Michael Comet's poseDeformer Slows With Each Fix

#1

Hello,
I am rigging a character with a simple bone structure and then using Michael Comet’s poseDeformer script to create psd’s for the problem areas (elbows, back of knees, armpits etc). The setup of the script is pretty well documented online, and the instruction through the process in the UI is incredible. However, after each fix, the calculation time increases almost exponentially! For the first poses that I DO get to work, the psd’s work like a dream, but after about 4 or 5 poses, I’m waiting upwards of five or ten minutes after clicking ‘Create/Edit Pose and Readers’. My only rationale is possibly the fact that I’ve been creating a new ‘poseDeformer’ node for each pose. Is it possible to have more than one pose associated with 1 poseDeformer node? I’ve already checked for a mirrorData node (with DAG objects only unchecked) as I know that slows down the process as well. Am I missing something from this process?

thanks in advance & cheers,
Edan

#2

If this has already been discussed, can someone point me in that direction?

#3

Anyone? I’m also getting some weird override issues where I put the shoulder in a downward position, and if the elbow is bent, the elbow’s corrective shape doesn’t activate, even though it’s parented under the shoulder i.e. moves with the shoulder’s motion. Thanks in advance.

#4

it is kind of difficult to help you without seeing some image or a scene file.Can you post something?

you do know that PSD is not for the faint of the heart?: it takes times and is expensive
Have you thought of editing the membership of your posedeformer( only using the point that are used)?

Personally I use cut up part of my character to do corrective shape( this is not the best method by far but for me its enough ).I also not use the default the geometry given to me to do this job, if you work with less point you can sculpt more quickly. and you can also create a better topology for your need.

the deformer is awesome and use radial basis function and angle between vector to mix the different shape in its database.

I was more comfortable with additive process hence my work on custom posereader with spherical coordinates and geodesic uv range driver.

Do you have setup properly your posereader?

#5

You should be able to have 1 pose deformer for your entire rig. I think. It has been a looooong time since I looked at this stuff but that should be true. Definitely you only need one per region, ie: for a foot area, you only need 1 deformer tho it might have front, back, left, right etc poses in it. I honestly can’t remember if you need more than one per region or not… heh… I should dig up my code again.

If you mirror the deformer there is a mirror data node that you will want to delete because otherwise things will get super slow.

I will also comment that while I was originally playing with the proper math for it, the actual deformer is really simpler and just does some basic rotation/dot stuff to make it work, tho it seems to work well enough… so it doesn’t really do proper LU stuff etc…

good luck…i’d help more but like I said it has been a looooooong time.

ps: thank google alerts for letting me see this post heh…

#6

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.