Orientation in the tracker's local coordinate system

[tomassykora]

Hello,

From what I've seen, the orientation obtained by GetDeviceToAbsoluteTrackingPose seems to be in the coordinate system of the lighthouses although I'm not completely sure about that (either that or my obtaining of the euler angles from the rotation matrix is incorrect). How can I get the orientation in local coordinate system of the tracker? Is it possible? I couldn't find an answer to this in the documentation and the data I'm obtaining are a little weird, I cannot decide which directions correspond to specific axes and it's a mess (e.g. when I'm rotating in the yaw direction with the tracker, the roll value is changing also). This is what I use to get the euler angles from a vr::HmdMatrix34_t matrix:

double sy    = std::sqrt(std::pow(mPose.m[0][0], 2) + std::pow(mPose.m[0][1], 2));
double roll  = std::atan2(mPose.m[1][2], mPose.m[2][2]);
double pitch = std::atan2(mPose.m[0][2], sy);
double yaw   = std::atan2(mPose.m[0][1], mPose.m[0][0]);

Thanks for any help.

[ahtur5]

+1 ! Is there a diagram showing the local coordinate spaces of all the tracked devices : HMD, controllers and trackers ?

[bddckr]

openvr/headers/openvr.h

Lines 48 to 56 in cdc39c0

	// right-handed system
	// +y is up
	// +x is to the right
	// -z is forward
	// Distance unit is meters
	struct HmdMatrix34_t
	{
	float m[3][4];
	};

And be aware of

openvr/headers/openvr.h

Lines 243 to 250 in cdc39c0

	/** Identifies which style of tracking origin the application wants to use
	* for the poses it is requesting */
	enum ETrackingUniverseOrigin
	{
	TrackingUniverseSeated = 0, // Poses are provided relative to the seated zero pose
	TrackingUniverseStanding = 1, // Poses are provided relative to the safe bounds configured by the user
	TrackingUniverseRawAndUncalibrated = 2, // Poses are provided in the coordinate system defined by the driver. It has Y up and is unified for devices of the same driver. You usually don't want this one.
	};

but that's easy as API that uses it takes a parameter for it.

[ahtur5]

That indeed could be the local coordinate system of the HMD, wrt to it's physical shape basically. But what about the controllers ? What's the "front" and the "up" of a controller ?

[echuber2]

If you study hellovr I think you can work out whatever you need about this. The tracking is relative to the room floor center as origin for standing (where the room is defined by your tracking space), and for seated mode it's calibrated with the head at the origin. (I'm recalling these details from memory as it was a year or two ago... something may have changed since then.)