Odjin Run Animation

Work in Progress of a running animation for Odjin. The wings are not animated yet.

The model is a game model hence I can not use IK-Solvers currently to fix the hopping as the engine doesn’t have IK support ( at least not yet ). I can though put some more key frames in between but this has to wait until the animation by itself is working.

The animal is supposed to run quickly. The animation is tuned down in speed to see the individual parts of the animation better. The animal is of rather small weight and roughly the size of a horse. Hence the animation has to give the feeling of it being quick and agile. The idea is to have a “realistic” animation like the animal would exist here and now.

http://epsylon.rptd.dnsalias.net/videos/preview_run_side.jpg

Critics looked for ( in particular but all are welcome ):

  • Tried to add a bit of “weight” to the head while waving up and down. Is it too much or too little?
  • Same for the tail. Too much weight or too little? More or less waving up and down?
  • Front claws contacting the ground: larger time space between the contacts or is it ok like this?
  • Flying phase: Too long or too short?

In my opinion there is something not fully flowing yet with the front legs and perhaps the tail waving but I can’t wrap my head around what it is or how I can improve it.

Known problems: Feet hopping. Requires IK-Solver which I can not use due to the nature of the project.

‘Weight’ on the head might be a bit much. But it’s rather the sudden movement in the neck that bothers me. I think it should be in action all the time.
The center of the body and wings appear stiff. The movement should run through the whole body. Running is a bit like a series of jumps, so if your drew a line throught the whole body, it would change angle like riding on a wave. Or at least I imagine it would :wink:

I was about to say that, his body stay still. If the legs are running on ik, then it’d be ideal to make the body undulate in a ‘back-legs-pushing-off > front-legs-pushing-off < back-legs etc.’ movement.

I guess you refer to the small time frame where the neck goes from a curve into nearly a straight line, right?

The center of the body and wings appear stiff. The movement should run through the whole body.

The wings are not animated yet so they are currently stiff and move with the “origin” bone. This gives most probably the stiff look there. What goes for the “origin” bone it is more or less at the same altitude but rotates. Maybe with the proper wing movements in the end it might hide this a bit. I would be happy if I do not have to touch the altitude of this “origin” bone.

Running is a bit like a series of jumps, so if your drew a line throught the whole body, it would change angle like riding on a wave. Or at least I imagine it would :wink:

I tried to get this feeling. I found on the net one image series of a cat running and tried to get a grasp on how such a movement cycle works. The image in question is here. I could not find something better so far. I got the impression that the “origin” bone ( somewhere in the middle of the spine ) more or less had the same altitude. This somehow made sense to me as pushing upwards ( “origin” moves up ) instead of mostly forward would slow you down. That said though I’m not a zoologist and might miss out something there.

The body behind the wing seems to be moving properly, but that wing slapping him in the hip is a problem. You might want to remove the wings temporarily while you get the body motion straightened out.

The animal is galloping, the front feet should be more in synch with each other, and the rear ones less so. Ideally, a well trained running animal would have its feet perfectly in synch during a gallop, but since animals, like humans, are left or right handed (footed? sided?) the “ideal” is probably never seen in real life. Still, it looks a bit like the front is trotting while the rear is galloping.

A bigger problem is the weight of the animal. When the front feet come down and touch the ground, they take up the entire weight of the animal, the shoulders should raise and the torso sink some. Then the front legs push off, shoulders pressing down, while the rear legs take the weight, and the hips raise and the torso sinks in the rear. The front feet are off the ground as the rear feet push off, hips pressing down, torso moving up, and carrying the rear feet up off the ground as the rear feet rotate forward for the next pace.

I think what looks odd about the head is the undulating motion of the torso and neck does not carry past the neck joint just under the horns. This part seems to be fixed horizontally even while it is moving up and down and front and back. Maybe if it had a bit of rotation, too.

It also seems to be breathing once each pace. That will probably cause hyperventilation. You might consider having it breath once every three or four paces.

I’m trying to understand this right now. When my dragon comes down on his front legs the origin of the shoulder bone is stopped from moving down ( with a subtle under swaying to avoid quick changes ). So should the shoulders now move up while the hind legs are brought into place ( and the torso sinking a bit )? I’m just a bit confused the shoulders are already moving up while the hind legs push off. Would this mean to distribute the up motion onto those two phases?

I think what looks odd about the head is the undulating motion of the torso and neck does not carry past the neck joint just under the horns. This part seems to be fixed horizontally even while it is moving up and down and front and back. Maybe if it had a bit of rotation, too.

In my game engine the first-person camera is not programmed but follows a special bone in the animation. I had before rotating the camera slightly up and down at the turning points but this causes heavy and erratic movement of the attached camera ( gives you nausea ). I then though a dragon would most probably try to “steady” his head while moving to not get nausea from this bobbing motion. But maybe I overlooked something while doing the head animation there. Maybe a “focus” point would help. I’ll try out once to add back this subtle motion to see if I get it there without nausea-effect.

It also seems to be breathing once each pace. That will probably cause hyperventilation. You might consider having it breath once every three or four paces.

In fact what it does is not breathing. I made there a little “eye-candy” motion that is a small up and down of the mouth to make it not stiff there. Maybe I missed what I wanted there.

I got notified that the picture of the running cat has been messed up ( did I tell you already how much I hate the Post Widget on this forum? One of the worst I’ve seen in my entire life! ). So here now the image:
http://www.tvcd.com/TextTut/CatRun/Images/CatRunningLong.jpg

In the last two cat pictures, where the front paw is just touching the ground and the rear paws are still in the air, you can see the points of the shoulder blades poking well above the curve of the back/spine. Then it looks as if the front legs aren’t so much pushing the body forward as they are pushing the front part of the body up a bit as the back legs touch down under the center of gravity and do the heavy pushing forward as the cat’s back straightens out.

He certainly does alternate paws in front on the landing. Seems odd. Especially since his front paws are so close together during the off the ground stage.

Would it work to have the camera track an empty that isn’t parented to the dragon, but is simply moving along a similar path? I’m sure the dragon would hold his head as steady as he could, but gravity and the running motions would cause some unavoidable bobbing up and down, as well as some slight pitching of his head.

As to the mouth motions, to me it looks like breathing.

can you animate with IK and then bake it? (I’m not sure if this means it is no longer using IK

@Orinoco:
What goes for the shoulder part I have to first experiment a bit. The model is made with a skeleton laid out for a game instead of an animation. Hence the entire shoulder motion is tricky to do with what I have here. Maybe I can do something there, let’s have a try.

What goes for the alternating motion I could try to displace the movement a bit so the contact times are more different. I had first though about making it close to each other like the back but then it gets already close to Spyro ( although this one has a pure 4-leg gallop if I remember correctly ). I hope I got you right on this one there.

The problem with the camera is that I want to do something special in my game which other games did not do yet ( or only tried in a very limited fashion ) and this is to have a true first-person view, hence attaching the camera to the head/eyes of the character you play. So I am not able to separate the camera from the head as otherwise tricks like Rag-Dolling or IK ( where no animation data is used ) would break the first-person impression I’m looking for. This simply requires more work at the animation end but It should be worth the fuzz.

@yogyog:
I never worked with baking animations. How exactly does it work? Does it create key frames in fixed intervals? I would like to avoid having tons of key frames as this could possibly slow down the game engine.

I was thinking more of a galloping horse than a cat, although, not having seen that stop motion photography before, I thought cat gaits would be similar to horses. Perhaps if they don’t shift their center of gravity a bit to the left and right they’d wind up kicking themselves in the shins? In the third image one rear paw is well forward of the corresponding front paw, and in the fourth image the same appears to be the case on the opposite side.

I suspect you are going to have to do a lot of creative faking to get the true first person view you are after. Our brains do a lot of processing to make the external world appear stable despite the motions of our head and eyeballs. You’ll need some way of damping out sudden moves so the camera image doesn’t jerk around and some way of anticipating turns by looking in a direction before turning in that direction. It will be quite an accomplishment if you manage to get it to work well.

Tried now reworking the animation a bit. I had been nearly done when I messed something up so I had to step back and try to fix this again. Hopefully I managed to make it better and not worse. The following I did:

  • Changed the front feet to contact with more time in between
  • Added a slight up/down motion to the root bone
  • Removed the phase in the head animation where the neck has been nearly straight making the head to bump up a bit higher than before
  • Added that the shoulders move upwards while hitting the ground ( without directly lifting the body-arm bone )

Still fighting a bit against a blocky look there. Here are two videos of the animation. One again from the side and the other using the head-mounted camera ( using an infamous cube-run ). To me the part where he hits the ground might be a bit “stiff” but I’m not sure about it.

Side Animation | Eye-Camera Animation

@Orinoco:
All games are build upon faking as much as possible :wink: . The camera code in the engine can take care of this if required but games often require a subtle movement as a static image ( like our brain provides it ) is felt “unnatural” by a player. Movement is assigned to you being alive. Having a completely non-moving view usually indicates that you’re dead or something similar. :eek: