This is for Frost 1.0.5.44250 used with Thinking Particles 4.0.5.7.
I ran into an issue with getting Frost instanced geometry to acquire the orientation channel from Thinking Particles. It is definitely reading something from TP as changing the alignment will affect the frost orientation, but the result does not match up with what the particle is displaying.
Looking at the debug log in TP, it is using a 9 value matrix to define the alignment (3 per axis). I figured Krakatoa would use the same orientation format as Frost so I saved out a PRT and got a 4 value matrix there (assuming 1 per axis plus an up value). I’m curious if there is a step I am missing to translate the data properly.
On a less pressing note, I noticed that the orientation type has a tendency to switch itself to “specify orientation” when adjusting the TP alignment with frost active.
Got a simple scene file attached (max 2010 format), just one particle rotating 360 on the X axis. Orange object is the TP data, Red is the Frost result.
The Orientation channel of PFlow is a Quaternion (4 values, angle and a Point3 value as the axis to rotate about). So Frost expects the Orientation to be a Quat (Point4).
When Krakatoa is saving Orientation from TP, it takes the Alignment channel and does a conversion to the desired format, but it looks like the value is not right.
Also, Krakatoa does not save a Tangent channel from TP correctly (it saves [0,0,0] no matter what), otherwise you could use the Normal X Tangent to build an Orientation value in KCM. And the Normal value stored in a PRT is also wrong, particles oriented to the World produce a Normal of [-0.841309,0.540527,0] instead of [0,0,1] and the Orientation is [0,0,0.479492,0.877441] instead of [0,0,0,1]…
These look like bugs to me and should be addressed soon (Darcy was working on Krakatoa / Frost / TP improvements this morning).
ah ok, good to hear it’s not just me. Thanks for the rundown on how the rotations are calculated, I had been playing with recreating the desired data in TP and saving it to a custom data channel, but got stuck at the point where the individual values needed to be composed together (using the tangent data sounds like an interesting thing to try once that one is ironed out). The rig was simple enough to remake in pflow so we have a solution for the moment.
Glad to hear more improvements are in store for the TP connection. I’ll toss in a +1 for the request of letting Frost pick individual TP particle groups and mention that the prior version of krakatoa that let you specify groups to record in the interface (as opposed to using renderable flags in TP) seemed quite a bit easier to manage in rigs with lots of nested groups…
That Rollout in Krakatoa is STILL there (but under Krakatoa Explorers), and it does set the Renderable flag for you. So we did not remove it, we just moved it next to the PRT Loader/Volume/Fume UIs, and it still does the same, just through a more convenient UI.
In the Krakatoa GUI, press [Explorers] at the bottom of the Main Control rollout and you will find the “Thinking Particles Groups” rollout.
