How does elevation smoothing best work?

[richardpbull]

Could you explain a little how this elevation smoothing works please? Looking into it a little it is a definitely a mistake to rely on single values. Often elevation can vary by 10m over a distance of 2m simply because of the relative accuracy of location over spot altitude in GPS devices.

Do you have advice for reasonable settings for the thresholds?
many thanks!

`
/**
* Apply elevation gain/loss smoothing? If true, the threshold in
* ELEVATION_SMOOTHING_THRESHOLD and ELEVATION_SMOOTHING_SPIKES_THRESHOLD (if not null) applies
* @var bool
*/
public static $APPLY_ELEVATION_SMOOTHING = true;

/**
 * if APPLY_ELEVATION_SMOOTHING is true
 * the minimum elevation difference between considered points in meters
 * @var int
 */
public static $ELEVATION_SMOOTHING_THRESHOLD = 2;

/**
 * if APPLY_ELEVATION_SMOOTHING is true
 * the maximum elevation difference between considered points in meters
 * @var int|null
 */
public static $ELEVATION_SMOOTHING_SPIKES_THRESHOLD = 50;

`

[richardpbull]

This provides a good visual example of smoothing noise from tracks' x or y points...
https://www.gpsvisualizer.com/tutorials/track_filters.html

For APPLY_ELEVATION_SMOOTHING, ELEVATION_SMOOTHING_THRESHOLD & ELEVATION_SMOOTHING_SPIKES_THRESHOLD

Would increasing ELEVATION_SMOOTHING_THRESHOLD might eventually result in points being ignored and the cumulative distance then decreasing? Or is it only the elevation that is not included in the cumulative elevation change summary?

[richardpbull]

Out of interest this is what happens when I increase the ELEVATION_SMOOTHING_THRESHOLD value from 0-15 on a GPX file from a 103 km mountain track.

I guess this is much easier for cycling (on road or even off road) as you know that the track is not going to exceed a certain gradient, whereas hiking is is easy imagine plenty of lumps in the trail.

For me we use Scarf's rule to estimate time on the trail, which is:

equivalent distance = x + α·y where α=7.92, x is horizontal distance, y = +tive vertical increase. Then we estimate a flat running time like 12 km/h for elite runners.