Incorrect default biped orientation

TackY

Hey!
Got my setup this week and have got it running, had similar problems as jump did with the wifi/chordata net not showing up but managed to solve it - I now bumped into errors in blender.

This is what my blender armature looks like when I'm in N pose:

It boils down to the dorsal and base sensors seemingly having very weird orientations, doing pose calibration doesn't change anything. I'm guessing pose calib doesn't really help if this is what the starting point looks like.

The sensors react nicely to my movements though.

I tried looking around in the code base but I obviously have no real grasp of it, haven't looked at it at all before now.

Will it help if I play around with the trunk_rot in get_JCS_orientations in pose-calibration.calibration.utils ?

daylanKifky

Hello @TackY and welcome to the forum!

The Armature normally looks completely entangled before running pose calibration. After a proper calibration sequence you should have a correct posture.
Could you send me a video of the person doing de calibration sequence? It you are able to put it in a picture-in-picture fashion with an screen capture of the blender GUI it would be great (so I can also see when the buttons are pressed, and how the armature reacts)

TackY Will it help if I play around with the trunk_rot in get_JCS_orientations in pose-calibration.calibration.utils ?

I guess you are a programmer, do you have experience with blender python scripting? Please tell me a little more about what you do

TackY

I did the static calibration, but there was absolutely no change to the armature so I got discouraged from continuing with the specific trunk/arm/leg calibrations. I'll see if I can put together something to show. How long should the static calibration run for, roughly?

I've made some blender addon coding for hobby work. I'm not a proper dev, I do data science/research.

daylanKifky

Yes, there's a bug in the current version of the notochord OS, specifically, the control server fails to initiate the transmission of raw sensor data, causing the calibration algorithm to fail, because the addon is executing it with no samples (you can see in the console that each phase of the calib starts and ends at sample 0)

Updating as @zaanen described shoudl solve that.
And, as I already commented, we are working in a new version of the OS with fixes to this and other problems we discovered thanks to your feedback

TackY How long should the static calibration run for, roughly?

This version of the addon only runs the calibration when it has the data for the complete functional procedure. We will add support for partial sequences (for example only static) in future releases
A couple of seconds for the static part is ok

TackY I've made some blender addon coding for hobby work. I'm not a proper dev, I do data science/research.

Oh great! Even if it's basic, it's not common to find people with bpy experience 👍️

TackY

Actually getting a proper error/crash on the calibration if I continue with trunk/arm/leg calibration.

blender-console.txt

13kB

The actual calibration logs are empty. Partner clicked the buttons and I did the movement.

zaanen

Got the same error at first.
The solution then was to update the system by hooking up the pi with Ethernet cable.
Go to the http://notochord.wlan
go to the second tab and press update all.

After that and waiting for a while my errors where gone.
You can also check if in the last step from the right leg the correct wifi is still connected.
It switched on my from time to time because its looking for internet.