Research: exhaustive Nimso Project TX study (82 video transcripts) + Boneworks/BIMOS rig reference

- research/nimso-project-tx.md: chronological per-video notes 2017-2026, topic synthesis,
  VR Messs scope, proportions & calibration (Winter/Drillis-Contini/Dempster tables verified),
  open unknowns
- research/sources/nimso-transcripts/: yt-dlp auto-caption transcripts + dated index (re-pullable)
- research/vr-physics-rig-hexabody.md: BIMOS/HexaBody/Boneworks reference, Box3D native joint
  mapping (spherical/motor/prismatic/distance), superseded-as-target note
- AGENTS/HANDOFF research pointers

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

research/sources/nimso-transcripts/9IDhcrzJtY4.txt

@@ -0,0 +1,3 @@
+# ProjectTX - IK-Less Control, WHY and HOW I'm not using IK! (9IDhcrzJtY4, 612s)
+
+[Music] me name Sony welcome back to another video of project TX this video is not actually about the project itself but rather about the IKEA system so bunch of comments on the last video in regards to what I meant exactly by no ika is involved with this character model here so obviously people assume you know this is pretty much an IKEA based calculation right and I mentioned obviously it's not an IKEA calculation so I'm just gonna explain exactly what the difference is in terms of what I'm doing as opposed to a sort of standard I okay and to do that obviously I need to explain what I K is so imagine the following scenario you've got an arm here which is lodged over here this is actually just a physics joint it's literally a standard object that I can just move around and it sort of looks like an hour and I've got an invisibly movable object here and I can just sort of float about but ignore that for a moment imagine this all now the way to control this arm if we were moving it in in using either physics or animation or whatever is we would rotate this to move this arm or the foot of the first part of the on the upper arm into position to where we want to go and then we'd rotate this you get the forearm to do whatever we want to do that but what if we know where we're standing and we want to get the hand point which is the end of the forearm here to actually match up with a certain position so let's say this position here well then we're gonna need to do some calculations because the problem with this is the only inputs we have is the angle here and the angle here or the other rotations are those two points though that little bit of information does not get as this point here I'd have to just sort of guess this and then guess that and guess this again and guess that and as you can see that I can get close to it but that's all just eyeballing it that we're just guessing we're just guessing if we want to get it to this actual position here we need to control this point and tell the system where that point is and then using that go backwards instead of forwards because remember we rotate this that moves the whole thing then we rotate this to get the hand point wherever the heck it wants to go inverse kinematics there's a backwards we decide where we want to be and then we move it we calculate backwards essentially so we've got a connect a kinematic object something that moves and then we do the inverse to calculate the angles that we need to rotate the object to the end result of course is this so we have a point which is the actual arm point and then we have a hand point which decide where where the end result is and of course it can limit itself based on the length of the arm and then we have a helper now the helper simply just rotates that part on the opposite on the sort of axis pointing inwards which decides which way to go now you'll notice this one is not a physics object so I'm not able to push it around right and that's because this is just purely I K so this is how inverse kinematics works you take one position there you take one position there you do some clever maths and the end result is you get the arm that it gives the system the rotations it calculates those rotations and the hand is always at the correct position as a result that's exactly what we want in terms of control but obviously we have a problem you see if we have a physics based system like I have here as I've shown in the last video if you haven't seen it please do check it out that project TX is the name if we have a physics based system we now want to decide what happens when a physics object is here so what you'd expect to happen is either pushes in it pushes to the side or it pushes to the other side that's what we expect to happen but how do we calculate them so I'm gonna put my hand here what you'd imagine to happen either we have it moving to the side this way we have it moving to the side that way all and this is the worst case we have the whole thing moving up which means we are going to have to control these two points we're gonna have to do a bunch of calculations it it doesn't really work out very well in terms of how we want to control things in an easy way over to this side now this side it may look like the same thing as you can see very similar thing happens I move this it rotates the a central part I move this it sort of car gets that position however the thing with this it purely physics there's actually no inverse kinematic calculation going on so this is very similar to the way I'm doing my arms here except this is without these soft fist trackers so it's a little bit hovering as you can see look everything just sort of floats in place so that's without all of my physics training but you can see now with this I can't push the elbow I can push the elbow exactly where I want it to go and that's precisely the reason that I wanted to use this type of logic there is absolutely no I care going on here there's no inveLine truncated