Quaternion rotation nodes have inconsistent application of Euler angles

[tobyk]

Issue

Both ‘Rotation’ [3D.Quaternion] and ‘Rotation (Vector3)’ [3D.Quaternion] claim to process euler angles by yaw x pitch x roll (check the tooltip).

However they give different results. Seems like one of them is wrong.

QtoEulerCheck.vl (27.5 KB)

Related

Rotation (Vector3) Quaternion switches axis?

Quaternions are confusing, and more so once if you are trying to understand resulting angles by passing the Euler angles. The rotation vector 3 node needs the order of the inputs rearranges from that of the Quaternion from Euler angles A picture… thousand words: [callmenames-2021-07-21-Application_2021.07.21-23.28.03] I think it means that the axis has been flipped by a right angle along the (x axis?). Does the Quaternion.Rotation(Vector3) need to have its pin relabeled

[motzi]

the nodes apply the transformations differently, as can be seen in the code:

public static void RotationVector(Vector3 pitchYawRoll, out Quaternion result)
{
    Stride.Core.Mathematics.Quaternion.RotationYawPitchRoll(pitchYawRoll.X * Float32Extensions.cyclesToRadians,
        pitchYawRoll.Y * Float32Extensions.cyclesToRadians,
        pitchYawRoll.Z * Float32Extensions.cyclesToRadians, out result);
}

this is probably wrong, as the Vector3 pitchYawRoll is input in the function called RotationYawPitchRoll in the wrong order (x,y,z instead of y,x,z)

[motzi]

i created an issue on github

[Elias]

Thanks for the report. Will be fixed in upcoming preview (>= 5.3-167).