Thanks LovePyrs, didn't know about that, I'll have to consider maybe going that route. I also really wanted to remap V2 to the V3's uv's because the v3 mapping has less seams.

This is frustrating because the info must exist, but we can't get to it. Without this, a method of collision detection via python wouldn't be possible. I tried "WorldVertex()" also and got the same results as with "Vertex()". I tried each method with the casual business man using this script; ########################################### scene = poser.Scene() actor = scene.Actor("Left Forearm") geom = actor.Geometry() vertMS = geom.Vertex(22) vertWS = geom.WorldVertex(22) print "model position: ", vertMS.X(), vertMS.Y(), vertMS.Z() print "world position: ", vertWS.X(), vertWS.Y(), vertWS.Z() v = actor.Parameter("Bend") v.SetValue(-100) print "moved model position: ", vertMS.X(), vertMS.Y(), vertMS.Z() print "moved world position: ", vertWS.X(), vertWS.Y(), vertWS.Z() ########################################### This script prints out the model space and world space xyz coordinates of vertex #22 in the left forearm before and after the forearm is moved. In each case, the locations remain the same. If you do figure out a solution to this, please post it! josiah

If I am not mistaken in what you are saying, I ran into this when I tried to create a script that moved the vertices of a figures bent elbow objects into alignment with an ideally shaped bent elbow object made in 3dsmax. I used this approach because it is hard to create a morph on a straight peice that looks the way you want it when bent. Anyway, instead of moving the vertices of the mesh I saw in the scene, it moved the verts from the "original" object which in turn doubled the bend! It seems that the Poser keeps two versions of the object meshes, the original object and the modified "virtual" object you see in the scene.I believe Poser creates a new pose or new morph by moving the vertices you see in the scene(not the ones in the base object). However when you try to get a vertex coordinates, Poser wants to give you the ones of the unaltered base object and not of the "virtual" one you see in the scene. I can't say I know away around this. I wish I did! If I am off base here on what you mean, this script will load a pose; ####################################################### scene = poser.Scene() scene.LoadLibraryFigure("Runtime:Libraries:character:People:Business Man.cr2") scene.LoadLibraryPose("Runtime:Libraries:Pose:Action Sets:Stalker.pz2") scene.DrawAll() ####################################################### (paths for Poser 4, 4 lines of code, all start with "scene"). I hate these little reply boxes!

Ok, the generated exception seems to be related to the time before looping the "Update" function. On my machine, if I set the ms in the line; root.after(250, self.Update) to less than 250, a exception will be generated if I try to reposition the TK window. If I set it for higher than 250, no problem. Again, this is only when using; scene.ProcessSomeEvents() to let Poser run along with the script. ProcessSomeEvents takes an argument as to how many poser events to process before it returns, but it seems to work the same with, or without a number specified. Could it be that when the loop time is short, ProcessSomeEvents is called again before it is finished? But that dosen't seem right! I wonder, if there is some way to use "after_idle()" to wait untill all events are processed, before calling "Update" again instead of using a ms setting? I have tried, but I am not sure I fully understand how after_idle works. Any ideas?

Yes it seems to works good, however you need to call "scene.ProcessSomeEvents()" to keep the TK window on top AND work in the main Poser window at the same time. Unfortunatly, this sometimes randomly generates a "C Exception" when you try to move the TK window. I wish I could figure a way around this. I could just make the TK window open at 0 x 0, but you still might want to move it at some point. So for now, I have "scene.ProcessSomeEvents()" commented out, and the TK window stays on top just fine. By the way, I did get Bushi's example to work by adding the line "self.master = master" inside of " def init(self, master):". Thanks Ockham!

This version seems to work fine, except that it sometimes, almost randomly, generates a C exception when moving the TK window, and this seems to be only when calling "scene.ProcessSomeEvents(1)". Perhaps I'm wrong using "root." in the Update function? or maybe the 10ms interval needs to be much larger? This does seem to work REALLY well! I need to experiment more I guess. (This is based on Bushi's example of "poser friendly script".) ############################################ from Tkinter import * import poser scene = poser.Scene() class App: ~def init(self, master): ~~frame = Frame(master) ~~frame.grid() ~def Update(self): ~~scene.ProcessSomeEvents(1) ~~root.lift() ~~root.after(10, self.Update) root = Tk() app = App(root) app.Update() root.mainloop() ############################################

Thanks Ockham! I have started to use the try/except way too much, and looking back over some other stuff I can see where I've made the same error! Your version works perfectly (of course)! Thanks again!

Ok, I found that the origin was moving correctly, but the "origin of an object is not it's location in 3d space but a reference to its own center! So by moving the origin, I am just moveing the point about which transforms occur, not the object itself!

Thanks Ockham! Your help is invaluable! I found that I needed "scene.DrawAll()" to "update" the view. The camera was switching, I just didn't know it. The parameter change code you provided is exactly what I needed! Works fantastic! Thanks again for your time and help! josiah

Wow thanks stewer! That makes alot of sense. I appreciate the help from both of you immensely. Now I am stumped on a tkinter problem: I have a button that runs a function. The function runs once when the button is depressed. I want the function to loop as long as the button is held down. Is there some way I could capture mouse events or something to do this? I want to use something like; while (mouse1 pressed on myButton): I really do try figure this stuff out on my own before I ask for help, but this has me stumped. Maybe its not possible? Thanks again!

DoH! I feel so stupid! Thanks a million ockham! I got it working now! Josiah

Ockham, thanks for your help on this, I appreciate it! You certainly know your stuff! Yes I think that would work, but it would be a lot of trouble for a lazy person like me to line up a very small box with a single vertex in 3D space. Also, when you would zoom in or out with the camera, the size of the box would appear to change size because you are closer or further away from it. Thats why I was hoping to keep the box a fixed depth away from the camera, then you could zoom in and align it with a vertex in 2D space and it could then look along the Z axis until it returns a vertex position. However I couldn't find a way to keep a prop a certain distance away from the front of the current camera. Your idea is great, the best yet, I just wish there where a little easier or faster way to select a vertex. I wonder why CL didn't incorperate vertex movement with "soft selection"? Magnets are hard to work with big time! As far as searching for the nearest vertex closet to the origin, couldn't we take and add or subtract one from each coordinate? This would be 26 positions to sort. However the number to look at would grow exponentially as the positions away increased, but I don't think thats a problem. I think you would have to have an array of the verts of the current body part, and the the box origin could be compared with each and return the closest vert index. The only way I can think to indicate to the user which vertex is selected, would be the vert number displayed on the dialog somewhere. The softselection idea is not that far fetched either, but thats off in the future maybe.

Arrg! Poser doesn't seem to supply enough python functions to create this! Maybe if I start simply and refine? Ok, after looking around the python methods, It looks like this COULD be done; A dialog in which YOU SELECT: 1. the bodypart you are working with. 2. the vertex number you wish to move. Below this would be X,Y and Z, plus and minus buttons to move the vert along each axis in either direction. The problem would be "what is a easy way to determine the vertex number I want to move?",there are ~1000 to choose from! josiah

Hmm.. Good idea Ockham, working on an idea... Any method to return the current camera depth (Z)? Heres a new idea; A small cube with a wireframe texture so it could be seen. this cubes z (depth) is "camera Z + 10". This will keep the cube the same size at any zoom. It should also have a X and Y boundries that are determined by the camera cone so if the camera rotates to where the cube would fall out of view, the cube will be "dragged" along the viewport edge. Zoom in to where you can see the verts you want to manipulate, activate the script and the cube appears. Drag the cube over the vert you want to select using the mouse. Press "Pick", and the script will look along the z axis until it hits a vertex, the it returns the vertex index number. The x,y, and z -/+ buttons become active and the vertex can then be moved with the "set" functions. What do you think?

Thanks! I have them working now. I appreciate the help Ajax and lesbentley! josiah