Is this just 100% random or is there some pattern to it?  I just added a pair of symmetrical bones to a figure (M4) and the first three tries, I got the bone flipped problem, the fourth try the bone works correctly.  I'm nearly certain I took exactly the same steps each time.  What is the trick?  :cursing:

awesome, I changed joint rotation order and there went the bug (4th time in a row).

You got gimbal lock.

With a new bone, every time?
ps got it just now on the 4th try, and did NOT change joint rotation that time

Yep happens with bones that are set at 90 degrees.
Would you like to see a picture?

If it will make sense, because I don't understand what you're saying.  90 degrees to what?

Why are you showing me a pic from Lightwave?  Rotating the joints in the Setup room, I can see plainly that the bone is moving on a different axis with each of xRotate, yRotate and zRotate.

Why not?  Gimbal lock is a phenomenon that happens in Poser,  Lightwave,  and even pilot instrument panels. 

Okay, since I'm not flying a plane or using Lightwave, do you have any information on avoiding this bug in Poser?

You'll have to record pivot rotation.

That is not helpful at all, Patorak.  Thanks anyway I guess, maybe Smith Micro support will have an answer.

A brand new pair of bones, not edited in any way except to turn on spherical falloff zones, and the bug appears again when running Figure -> Symmetry.  :cursing:

Hi pjz99,

I'm not quite understanding what you mean by bone flipping. Do you mean it flips its rotation?

I mean, when I turn on spherical falloff zones for a pair of bones that were created (apparently) symmetrically, and then do Figure - > Symmetry -> Left to Right (or Right to Left) -> Copy joint setup information, one side's falloff zones are flipped and oriented at -180 degrees on two axes.

edit: in case that isn't too clear, if the falloff zones are not set up absolutely simply (e.g. if they are rotated or scaled) then this will produce an asymmetrical deformation, which is probably not good for a human character.

I know what you mean now, But i'm sorry I don't know how to fix it. I see it all the time in all sorts of figures so I just thought it was normal. Have you tried copy and paste? I'll go try and figure something out.

cheers,

Mike. 

Copy and Paste will paste in the wrong values - e.g. the symmetry axis will be correct, but the other two axes of rotation need to be multiplied by -1, and also translation values.  You can fix the values manually (I've tried this several times) but as soon as you do Figure -> Symmetry again, the values are broken again.  It is "normal" in that the more you look at different figures, the more you notice this bug, but it is not desirable behavior AT ALL.

yep its a bug in the poser rigging system which has been a problem since at least Poser 5 and probably earlier, and as patorak states has to do with the fact that the tools are so sensitive to gimble lock. this issue also accounts for the strange crazy spinning of spherical falloff fields when rotation angles approach 90 degrees (or multiples thereof) the only way I have found of avoiding it is to zero the rotations of both falloff zones and try a different rotation to get the same basic shape.

I checked Hiro3, James, and M3 and the fall off zones are all 180 degres opposite of eachother so if one zone says 31 degres the other side says 149 degres..  In every case for the up and down bones of the legs the falloff zones are rotated on the X axis 180 degres.  So that must be normal.  Did you mean something else?

The thing is, this problem is appearing for me with new bones that are drawn nearly perpendicular or parallel to the ground plane, before ANY modifications have been done to the falloff zones they become asymmetrical.

Create bone 1(several times I actually just drew the bone without moving the endpoints at all)
Create bone 2
Figure -> Symmetry -> Left to Right (or Right to Left)
... At this point the bones appear to be symmetrical...
Enable spherical falloff for one axis on bone 1
Enable spherical falloff for the same axis on bone 2
Figure -> Symmetry -> Left to Right
 ... at this point the falloff zones are no longer symmetrical (two axes show -180 degree rotation).  Nothing has been rotated or moved, how can gimbal lock be causing this?

Out of about 8 tries, I have not been able to get a symmetrical set of falloff zones added onto an existing rig.  I'm gonna do a humanoid rig from scratch and see if this reoccurs with a simple figure.

Mike if you look through the Top camera or Left/Right cameras, if the spherical falloff zones for opposed joints like left shoulder/right shoulder do not overlap absolutely perfectly then that is a serious problem - the rig will not produce a symmetrical deformation (SP3's leg rig e.g.).

Yes X and Z  rotations for the fallof zones will be 180 on one side while on the other they will be 0. It's always that way.

Oh, now that is weird the falloff zones should line up perfectly fron side to side.

No, it isn't always that way and should never be that way, because that's not symmetrical.  Look at M4's shoulder falloff params for example ... -54 y / 0 x / -5 z on the left, 54 y / 0 x / 5 z on the right side.  If the falloff zone is perfectly spherical, then one of the sides can be 0 and the other can be -180 and it won't appear to have a problem, but that is not symmetrical!

You can rotate Y and Z but then you have to have X at 0.  You can rotate X and Z but you must have Y at 0. You can almost never get all three to rotate, because it freaks out and spins out of controle.  If you want to rotate Y and Z then set both sides X at 0   I sure would like to use all three sometimes too, very irritating.

I think you are on the right track with the reason for this bug, try this with steph petite on the right shin bend, rotate the inner sphere Z rotate from 180 to 0 now do a symmetry from right to left, this flips the zrotate from 0 to 180  on the left leg but also this snaps the falloffs into alignment in poser. the key is that both the X and Z rotations were on or close to 180 degrees in one sphere which is bad, this causes the gimble lock issue.

Hmm guess its too late here, and Im getting bleary, for some reason I cant repeat that result but definately you are right about the flipping.

That's why I'm calling it "bone flipping" because that really seems to me what is going on, that a bone's orientation somehow can become flipped internally.  I haven't found any pattern except that, when adding bones off the chest on M4, it seems to happen 100% of the time....

pjz, is it only getting this problem when you use symmetry twice?  If that's the case can you try doing everything on one side only and then using the symmetry tool.

I've tried that, and it hurts worse (more work is wasted when the bug appears).  You can see this behavior on a wide variety of figures - SP3, V4, Sydney for example and I'm sure there are others.  Figure -> Symmetry ought to make all the falloff zones mirror each other, but for some specific bones - not random ones, something that happens 100% consistently when it appears - the falloff zones are "flipped" and not oriented correctly.

Smith Micro support has reached the stage of asking me for my serial number (which they already have attached to my login for their ticketing system).  Lightning quick support is always appreciated...  :cursing:

As others have stated it has always been like this, but irregardless the falloff zones should mirror over correctly. I have manually altered those rotations values to reflect true symmetry and the falloff zones go wonky needing maual correcting. When I did this I really didn't see much difference from the Poser symmetry and my manual edited symmetry for my rig...it still seemed to act the same. (ie; it will always follow the falloff zones)

No, it has not always been like this, and it is not always like this on every bone of every figure.  It's definitely abnormal behavior that isn't present in all rigs.  "Don't use the symmetry feature" is a shitty workaround, I want something better than that (maybe like, having the bug fixed).  Having an object's orientation flip to -180 degrees is not the same as 0 degrees and is definitely not symmetrical.

This might sound like a dumb suggestion, but what about building (or editing) a rig with PHI Builder, bringing it into Poser, and see if you can set the values without them flipping.  I remember something about this when I was new to Poser, some 5 years ago.- Starkdog  

Yeah I'll be fooling around with this a bit with a scratch rig, but a current task of mine is to add bones to an existing rig (e.g. M4).  Thanks though, I'll try your suggestion too, might be helpful for different purposes.  I really am beginning to think the real problem is that Poser's symmetry tool is just broken.

Here's another strange thought.  Have you tried this with DAZ's rigging tools for D|Studio?  I have yet to rig anything with the toolset, but it looks very interesting.  I bet if you submitted a ticket to E-F, um, SmithMicro, we might see this issue fixed by Poser 10!!! Hehehe, -Starkdog 

Quote - Here's another strange thought.  Have you tried this with DAZ's rigging tools for D|Studio?  I have yet to rig anything with the toolset, but it looks very interesting.  I bet if you submitted a ticket to E-F, um, SmithMicro, we might see this issue fixed by Poser 10!!! Hehehe, -Starkdog 

 
The Daz setup tools are super! You can even use all three rots at once on the falloff zones. Of course they dont exactly translate to poser.

I've looked into how this problem manifests in DAZ|Studio - and actually it doesn't manifest at all.  For example the SP3 shoe rigging problem I had a few days ago, the same CR2 is not interpreted the same way between Poser and D|S.  At any rate I don't own the D|S rigging tools and am not interested in buying them, because I'm certain they won't work around this problem - look at all the DAZ content that shows this kind of falloff bug (e.g. V4). 

Now that you mention that, it makes me realize the problem is as simple as the Poser symmetry tool might be broken, it's something fundamentally bugged in how falloff zone coordinate data is handled internally by Poser.  Falloff zones that are shown as -180/-180/0 in Poser can be shown as 0/0/0 in D|S (as they should be), when looking at the same CR2.

An update: Smith Micro support has gotten around to asking me how to reproduce this.  So fast the G-forces are killing me!