Script: Measure Rig Performance
As a rigger you have to get along with the animators, and a good startingpoint for that is to provide them with fast rigs. If the animators have to playblast for each time they want to see the actual animation, they are going to waste a lot of time!
I asked this question over at Rigging Dojo about a year ago:
I often find myself implementing a lot of stuff in a rig, stuff I think the animators may need. As I’m improving my rigs, I get more nodes and more calculations going on to achieve what I want – which in the end may result in performance issues with the rig when animating. I’m trying to find out if there’s some way to measure the performance of a rig-component? I find it really hard to find out which components that slows down my rig, as I have to add and animate all of the rig-components together, and then try to compare the playback-speed, remove components, re-check, etc. How do you guys approach this? Also, is there any documentation of speed differences for expressions vs nodes, extra joints vs corrective blendshapes, skinClusters vs wrapDeformers, constraints vs normal parenting, etc?
I got a lot of interesting answers in the thread, but one in particular from Josh Carey helped a lot. He tipped me off on dgtimer, which does exactly what I asked for. So I’ve just wrapped this into a clean little script.
The script measures the speed by playing off the scene and evaluating each DG node’s performance, which means that for the object(s) to actually be evaluated you have to animate each component (that you want to be included in the evaluation) over a given timespan, the more frames you calculate from the more accurate the results will be.
So if you have a rig that runs too slow, you can use this script to see which node(s) that does the hardest impact on the performance. You often find that it’s just a couple of nodes that really takes onto the percentage, stuff like wrap deformers are something you want to avoid (at least in the proxy-rig) 🙂
I’ve uploaded a video to demonstrate how it can be used, in the video I’m using it to compare normal parenting, direct connections, constraints and skinclusters.
3,924 total views, 4 views today
Script: Delete Flat Keys
When you’re working with animation you eventually end up with a lot of extra keys, by that I mean flat keys, keys we don’t need.
See the image below, I’ve marked the flat keys with red dots:
With the script you just select all of the objects that you want to clean the keyframes on, then you define the angle threshold, and the script will go through and delete all of the keys with a lower angle threshold than defined.
When run on the example above, we end up with this:
Script: Hide/Display Local Axis
It can be a hassle to turn on/off the local axis of multiple objects simultaneously, the reason for this is that by default there’s no on and off option on “Display > Local Rotation Axis“, it’s just a toggle. This means that if you select a bunch of objects and toggle the local axis, the objects that already have the local axis visible will be hidden, while the rest of the selection’s local handles will be visible. This can be frustrating at times, especially if you have joints/objects that have the same position and orientation.
This is a really simple script, it basicly just adds the flag that defines on/off to the command that toggle’s the local axis. I’ve also added an option for it to go through the entire hierarchy, definitely a timesaver.
Script: Find Duplicate Names
When dealing with rigs you often end up having a ton of nodes, and it’s easy to end up having several objects with the same name. You may think that Maya automaticly renames those objects for you, and it does, but only if the objects are in the same hierarchy. However, if you have the objects in separate hierarchies they can share the same name.
This isn’t really a problem in it self because Maya will just use longnames to distinguish them apart, but you may run into problems if you’re using scripts that doesn’t take longnames into consideration.
In the image above we have two objects named “pCube”. So what Maya does to distinguish them apart is to look at their parent, if the parent’s name is unique it will use that in the longname when dealing the objects. So if I were to select the pCube in root2, Maya would read it as “grp2|pcube”, and if both of the mid-groups was named “grp2” then maya would read the pCube as “root2|grp2|pCube”.
I’ve made a script that finds (and renames) all of the objects in the scene sharing the same name, the download link is at the bottom 🙂
2,742 total views, 2 views today