ANT Forum

I'm making tools to process FIT-data from the VIRB Ultra 30, which has worked well so far (matching videos with FIT-files, extracting data corresponding to recording session etc).

Currently I'm trying to find the best way to get heading. Getting it from `gps_metadata` (id: 160) seems the easiest since I won't have to deal with three_d_calibration messages for the magnetometer, but I observe two odd behaviours:

1. Lots of outliers: large jumps for the heading values, despite the camera wearer standing still. Could possibly be filtered out with some arbitrary threshold. May simply be that e.g. the magnetometer is acting up or is affected by a stronger magnetic field in the vicinity.

2. Static magnetometer value if speed is zero, until the camera wearer starts to move again, e.g. (raw output below from a simple cli-tool I made SPE = speed, HDG = heading):



Point 2 still seems odd to me.

Does anyone know what the "heading" value in gps_metadata is derived from? I'm currently working on processing the magnetometer data directly for this reason (I still need to sync this to the gps track).      

Further updates. As soon as speed is zero heading becomes static until the camera starts moving again. Does this mean heading in "gps_metadata" is derived from gyro/acceleromoter, rather than magnetometer? (all three?)

Is it perhaps better to ask Garmin directly?

Example (unprocessed fit data, note points marked by arrows):