Re: Huge difference in elevation gain numbers, Garmin 800 versus GC

[G Coco]

Hi,

Can you upload the original FIT file to look at please?

Thanks

On Sunday, August 5, 2012 11:33:06 PM UTC-4, Charles Seely wrote:

Love GC.  However, I just did a ride and uploaded to GC. The average speed, cadence, heart rate, maximum-speed, cadence, heart rate and power etc all match. However, there is a huge difference between elevation gained. Garmin shows 5092 feet climbed and GC shows 4373 elevation gain.  Ride is a well known one in the Bay Area and i uploaded it to rava where it shows the same as Garmin and all segments are well known, so its easy to check. Ive seen differences before but nothing life this.  any advice?

[Charles Seely]

here you go. thanks

[Robert Chung]

Could it be the hysteresis constant?

[Mark Liversedge]

Almost certainly. It is currently 3.0 meters, I must confess I think that is quite conservative, but its been that way since Christmas Day 2009, so we might want to think hard before changing it. 

Here is the code:

        foreach (const RideFilePoint *point, ride->dataPoints()) {

            if (first) {

                first = false;

                prevalt = point->alt;

            }   

            else if (point->alt > prevalt + hysteresis) {

                elegain += point->alt - prevalt;

                prevalt = point->alt;

            }   

            else if (point->alt < prevalt - hysteresis) {

                prevalt = point->alt;

            }   

        }   

Regards,

Mark

[tetonrider]

Just as another data point, I happened to do a ride this weekend where I was carrying both a PC7 and Edge 500, and I'm a long-time user of both WKO+ and GC. 

I find that elevation data in WKO+ and GC tends to match up (when data is from the same device). As an example, I compared the PC7 data in both WKO+ and GC, as well as the Edge 500 data in WKO+ & GC. The discrepancy for the same data file in the 2 programs is on the order of 0.7% (~50' discrepancy over ~7000' elevation gain). The files from the PC7 and 500 report different elevation gains, but within 5% (7000 vs 7300).

I've found values reported on the Garmin device itself to generally overstate elevation gain (Garmin Connect would match this), perhaps suggesting that the trigger for an actual elevation "gain" is lower. I no longer pay attention to reported gain on a Garmin (or use Garmin Connect) for that reason. I have learned not to pay much attention to either.

-eric

[Robert Chung]

On Wednesday, August 8, 2012 3:09:07 PM UTC+2, Mark Liversedge wrote:

Almost certainly. It is currently 3.0 meters, I must confess I think that is quite conservative, but its been that way since Christmas Day 2009, so we might want to think hard before changing it. 

From a historical perspective, the first commercially available cycling altimeters were barometric and they could jump around a bit. Because of that it sort of made sense to have a hysteresis setting -- but the setting was empirically chosen for the idiosyncrasies of that kind of device. I don't know enough about GPS or GPS+barometric altimeters to know whether the same hysteresis setting ought to apply. Perhaps rather than be a global constant ...

[sean hunt]

I have noticed a large difference between elevation gain on Golden Cheetah, Garmin Connect, map my ride, Strava and RideWithGPS.

Was the Elevation hysterisis ever changed?

Could it be made adjustable?

[Mark Liversedge]

It wasn't, it could. We could add it as a special metadata field.