Yeah, ILM based the AT-AT movement from an Elephant's walk, .... which was shown in the Making of Empire (I recorded on Video yonks ago (mid 80's) hosted by Mark Hamill ) ....
The AT-AT does look deceptively simple to do, .. until you do one, ... So I for one appreciate the effort you're putting into this Stonks .... keep going mate, you're doing a great job so far :thumb: ....
That would be SPFX. A tarted up copy is included in the extras disc of the Bluray release. The full original version is in on Youtube if you do a search. I didn't turn up a clip of the elephant walk cycle sequence, but I did find this - http://www.youtube.com/watch?v=ZNJmF3JOeqk
I don't think, given the model's original articulation, that it is in fact capable of turning without dragging something along the ground, due to the fact that there isn't any visible joint that isn't pivoted solely along the yz plane. That's just going from the references I have, which aren't very good.
The walker must be able to pivot the hip joints in xy, given that a walker sidesteps in TESB to shoot a snowspeeder. That said, it's not at all clear how it works, given that drive shaft thing along the underside. Even if the hip connection to the drive shaft could pivot in xy (and it might look pretty odd), the ankles and legs absolutely cannot, so you're still left with some pretty non-organic animations required for the turn; it's not going to be as easy to animate as an elephant. I'm looking forward to seeing how you handle it. It was never a problem in the movies since walkers never turned in the movies.
I had thought that maybe the entire forward and rear drive assembles(drive motors, hips,legs,and feet) could pivot off axis to a certain extent independent of one another. I assume a simple and perhaps more novel solution is at play here, but that's how I always saw it as operating.
If the entire fore and rear assemblies move, they will have to take into account the tube of the drive shaft mechanism, which does not move (at the extremities).
The real problem is the lack of 2 points of xy rotation in the system. Even if the hip can turn in xy, the ankle cannot without modification, leading to an inability to cleanly turn the whole vehicle. It's the reason I had to split the footpad and make the bottom rotate on the AT-SE; the lack of a real ankle is a problem with making these walker leg designs "work."
You can fudge it for TESB, but you can't really with a 3D model, at least one that isn't meant to be static/only go in a straight line.
I would like to make clear the current animation looks really great, but if it follows the existing layouts seen in the refs, it'll only work in a straight line unless Stonky modifies the original design to allow for turning.
That is interesting. I havent looked too closely at the cutaway diagrams. The only one that I reckon to be correct is the one by Binkerman.:)
Wasn't aware that a driveshaft traveled through both assemblies. I just figured that each drive assembly had its own dedicated systems and they were linked via computer and terrain plotting computers.
You are probably correct though. As I said, I havent looked too closely at it.
Stonky, your animation cycle is looking spot on. It looks solid, heavy and definite: not slow and lumbering. The chin blaster timings are great too. And I approve of the wide angle shot
... a walker sidesteps in TESB to shoot a snowspeeder.
That's my favorite AT-AT moment of the film. I tried to work out how the model was hinged to allow that shot. The best I could come up with was that the hip joint, which is almost completely obscured may be a ball joint. The alternative was that they bodged it for that one scene.
The rear right leg seems to rotate a few degrees then bend back at the knee. The rear left leg swings out to counter balance.
The ball joint coupled with the foot damping mechanism should allow a few degrees of turning on the move or a horse like sidestep youtube link. It's feet may be a bit big for that however.
That's my favorite AT-AT moment of the film. I tried to work out how the model was hinged to allow that shot. The best I could come up with was that the hip joint, which is almost completely obscured may be a ball joint. The alternative was that they bodged it for that one scene.
The rear right leg seems to rotate a few degrees then bend back at the knee. The rear left leg swings out to counter balance.
The ball joint coupled with the foot damping mechanism should allow a few degrees of turning on the move or a horse like sidestep youtube link. It's feet may be a bit big for that however.
you can see this in action in the game Empire at War too, though I don't know if it's the same joint you're talking about, in the game it's the last leg joint (if that makes sense), the one where the legs attach to the body near the centre
I always assumed that joint or the next joint would be a ball joint
the one where the legs attach to the body near the centre
Yep, thats the one I was calling the hip joint. It's visually and mechanically odd (from a real world perspective) but it does allow a nice range of movement and could be implemented with minimal design changes.
Thanks all for the kind words. I still have a lot of work to do both with regards to animation and the model itself, and work continues. It's interesting once you go from shot to shot in the actual footage to do some real analysis you can see how different each step really is, even amongst different walkers in the same shot (animated by different people which would explain the inconsistencies I guess). Now that I have the rough mechanics in the rig worked out I'd really like to get into some animation as the model moves along towards texturing and ultimately completion.
As for the turn, I think that you're right, the first assumption that you'd have to make is that there is a ball joint connection at the hip which is counter to all of the reference out there. If you can make that leap, it should be able to make a turn without sliding the feet on the ground I think, I'll have to try it and see what happens. Given the very little clearance the hip parts would have laterally because of the central cylindrical "drives" (I guess? Let's call them that, sure ) and the interior of the hull, each sidestep would have to be pretty small. The stepping pattern would probably want to change as well, from left left right right to left right left right.
Again, thanks for the encouragement, it means a lot.
We would guess that while there is a ball joint, it is not capable of all that much "balling". It probably can only turn a few degrees away from the central axis that would be defined if there were no such ball joint. However, that should be more than enough to allow sufficient turning without dragging feet.
yeah in empire at war it's small sidesteps basically, but it works
I always wondered what it would be like to show the AT-AT moving a bit faster, or climbing steep rocks or something, just to show it can handle more than flat ground
The AT-AT is suited for slow changes in incline, ignoring lesser changes in terrain, whereas the AT-TE is truly suited to moving along the uneven surfaces.
One thing can be said about the AT-AT... If you don't have high ground available, bring it along with you.
I seem to remember reading a passage in one of the godawful extended universe books about an AT-AT 'running' through a forest knocking down trees and making a thundering entrance to some such battle. I think that would be cool to see (and hear). Your model is AWESOME Stonky! and the animation is pretty damn sweet too!
Wow!!!!
That's a nice rig you have set up.
So you use Houdini eh?
I've heard good things about it; especially for use with dynamic and fluid effects.
I've been thinking of looking more into it.
Its seems the need to learn new software or keep up with new versions (every 18 months)
never ends
-sigh-
Thank you! the rig is definitely a work in progress but I think all of the broad strokes are there. I'm working on the "brackets" that attach to the top part of the leg close to the hip joint now, automating their positions based on the positions of the legs, we'll see how it works out.
Yeah, houdini - I just seem to keep going back to it. I like it for it's meticulous nature, perfect for this sort of thing. So true about software - I heard an analogy once that compared a new car to a software upgrade, and that the equivalent to the software changes that we've all come to expect would be to find the steering wheel in the trunk on next year's model.
I've got the moving "brackets" worked out for the "hip" joint, and in between moving forward on the unfinished model I've been kicking around some ideas for displacing snow for the footsteps/foot prints. I'd also like to drop snow from the feet as they leave the ground and blow snow around from the ground as well. I'm also working on a system to dispel smoke from the main canons when they fire. And I still have to texture the thing.
very nice, the only thing I can add to this is the walk cycle seems a little fast when compared to the movies (although been a while since I watched Empire so could be wrong ) other than that, its all fantastic work!
Agreed, it's too quick - there also shouldn't be a pause between each step (like I have here). As each step settles the next one should already be "decompressing" in preparation for it's move. It's interesting, it looks like (based on everything that I have been able to cobble together, from various sources) that they originally screwed each step into the ground when the the foot made contact with the ground, using a screw that entered the foot from the bottom. The foot "compression" was animated by tightening the screw, compressing the spring loaded foot. The screw would also keep the planted feet in position as the walker inched forward. When it was time for the foot to decompress the screw was loosened a bit each frame, decompressing the spring loaded foot until it was completely unscrewed when it was time for the foot to leave the ground. I can't even imagine how laborious it must have been, I've got it easy.
The animation that I posted previously was also 30fps instead of the proper 24 fps which will be adjusted in the next posting as well. Each of my steps was 24 frames from lift off to landing, which would be less than a second playing at 30fps. Upon looking at the footage, I'm going to look at 36 frames a step (from lift off to landing), playing at 24fps and see how that looks. I see steps taking anywhere from 30 frames all the way to 50 frames in the actual film, quite a range to choose from, but the average appears to be around 36.
Very nice m8, looks great. Brings a tear to my eye though as I had one of these and it got put to one side while moving house ect, I never got to finish it as my hard drive with all my stuff on stopped working, lost the lot . Got disheartened for ages but seeing yours has got me thinking again lol. Gr8 work... Here was my effort and my only picture of it. http://www.scifi-meshes.com/forums/showthread.php?74745-At-at/page2
Posts
Yeah, ILM based the AT-AT movement from an Elephant's walk, .... which was shown in the Making of Empire (I recorded on Video yonks ago (mid 80's) hosted by Mark Hamill ) ....
The AT-AT does look deceptively simple to do, .. until you do one, ... So I for one appreciate the effort you're putting into this Stonks .... keep going mate, you're doing a great job so far :thumb: ....
Jas
I love the old behind the scenes footage from the original trilogy, thanks for the link DCB!
It'll turn like an elephant I guess, it's legs can come out to make that sort of maneuver, they're not constrained in any way.
Thanks Binkerman, much appreciated!
Here's something I threw together as the rig continues, still LOTS to do with the rig and the model itself. Animation is still very rough, etc., etc.:
The walker must be able to pivot the hip joints in xy, given that a walker sidesteps in TESB to shoot a snowspeeder. That said, it's not at all clear how it works, given that drive shaft thing along the underside. Even if the hip connection to the drive shaft could pivot in xy (and it might look pretty odd), the ankles and legs absolutely cannot, so you're still left with some pretty non-organic animations required for the turn; it's not going to be as easy to animate as an elephant. I'm looking forward to seeing how you handle it. It was never a problem in the movies since walkers never turned in the movies.
The real problem is the lack of 2 points of xy rotation in the system. Even if the hip can turn in xy, the ankle cannot without modification, leading to an inability to cleanly turn the whole vehicle. It's the reason I had to split the footpad and make the bottom rotate on the AT-SE; the lack of a real ankle is a problem with making these walker leg designs "work."
You can fudge it for TESB, but you can't really with a 3D model, at least one that isn't meant to be static/only go in a straight line.
I would like to make clear the current animation looks really great, but if it follows the existing layouts seen in the refs, it'll only work in a straight line unless Stonky modifies the original design to allow for turning.
Wasn't aware that a driveshaft traveled through both assemblies. I just figured that each drive assembly had its own dedicated systems and they were linked via computer and terrain plotting computers.
You are probably correct though. As I said, I havent looked too closely at it.
That's my favorite AT-AT moment of the film. I tried to work out how the model was hinged to allow that shot. The best I could come up with was that the hip joint, which is almost completely obscured may be a ball joint. The alternative was that they bodged it for that one scene.
The rear right leg seems to rotate a few degrees then bend back at the knee. The rear left leg swings out to counter balance.
The ball joint coupled with the foot damping mechanism should allow a few degrees of turning on the move or a horse like sidestep youtube link. It's feet may be a bit big for that however.
I always assumed that joint or the next joint would be a ball joint
Yep, thats the one I was calling the hip joint. It's visually and mechanically odd (from a real world perspective) but it does allow a nice range of movement and could be implemented with minimal design changes.
As for the turn, I think that you're right, the first assumption that you'd have to make is that there is a ball joint connection at the hip which is counter to all of the reference out there. If you can make that leap, it should be able to make a turn without sliding the feet on the ground I think, I'll have to try it and see what happens. Given the very little clearance the hip parts would have laterally because of the central cylindrical "drives" (I guess? Let's call them that, sure
Again, thanks for the encouragement, it means a lot.
I always wondered what it would be like to show the AT-AT moving a bit faster, or climbing steep rocks or something, just to show it can handle more than flat ground
One thing can be said about the AT-AT... If you don't have high ground available, bring it along with you.
Wow!!!!
That's a nice rig you have set up.
So you use Houdini eh?
I've heard good things about it; especially for use with dynamic and fluid effects.
I've been thinking of looking more into it.
Its seems the need to learn new software or keep up with new versions (every 18 months)
never ends
-sigh-
Yeah, houdini - I just seem to keep going back to it. I like it for it's meticulous nature, perfect for this sort of thing. So true about software - I heard an analogy once that compared a new car to a software upgrade, and that the equivalent to the software changes that we've all come to expect would be to find the steering wheel in the trunk on next year's model.
I've got the moving "brackets" worked out for the "hip" joint, and in between moving forward on the unfinished model I've been kicking around some ideas for displacing snow for the footsteps/foot prints. I'd also like to drop snow from the feet as they leave the ground and blow snow around from the ground as well. I'm also working on a system to dispel smoke from the main canons when they fire. And I still have to texture the thing.
Agreed, it's too quick - there also shouldn't be a pause between each step (like I have here). As each step settles the next one should already be "decompressing" in preparation for it's move. It's interesting, it looks like (based on everything that I have been able to cobble together, from various sources) that they originally screwed each step into the ground when the the foot made contact with the ground, using a screw that entered the foot from the bottom. The foot "compression" was animated by tightening the screw, compressing the spring loaded foot. The screw would also keep the planted feet in position as the walker inched forward. When it was time for the foot to decompress the screw was loosened a bit each frame, decompressing the spring loaded foot until it was completely unscrewed when it was time for the foot to leave the ground. I can't even imagine how laborious it must have been, I've got it easy.
The animation that I posted previously was also 30fps instead of the proper 24 fps which will be adjusted in the next posting as well. Each of my steps was 24 frames from lift off to landing, which would be less than a second playing at 30fps. Upon looking at the footage, I'm going to look at 36 frames a step (from lift off to landing), playing at 24fps and see how that looks. I see steps taking anywhere from 30 frames all the way to 50 frames in the actual film, quite a range to choose from, but the average appears to be around 36.
Thanks again!