FEATURE REQUEST: How many of YOU would like to see continuous Alti-Autocalibration?

maybe they can put the few lines of code from an old fenix 3 firmware back in again.

When I bought my F3, this functionality had already been removed. But I know from early owners that it was unusable. If you want autocalibration during an activity to work, it must be implemented intelligently. It was not in the F3.

AllanOlesen67 thanks for the Info! before i was thinking for getting a used F3 and put an old firmware on it.

i don't say that gps-elevation is best point. i would not by a watch without barometric altimeter. but the starting calibration is mostly wrong, when you start your activity for example in a narrow valley. and every watch of my friends catches up and recalibrates when getting higher with better gps signals. except the fenix. on a hike it is not a problem, but for example when trailrunning i hate it take time for recalibrating.
and i don't want it, as a fixed feature. but the option to turn it on and off would be great.
so for example Autocalibration options:
OFF
on start
continuously

Hi,

What I'm trying to say is that your friends watches should have something else such a better algorithm .
Better gps signals + continuous calibration alone "could" lead to an even bigger error. Depending on the sampling of your "desired" continuous calibration "would be like" you didn't even needed a barometer...
Fact is : GPS altitude per se is the worst metric possible therefore it should be used very wisely for calibration, otherwise it throw away a barometer given precision.
For climbing, altitude is key so you would need a very "absolute" precise measure all the time.
But for purposes such as calculating a "total ascent/descent" (sports in general) we needed only a "relative" precise measure because we would looking to calculate altitudes differentials. With no cold fronts we even didn't need any calibration at all. But even in this Garmin is failing so I'm not defending them at all.

Regards,

Hi,

What I'm trying to say is that your friends watches should have something else such a better algorithm .
Better gps signals + continuous calibration alone "could" lead to an even bigger error. Depending on the sampling of your "desired" continuous calibration "would be like" you didn't even needed a barometer...
Fact is : GPS altitude per se is the worst metric possible therefore it should be used very wisely for calibration, otherwise it throw away a barometer given precision.
For climbing, altitude is key so you would need a very "absolute" precise measure all the time.
But for purposes such as calculating a "total ascent/descent" (sports in general) we needed only a "relative" precise measure because we would looking to calculate altitudes differentials. With no cold fronts we even didn't need any calibration at all. But even in this Garmin is failing so I'm not defending them at all.

Regards,

marcossaad_ you're absolutely right. i totally agree with every point you said.
like you said, for example doing a normal run, i don't really care if the absolute altitude is correct or not. the ascents are counting.
doing lets say mountaineering, i have 2 data fields running. elevation and GPS elevation. and when getting out of a narrow valley on a "mountain crest" (hope thats the right english word for it),
the GPS elevation is awesome exactly. because there are no reflection possibilities and good a good view in every direction to the sky.

maybe i am completely wrong but i thought, if even the oldest garmin GPS devices (talking about the years around < 2000) could give you a tolerance value for vertical accuracy (+/- 10 meters for example) i could not be so hard, to code an algorithm something like: (example)

- how is the accuracy of the GPS vertical elevation -> +/- 50 m
-> do nothing

- how is the accuracy of the GPS vertical elevation -> +/- 25 m
- whats my gps altitude -> 1550 m
- whats my barometric altitude -> 1530 m
- is it in in tolerance -> ( > 1525 m, < 1575 m) -> yes
-> do nothing

- how is the accuracy of the GPS vertical elevation -> +/- 10 m
- whats my gps altitude -> 1550 m
- whats my barometric altitude -> 1530 m
- is it in in tolerance -> ( > 1540 m, < 1560 m) -> no
-> recalibrate it to the GPS value or the middle of difference -> barometric value jump to 1540m -> correct barometric pressure to sea level.

and so on and so on. suunto for example does nothing else. even when turning fused altitude on, when the accuracy tolerance is not good, they do nothing.

Another point would be, you would get Storm alerts when being outside and doing activities. the only points when i get an storm alert is when being stationary.

and for note: i don't would like to see this as default. but the option for it: Autocalibration
OFF - if you would like to do it manually
AT START - if you just want to have it, as it is.
CONTINUOUSLY - if the absolute elevation is important to you.

(hope you can understand my not so good english) :-)

Mistamb, your algorithm would be a disaster. It would cause jerky corrections of altitude, leading to a strange altitude graph for the full session. This was probably the algorithm, which Garmin used in the Fenix 3, until they ditched it.

If you want to do it right, you need to look at the long term deviation between at GPS and barometric altitude and do more smooth corrections along the way.

AllanOlesen67 i know what you mean. but let's say, it will do this on start, as long as it has a quite good altitude. then take a brake for lets say 15 minutes or even better 30 minutes. it would not really effect battery consumption and yes, is would give you a little jump up or down, every now or then.
if the weather not change, there is nothing to correct. but if so, yes you would see a jump. as the suunto ambit series does, or the polar vantage v. (it would not be as smooth as the suunto spartan or 9 series,) but it would be the more correct altitude.
lets take this an example:
there is correction in air pressure with 7 hPa which equals around 56 meters of elevation along the way. the peak was of around 3 meters. where my watch was around 70 meters off at the peak.
and lets say there are days you don't wanna take care and manually recalibrate your watch every now and then and just do sports or enjoy the mountains.
ciq.forums.garmin.com/.../1467196.jpg

Implemented correctly, so it doesn't result is weird stepped corrections like allanolesen67 says, it would be a great feature... I've just done a 55k trail run along the coastal path, calibrated altimeter via gps pre start, and at one point my Garmin said I was -18m below sea-level, which was incorrect. I've never trusted the altimeter accuracy during a run, but use it as a ball park figure...

I've enjoyed reading the thread so far and I'd like to add a few comments to the discussion:

1. GPS altitude is not usually very good. I have had numerous GPS watches that do not have a barometric altimeter and the total ascent has always been >10% off. Usually massively off of you have numerous ascents/descents.

2. A good way to see how the GPS altimeter is working is to add the data field for it next to the altitude field on a data screen. Mine is almost always >5 metres off.

3. GPS altitude lags hugely. I do a lot of hill running and I know the altitude of my local peaks . To get an "accurate" GPS altitude in a peak I have to stop for at least 2+3 minutes.

4. The Fenix implementation of altitude is strange when outside an activity and this could be easily improved. For example, I wore my watch to bed last night. It obviously knows I'm sleeping as it's in the sleep period but my altitude changed 3 metres during to a slight change in barometric pressure. They need to improve this locking. I think it does do some of this but it's not 100% perfect.

5. The 910XT had a feature where altitude could be calibrated automatically from a saved location. So, for example, I have a saved location for "home" which has a defined altitude of 29 metres. When the watch had a GPS fix close to this location it would automatically calibrate the altimeter. I also had this set of a couple of local high peaks in the area and that dealt with any shift in pressure during a run. I have requested that this could be added a number of time but alas nothing has happened (they did implement my request for a "lap" option on the paused activity screen so they do listen!)

6. Finally, DEM data. This is really under utilised. The F5 series has DEM data in the maps and this could be used to calibrate the altimeter regularly when you have a GPS fix. This is hugely preferable to using a GPS altitude.

Anyway, all good stuff...hope some sort of improvements will be made on the 5x but I think I'll probably buy the 945/Fenix 6 and hope those firmwares will be improved.

Cheers,

D.

Regarding 5.
Actually there is auto cal to saved locations. If you change alti before start of an activity, this is stored somewhere. I don't think it's possible to see saved locations anywhere. I think that was the conclusion in earlier discussions.

If you have auto cal to on, there is an automatic calibration with gps during pauses in the same way as before start.
This calibration has been so bad that I've not been using autocal in a long time ( can have been changed in later sw ), even though I really would like to use it for saved locations.
I got jumps for 100 m to totally wrong values.
I wonder if it calibrats to saved locations during pauses also? Never tested