg, bent at the knee. The thigh and shin sections are taken to be of equal length marked as m, and we know its offset off the vertical (blue) marked as d and the angle of the knee marked as a.

Red - the target digitigrade leg, bent at the knee and at the ankle. The thigh, shin and foot sections are taken to be of equal length marked as n, for which we are looking to discover the offset off the vertical (blue) marked as c and the angle at the knee marked as b. By definition, the angle at the heel is taken to be equal to that of the angle at the knee, so we don't need to calculate it separately.

Blue - the "vertical," forming the straight line between hip joint and heel joint. It's there for calculation purposes only, and its length is given as x, which doesn't actually come to play in anything I'll list here, but it's on the picture so I'm mentioning it.

What we know: the lengths of m and n and the size of the angles a and d/
What we need: the size of the angles c and b. As such, the sizes of those angles are given as:



Now, those aren't easy formulas to calculate by hand, especially considering there's some pretty serious trigonometry in there that isn't even solvable by hand (unless you have tables of pre-calculated trigonometric functions on hand, which doesn't really count as "by hand"), but I have my doubts its actually too costly for machine calculation, even in real time. I dare say the cost of a simple waving piece of cloth such as a cape is vastly higher, with numerous vertex calculations (which are undoubtedly singularly simple, but much more numerous) as opposed to basically just two.

I'm sure a smarter man than me can come up with a better, easier calculation that doesn't go through as much trigonometry, but that's all I have at the moment.

Mind you, that's actually only one step of three, but I dare say it's the most complicated one. It's also the one that's likely to cost the most collisions of skeletal limitations, so it's the one I decided to do. I might be convinced to do the others at SOME point, but that's not very likely.

*edit*
OK, I had to laugh at my tag typo. I mistyped [/img] to [/omg]


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

Quote:
Originally Posted by Samuel_Tow View Post
I dare say the cost of a simple waving piece of cloth such as a cape is vastly higher, with numerous vertex calculations (which are undoubtedly singularly simple, but much more numerous) as opposed to basically just two.
Mmm, it's not quite that simple to compare, I think. Cape systems could probably be done with a vertex shader on the graphics card, whereas a plantigrade -> digitigrade transformation would have to be a transform in the animation system on the CPU. While a cape system would have more vertices that it's working with, it might still be faster than the trigonometry.

Quote:
Originally Posted by Samuel_Tow View Post
I'm sure a smarter man than me can come up with a better, easier calculation that doesn't go through as much trigonometry, but that's all I have at the moment.
The second equation can be reduced from 3 calls to sine and 1 call to arccosine (4 trig functions), to a call to sine, a call to cosecant, and a call to arcsecant (3 trig functions):
ArcSec ( (4nx) / (3n² + m² Csc(d)² Sin(a)² )

Edit: Invert the fraction within the ArcSec, and change ArcSec to ArcCos, and it should work fine, too.


http://www.fimfiction.net/story/36641/My-Little-Exalt

 

Posted

Quote:
Originally Posted by Fleeting Whisper View Post
The second equation can be reduced from 3 calls to sine and 1 call to arccosine (4 trig functions), to a call to sine, a call to cosecant, and a call to arcsecant (3 trig functions):
ArcSec ( (4nx) / (3n² + m² Csc(d)² Sin(a)² )

Edit: Invert the fraction within the ArcSec, and change ArcSec to ArcCos, and it should work fine, too.
Hence why I said a smarter man I never got the grasp of secant functions or how they work. There's theory I could read up on, I suppose, but I remember being really "scared" of them back in the day.

Quote:
Mmm, it's not quite that simple to compare, I think. Cape systems could probably be done with a vertex shader on the graphics card, whereas a plantigrade -> digitigrade transformation would have to be a transform in the animation system on the CPU. While a cape system would have more vertices that it's working with, it might still be faster than the trigonometry.
I guess it's a question of computing time. Do those look like they're very slow to calculate? I mean, slow enough to cause performance degradation? Do we even know how often skeletal positions are recalculated? I would stab-in-the-dark guess once per frame, which would put them at 60 times a second, and I guess speed would depend on the proficiency of whatever internal calculator is used.

However, you have to admit that as far as infeasible rudimentary kludges go, this one is pretty creative

*note*
I actually came up with another biology problem when thinking about this. If we map the digitigrade to to the plantigrade heel, then the digitigrade leg would be standing on air when the plantigrade foot is standing on its toes. If we map the digitigrade toes to the plantigrade toes, then the digitigrade foot will always be forward of where it ought to be (and look like the character is leaning back), and it would STILL look odd if the character stood on his heels.

Then again...

This already happens. For instance, try this - play a Monster Legs character and hit rest. This has your left foot standing on its toes with the heel in the air. Your hoof/claw will end up staying up in the air. That's in-game right now, so I intend to use it as precedent. Furthermore, I talked about problems with crouching, but there are problems with crouching right now. Kneeling to rest will have your digitigrade heel clip clean through your thigh, folding your knee a full 180 degrees which no believable knee joint should allow. Again, it doesn't look good, but I'm using it as precedent.


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

Quote:
Originally Posted by Samuel_Tow View Post
Hence why I said a smarter man I never got the grasp of secant functions or how they work. There's theory I could read up on, I suppose, but I remember being really "scared" of them back in the day.
What's to be scared of? Secant = 1 / Cosine. Cosecant = 1 / Sine.

Quote:
Originally Posted by Samuel_Tow View Post
I guess it's a question of computing time. Do those look like they're very slow to calculate? I mean, slow enough to cause performance degradation? Do we even know how often skeletal positions are recalculated? I would stab-in-the-dark guess once per frame, which would put them at 60 times a second, and I guess speed would depend on the proficiency of whatever internal calculator is used.
I don't know. There's not really a way to know without comparing the two using performance monitoring tools.

However, as I was walking to lunch, I realized: This math is really only good for transforming the bones. A proper digitigrade rig would have different animation, too, which has not yet been addressed. I think, like my first assumption, that you'd simply need an animator to go in and do it manually.


http://www.fimfiction.net/story/36641/My-Little-Exalt

 

Posted

Quote:
Originally Posted by Fleeting Whisper View Post
What's to be scared of? Secant = 1 / Cosine. Cosecant = 1 / Sine.
It's the math surrounding them that I'm not familiar with. Trigonometry in general is less about knowing the functions and more about knowing how to combine them, and I just don't.

Quote:
However, as I was walking to lunch, I realized: This math is really only good for transforming the bones. A proper digitigrade rig would have different animation, too, which has not yet been addressed. I think, like my first assumption, that you'd simply need an animator to go in and do it manually.
Well, of course, but David already pretty much stated that this is not likely to happen, which we should probably read as "forget about it." That said, what are our options?

1. Brand new animations? Not gonna' happen.
2. What we have right now? Works, but looks BAAAD!
3. Some tweak of the geometry without altering the rig? Possible, but it's limited by the rig itself.
4. Some kind of faux recalculation of the leg bones? Not ideal, but it's certainly better than what we have.

So, yeah, it's probably not what I'd wish for if I had SantaChrist's ear, but of the options that seem at least rational (which is funny, talking about irrational numbers), this one looks like the one that has the potential to produce the most interesting visuals without actually requiring reanimating EVERYTHING for.

Of course, one has to wonder if the technology to recalculate bones on the fly like this even exists, but if it did, I'd go for it


Quote:
Originally Posted by Arcanaville View Post
Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

 

Posted

Quote:
Originally Posted by Samuel_Tow View Post
Well, of course, but David already pretty much stated that this is not likely to happen, which we should probably read as "forget about it." That said, what are our options?

1. Brand new animations? Not gonna' happen.
2. What we have right now? Works, but looks BAAAD!
3. Some tweak of the geometry without altering the rig? Possible, but it's limited by the rig itself.
4. Some kind of faux recalculation of the leg bones? Not ideal, but it's certainly better than what we have.
The big difference between a plantigrade gait and a digitigrade one is that, for plantigrades, there is relatively little flexion of the ankle during the stride, while for digitigrades, there is a range of about -30° during the forward movement of the leg to +90° at full extension behind. While adding the rotation itself isn't difficult, it changes the length of the stride, throwing off the way various attack animations integrate with body movements.


"But in our enthusiasm, we could not resist a radical overhaul of the system, in which all of its major weaknesses have been exposed, analyzed, and replaced with new weaknesses."
-- Bruce Leverett, Register Allocation in Optimizing Compilers