[nativeforks]

Thanks! Currently, MBCompass can show both magnetic north using Android’s sensor fusion and true north (based on WGS84 geodetic coordinates).

Adding a magnetic compensation map sounds like a great fit for improving global accuracy without changing the app’s core goals. Thanks for the suggestion.

[jacquesm]

YW, there are some pretty compact representations possible of that data but it will come at a considerable expense in computational overhead.

https://www.ncei.noaa.gov/products/earth-magnetic-model-anom...

Is a good starting point.

[xavieralexandre]

You are linking to the magnetic anomaly grid, which is primarily intended for geophysical research and modeling local variations.

For a basic compass app, you can rely on the World Magnetic Model (WMM) instead: https://www.ncei.noaa.gov/products/world-magnetic-model

From that, you can pre-compute a low-resolution declination grid. NOAA even provides one here: https://www.ngdc.noaa.gov/geomag/calculators/magcalc.shtml#i.... That’s only a few KB of data and requires just a simple declination subtraction based on lat/lon, similar to how it’s handled on nautical charts. This works fine as long as you stay away from the poles (and Alaska).

[jacquesm]

Near ore deposits the map resolution needs to go up considerable beyond the 'few KB of data' to keep it working. Nautical charts are simpler because there is as a rule a bit more distance between the compass and any anomalies but in certain mineral rich areas of the world that definitely is not the case.

[xavieralexandre]

Thanks. I was mostly basing my answer on my experience with nautical charts, where declination corrections are typically uniform over relatively large areas and local anomalies are less of a concern. I hadn’t considered how much more variable the field can be in terrestrial use.