Dodgy elevation on Spartan Sport WHR

isazi

@RiphRaph please search the forum, as there is a long discussion on this issue. Short version, the non baro watches have higher thresholds for computing elevation/descent than baro watches.

freeheeler

@isazi
e. g. this contribution from our friend, the former user…

"The actual algorithm is far more complex than that. The barometer resolution is 1m or probably even higher on modern Suunto watches, but you can try going up and down 2-3 meters and your watch will count zero total ascent even if you repeat that 100 times.

There’s a threshold, which is 3 meters for baro watches and 7 meters for watches without baro, and the total ascent / descent increment in chunks once the altitude change exceeds that threshold.

As an easy experiment you can customize a sport mode to have altitude and total ascent together on the same screen and just watch the two fields together. You’ll see that the total ascent lags behind altitude changes and in many cases it may not register the last few meters or a climb as you crest over a hill even if those few meters exceed the threshold."

https://forum.suunto.com/topic/3296/bug-of-positive-elevation-alti/56

RiphRaph

@TELE-HO Thanks so much for linking to this. So I conclude that my watch is behaving “correctly”, that it’s the 7m threshold that’s throwing off my elevation figures, and that there’s nothing that can be done about it except switch to a different watch.

Rather disappointing that the Tomtom could do it so much more accurately for about a third of the price, but there it is. One step closer to becoming a Former User myself.

freeheeler

@RiphRaph
I would observe the different trails you’re running and compare the total ascent to what you expect or measure in an external tool. Maybe the offset isn’t that big on other trails and you can happily live with what you have instead of investing in another watch.
It is difficult to figure out what the different watches do in the background, even within one watch maker. I think also Ambit3R and Spartan Sport have different ways to calculate total ascent.
But yes, as said, if you value a good precision of total ascent I would consider upgrading to a baro watch.

Side comment: A3PS and S9B, both baro watches do record different total ascents on rolling terrain… I did not compare recently as I gave my A3PS to a good friend… but back when I did comparison the S9B did calculate a littlebit too much.

Dimitrios Kanellopoulos

@RiphRaph if your watch didn’t have that filter it would produce unrealistic ascent’s.

Could I have a comparison of your prev tomrom with the Spartan to understand ?

isazi

@RiphRaph different watches have different algorithms, and you’ll see people complain a lot about elevation on other vendor’s forums too.

RiphRaph

@TELE-HO The offset is about as big on other routes as well. Just checked my other regular loop and the Suunto claims about 25% of the ascent that the Tomtom reported. I guess I will be looking out for a baro watch, because this is just nuts.

RiphRaph

@Dimitrios-Kanellopoulos It produces unrealistic ascents with the filter as well. QED.

Do you mean you want to see actual gpx files? I can provide those, but not via links in a post here as I don’t want to disclose location info on a public forum. Can I send them to you privately on this forum? I’m not yet very conversant with how things work around here.

Dimitrios Kanellopoulos

@RiphRaph said in Dodgy elevation on Spartan Sport WHR:
Send me a chat message please

You can upload them to dropbox / drive and pass me the link

RiphRaph

@Dimitrios-Kanellopoulos OK, just did that.

Dimitrios Kanellopoulos

roger that

jsr184

@Dimitrios-Kanellopoulos even with baro it looks weird when I see my data in QS.

Elevation gain/loss is over 600 meters:

But the watch says:

Is Ascent/Descent same here as total elevation gain/loss. My understanding is that the Ascent/Descent are totals from whole activity:

The altitude uses baro only here (no GPS, it’s indoor, it’s always either going up or down without GPS).

@RiphRaph I think my issue may be related/similar but if not then sorry, I didn’t mean to hijack your thread

Dimitrios Kanellopoulos

@jsr184 QS uses the minimal 1m calculation for charts. You can edit this on settings.

Your hand moves more than 1m FYI its logical. That is the case why all companies use filters. See my reply above

RiphRaph

@Dimitrios-Kanellopoulos “the diff is so small”?

On the 8 km route, Tomtom reports 68 m of ascent and the Spartan reports 22.
On the 10 km route, Tomtom reports 86 m and Spartan reports 14.

In what world is this a “small” difference?

Dimitrios Kanellopoulos

@RiphRaph I am sorry. I read 44 vs 52 as seen in the pics.

jsr184

@Dimitrios-Kanellopoulos is 1m means meter or minute? Don’t get angry please, but I don’t think I understood what you’ve been trying to say The thing is that if I do a climb of 9-10 meters up and then I’m being belayed down it’s usually 2-3 minutes of staying pretty much in the same location in 2D floor space. Hence, what QS shows looks more real. I don’t believe that my total ascent registered by Spartan was just little over 100m while I spent in the climbing centre over 8 hours doing just that - climbing (average route 7-10m high). Is it possible that my higher climbs are filtered by Suunto because it assumes no one suddenly can shot up in the air so far in such a small area? 622m vs 111m is even greater difference.

freeheeler

@Dimitrios-Kanellopoulos
all the tiny spikes in @jsr184 's recording could also be sealed sensor holes and pumping effect…

Dimitrios Kanellopoulos

@RiphRaph and yes I could say it’s small 30m of a miss. But that is personal and debatable.

But to be fair the tom tom has an external antenna and it’s accuracy is well comperable to an Ambit according to many AFAIK

If you are really that concerned about ascent in these small scales you should go for a baro device.

Again and honestly from an ascent freak, anything under 100m up and down “is flat” but that is very “bad” from me to say.

freeheeler

@Dimitrios-Kanellopoulos said in Dodgy elevation on Spartan Sport WHR:

anything under 100m up and down “is flat”

yes it is

Dimitrios Kanellopoulos

@jsr184 1meter, I am not angry just not putting emoticons. lol.

I am just replying fast.

If you want QS to show you a better ascent I would go to settings and put that filter to 3-4m.

4m is the ideal for all devices. That is not something from the top of my head. More news will come later as I am collaborating with some people and we train neural networks (for QS) and we are trying to publish those results, provide you this clevernes and GAP (on QS).