This is as much a note to myself as anything else. This has been nagging at me for years so now that I got around to finding an answer, I shall record it here lest I need to look it up again.
For a long time, I was finding that the distance of a track recorded on my GPS (a Garmin GPSMap 64s) would come out significantly lower when downloaded onto my computer and opened in Memory-Map. My theory was that this was something to do with how Memory-Map opens the GPX file, maybe losing detail that the GOS unit recorded. As an example, on my recent Tutuwai tramp, my GPS showed 16.2 km when I arrived at the hut. Imported into Memory Map, the distance became 14.5km.
A track, in the form of a GPX file is essentially a list of recorded waypoints. The GPS records a point at certain intervals, stringing them together for a track. The GPX file can them be downloaded and imported into various bits of mapping software. The recording interval can be set on most GPS units - Garmin offer the options of manually specified time intervals (e.g. every 5 seconds), distance intervals (e.g. every 50 m), and an Auto mode with options to record more or less data. These option determine how many points will be in the GPX file. Most software will calculate the total distance by drawing straight lines between the points.
Eventually I determined that no information was being lost when I imported the GPX file, so the discrepancies I was getting must be due to something the GPS was doing. Today I stumbled upon what that was: the GPS is doing something along the lines of recording the same trip in two ways. The GPX file is recorded at the specified interval, and this is the distance that shows up when the GPX file is imported. The GPS is also recording the trip at 1 second intervals and this is the distance that shows up on the trip odometer and when reviewing the trip file on the GPS itself. More data points = more detail = longer distances. Fewer data points in the GPX = less detail = shorter distances.
Why the GPS unit operates like this I do not know. Whether one method is more objectively accurate is a mystery to me. That being said, 1s is very very frequent, to the point where it may be capturing a lot of positional jitter from the GPS, potentially over-estimating distance. And setting the GPS to record points at a lower frequency can under-estimate distance by smoothing out small meanderings etc.
A few forums I read seemed to like using the "Most often" option of the "Auto" track recording setting for Garmin units, so I'll suppose I'll do that from now on. And if nothing else, maybe someone else will stumble upon this and answer there own questions about GPX anomalies.
No comments:
Post a Comment