I’m experienced using the Kinect V2 sensor with Unity. T-pose calibration was not needed. Since I am creating a standalone experience for a video wall I cannot expect users will be standing in a T-pose for calibration.
Is this necessary or optional? I am using the RealSense D435 sensor.
If necessary this is a showstopper and will need to research alternative solutions.
There is no need to t-pose type calibration using Nuitrack - BUT - you may find that the D435 is not optimal for skeletal tracking - it is really designed for close proximity details - as the further you get away from the d435 sensor the higher the level of errors in the depth data - this can result in very jerky and unreliable point data.