12 replies to this topic Post Order

[pauloc]

Hey Guys,

So with all these different tools to track your ride available, I've noticed huge differences in the ascent reported. We've always used my Polar as base but now a few guys are using iPhone apps and Garmins and all of the calcs are way out.

For example, my Polar will report 650m of ascent while Strava reports 1200m??

Cheers,
Paulo

[Just Keep Pedaling]

I won't pretend to know why.

It might be something to do with GPS receivers only getting info from satelites at long intervals to save battery - but that is just a guess.

[Zac.A]

On an MTB ride my garmin read 1500m but mapmyride read 850m...

I don't know but I think the online and phone apps use the roads and maybe not mtb trails?
have no clue other than that and what JKP said.

[eccentric1]

Polar used barometric pressure. GPS units should be accurate as they use triangulation.

[Gunzo]

My Nokia E72 indicates the accuracy of the readings.

This indicates that when the coordinates accuracy is within 20m, the height is still @ 50m. So I don't believe GPS' is as accurate as we like to believe. At least not a cellphones internal GPS. If it's a dedicated GPS with a decent antenna, different story.

Another issue: If the phone is in your back pocket, do you know what the GPS signal strength is? Optimally I think your phone should be open pointing towards the sky.

[Kranswurm]

I use a Sigma computer which uses barometric pressure.
It is accurate.

[Minion]

It's been discussed a number of times before. It's mainly to do with the algorithm that the device or software uses to smooth the data, not so much the accuracy of the sensor.

As an example, using the same data file from my Garmin, for the same ride, I can get ascent values varying from 500m to 1000m, depending on which software I load it into and which settings I use.

[pauloc]

Edman, on 09 January 2012 - 10:11 , said:

It's been discussed a number of times before. It's mainly to do with the algorithm that the device or software uses to smooth the data, not so much the accuracy of the sensor.

As an example, using the same data file from my Garmin, for the same ride, I can get ascent values varying from 500m to 1000m, depending on which software I load it into and which settings I use.

Interesting... So how do we figure out which program is most accurate then?

[Minion]

pauloc, on 09 January 2012 - 10:20 , said:

Interesting... So how do we figure out which program is most accurate then?

I suppose it's all a compromise - you don't want noisy data to to mess up the ascent values, but you also don't want to lose actual ascents in the smoothing.

It also depends on what you'd call an ascent. If you drive up a 10m elevation, is that a 10m ascent? What about a 5m elevation or a 1m elevation? What happens if it takes you 100m to go up the 10m or if it takes 1000m?


I'm not too concerned about total ascent, but if you use a program like Sport Tracks then you can change the level of smoothing until you get ascent values that you think make sense.

[Kona Kula]

A barometric altimeter will always be more accurate. The gps units are the ones with the issues due to the algorithms. This can be hardly noticeable or way out.

[Minion]

Kona Kula, on 09 January 2012 - 11:26 , said:

A barometric altimeter will always be more accurate. The gps units are the ones with the issues due to the algorithms. This can be hardly noticeable or way out.

They're different sources of error, though. A lot of the time, the difference (it's not really an error) is due to the smoothing used. The Garmin data I used for comparison came from a unit with a barometric altimeter.

Here's a good article on GPS elevation problems:
http://www.dcrainmak...device-gps.html

I would summarise by saying that if you see different ascent values, depending on the program used to display them, then it's due to smoothing errors. If you see wildly different ascent and descent values in the same program for a ride starting and stopping in the same place, then it's probably an accuracy issue.

[eccentric1]

Kona Kula, on 09 January 2012 - 11:26 , said:

A barometric altimeter will always be more accurate. The gps units are the ones with the issues due to the algorithms. This can be hardly noticeable or way out.

I had my Garmin ETrex GPS unit on when I flew into Beira Airport the very first time, I had Aircrew permision. I calibrated the altimeter with a 1:50 000 map before I left. It was cloudy over Beira. We were approaching sea level very fast and I died a hundred deaths. We broke through cloud and were still quite high. Could never understand what happened there.

[gogo@]

According to Endomondo my house is moving in the vertical... my total ascent versus descent seldom matches! I wouldn't trust cellphone gps systems like these for comparative purposes between different units. Even with with altimeters using barometric pressure for their measurements there are calibration issues. Still, it's fun, and reasonable to compare your own rides if you accept there are limitations. I'd be curious to know though how close some of the phone apps get and which are most accurate.