A:M Dave

Thank you, robcat2075. I am familiar with tagged structures. I was hoping not to have to go there, but will give it a go. Thought there might be a utility somewhere. Odd that most of the models are opening with little trouble.

I'm trying to freshen up my older models. Most are importing pretty well. I have one (ver 9.1), so far, that causes A:M to crash before it fully loads. It's not a big or complex model. Is there a way to step through the load to find out where it's hanging up? Is there some sort of conversion tool somewhere? I looked at the text of the .mdl file, which was not corrupted. First glance, all looked normal.

Very nice. So scanning this string, I'm still not clear... are you using a plug-in or some stock technique? I can't think of any way to get stock A:M to capture the "below the surface history" information. I believe much of the snow prints in Ice Age were done this way. Hold the phone... just saw the bit on the .hxt plugin. I'm working through the new features of v13. Too bad you say it doesn't work there. Are you going to upgrade it?

Rodney, Thanks for the opportunity. Please consider me for the free upgrade. I'm a long-time user of A:M.

Do Americans with Danish heritage count? I've lived in Stavanger, Norway and was fairly fluent at one time.

Situations: I've wanted to do this most when a wheel on each side of a vehicle needs to turn at a rate commenserate with the path that it is following. Usual solution: Ignore it and hope speed, distance, object occlusion and/or detail do not allow the viewer to see that the two sides are locked together around a corner as if they are on a solid axle connecting them. Typical construction is the vehicle has a center bone constrained to the vehicle path. An action with stride length gives the illusion of grip to the surface. Straight paths are fine, but curves are not accurate. Wanted: A way for A:M to adjust stride length differentially depending on instantanious radius of curvature with the path and offset of the wheel from center in the action. (I'll use the example of the front steering wheels of a buckboard) Things I've tried: 1. Make a path for each wheel. Each wheel is modeled separately (or mirrored). The rest of the vehicle is a separate model. Make a stride length action for the wheels. In Chor, make two paths equidistant from each other. (Rotoscoping can help in path drawing, but if the vehicle path turns an off-camber corner around a hill, inaccuracies creep in.) Constrain wheels to respective paths and drop stride length Action on them. Assemble vehicle to "float" between them by way of Translate To constraints from vehicle to each wheel. An Orient Like constraint will probably also be needed. Results: It takes some careful setup and layout, but the results are really close. Shortcoming: The biggest problem is keeping the two wheels opposite each other. I have been leaving one Ease alone and have been trying to adjust the other to match by eye. Is there a way to control the Ease of one path with the Ease of another? Sure wish a model spline could be used as a stride length path in Chor. Sure wish a path could be made in an Action. Sure wish there was an "copy offset curve" function. 2. Use a vehicle constructed on center with straight line stride length. Use Poses for wheel turning. In Chor, adjust rate of wheel turn with Pose over the top of the Action and stride length motion. Results: Doesn't work. Shortcoming: The Pose overrides the Action. I haven't found a setting that blends the two. 3. Use on-center constructed vehicle. Poses for wheel turning. Forget stride length. Constrain vehicle to path in Chor and manually adjust rate of wheel turn (probably with repeating Action). Results: Good for wheel slipping maneuvers. Shortcomings: Laborious for good "traction" matching. 4. Avoid this situation and don't worry about it. Results: No one will notice except me. I like the certainty of approach #1. I've been thinking about relating some attribute to the center of the curve or some throttling bone, but I've never used the formula or multi-relationship-driver capability. Is there a good example anyone knows of to go dissect?

I think I found the problem. In Action, I only explicitly set the Ease on the first path thinking that once the length of the Action was established, other path constraints would "assume" the length by default. Not so. Once I explicitly set the others, they worked in the Chor also. Now on to the next challenge, that of getting left and right turns to work well. I've got one scheme, but am going to try something else. I'll let you know how it works out.

Thanks for the reference. I had not seen that yet. Dhartman's approach is very similar to the one I was trying. I think he ran into the same problem I did trying to get the lead elements not to bunch up at the end of the path at 100% ease. I used separate, connected spline loops. He used a slick trick of an open-ended double loop. My elements' ease goes 0% - 100%; his 0 - 50% (for the 1st element). (If I had a trick, it was getting the copies of a spline to shift their first control point to different nodes.) I'm going to try his construction to see if that holds up. What I don't understand is why my action is breaking down in the Chor. Either of these modeling approaches should work just as well. Does my "broken action" work on anyone else's machine?

Does someone have an idea regarding what is happening here? Here are the evidence and facts in the mystery... Three attached files tell the story: 1. Really Broken.prj --- Look in Chor and you should see just the red plate go around and the others moving but v-e-r-y s-l-o-w-l-y. 2. Working Action.prj --- This Chor does what it should: all plates go around the path of the tread. 3. Good and Bad Action.wmv --- Knowing how demos go, here's a video showing what I got. The catches: The project with the broken Chor was working at one time (even after multiple opens and closes) Both Chor's in both project files were built in an identical fashion. Looking at the motion in the Action works in both files. Any ideas anyone? ...or... A more elegant (and robust) way to do this? Following are the things I've tried or thought of: a) Single spline with many plates distributed around with Ease. Problem: Ease greater than 100% is not allowed. Chaining all the plates to a lead plate (like a train engine with cars). The lead plate is Constrained to the spline. The others are "managed" into positions with kinematic links. Problem: OK in theory. Very messy in reality. Managing plates ends up being every frame keyframe ordeal and even then is problematic. c) Thought about making it all round then using some sort of deformation. Problem: I don't want the plates to change shape (squash / stretch). d) Animate an image over the top of a band. Problem: This would be cheating and the results would look too phony. [attachmentid=12433][attachmentid=12434] Good_and_Bad_Action.wmv Broken_and_Working_Project_Files.zip

Sydtocreation: I was working on my own tread project when I saw this. I think I can help. Turns out the Sherwood Forest Treadhed tutorial you referenced is close to what I was doing in my project. The problem is that version 11 does not allow Ease on a path to go past 100%. I didn't try an older version, but I seem to remember that previous versions DID allow this. (I might be wrong on that.) If all the tread plates are placed on the one spline and distributed around the loop with Ease percentages, then they will all stack up at the end of the action when they all reach 100% and the single end of the single spline. You could trick the eye by not having the same plate go all the way around, but rather only to the next plate position. If you are careful with modeling, and all your plates are identical, then a repeating action with stride length would not show the plates don't really go all the way around. I wanted the ability to have a different colored plate and watch it go around. I ended up construction a path spline for each plate. The spline has at least a control point everywhere a plate origin will be. Each new copy of the path spline was directly on top of the original. With each new loop, I named the group (so I could find it again). Through a process of hiding all but the loop I was working on, you can break and reattach ends and beginnings incrementing the break around the loop. This re-sequences the path loop and changes the position of the first control point (which will be 0% and 100% Ease). I see you posted this way back in August, but let me know if you need more.