Absolute altitude

[sar104]

Then it's a true mystery if they were both at -42 m.

But they weren't, were they? According to the OP:

Relative Altitude : +13.50
Absolute Altitude : -42.00

That was the reason for his question; he took off from ~ 0 m MSL, and at a barometric altitude of 13.5 m he found - 42 m MSL from GPS in the metadata.

 

[AlanTheBeast]

But they weren't, were they? According to the OP:

Relative Altitude : +13.50
Absolute Altitude : -42.00

That was the reason for his question; he took off from ~ 0 m MSL, and at a barometric altitude of 13.5 m he found - 42 m MSL from GPS in the metadata.

Ah - okay - I misread. Then what he sees may make sense. Ellipsoid error + GPS error could go that far down. 55 m seems a lot to me - however. I wonder what else is going on that we don't know about.

 

[sar104]

Ah - okay - I misread. Then what he sees may make sense. Ellipsoid error + GPS error could go that far down. 55 m seems a lot to me - however. I wonder what else is going on that we don't know about.

I agree that it seems to be on the high side of the expected error, but you can also add some tidal variation from MSL as well.

 

[marciano]

Thanks all for providing interesting information on this matter.

It would be more useful to store in EXIF data sea level altitude which can be taken from some maps like Google's. I use FPVCamera. Each waypoint I set displays ASL, AGL and height from Home Point (Relative Altitude).
Here's an example.

 

[AlanTheBeast]

I agree that it seems to be on the high side of the expected error, but you can also add some tidal variation from MSL as well.

Did he takeoff near the water line at low tide?

 

[sar104]

Did he take near the water line at low tide?

Don't know, but that was what I was wondering.

 

[Lescenatt]

No you c

I was flying over a beach. Looking at P4P metadata i found Absolute Altitude = 42m below sea level, so it seems I was diving
Relative altitude refered to home point altitude was okay
Is it possible to calibrate barometer/GPS to get more accurate data?
Compass is already calibrated and GO4

Got from metadata:
GPS Altitude : 42 m Below Sea Level
Relative Altitude : +13.50
Absolute Altitude : -42.00

GPS and barometer match the same altitude? Mmmmmm

No you canT set nothing.
Only the barometer work with altitude.
The Gps do NOT give you the altitude.
Also the ground sensore give you help to rivelate the altitude...that may be a different of the barometer.

 

[sar104]

Thanks all for providing interesting information on this matter.

It would be more useful to store in EXIF data sea level altitude which can be taken from some maps like Google's. I use FPVCamera. Each waypoint I set displays ASL, AGL and height from Home Point (Relative Altitude).
Here's an example. View attachment 81193

Agreed, but the aircraft does not (currently) consult Google Earth or any other source of DEM, so that's not possible.

 

[marciano]

Did he takeoff near the water line at low tide?

Yes, from the DJI box over the sand, maybe 1 foot ASL.

 

[andy_k]

As silly as it may sound, just because you are standing by the water doesn't necessarily mean you are at 'sea level' - it's a worldwide calculation

 

[AlanTheBeast]

As silly as it may sound, just because you are standing by the water doesn't necessarily mean you are at 'sea level' - it's a worldwide calculation

MSL is a geoid, but the geoid is not a spheroid. The geoid is a lumpy (for very fine levels of lumpy) thing. If you're at the beach (on the ocean of course), then MSL is always between high and low tide.

If you can find a local measuring station, its -0- mark is sea level (MSL). This is measured over long cycles to establish the mean. (20 years or so). The MSL for that beach would be that point and that point should fit the geoid model quite closely (cm's).

If it's GPS, then all bets are off. It's accurate, but its "sea level" is its ellipsoid (WGS-84) whereas "MSL" for, ie your airport elevation, is the "geoid".

1-Mean Sea Level, GPS, and the Geoid

 

[Racklefratz]

GPS is horribly inaccurate for altitude and can swing +/- 200 feet or more in short time periods.

That's an incorrect statement, and you're mixing terms you don't understand.

Re-read the Garmin discussion you referenced. Nowhere does it mention the word "altitude". The discussion is about GPS accuracy of ELEVATIONS, not altitudes. Elevations refers to ground-level, as in the term "field elevation". Elevations and altitudes are two different things.

The dirty little secret is that GPS ALTITUDES are extremely accurate, so much so that, in real aviation, the FAA has been phasing out ground-based navigation facilities in favor of GPS navigation for decades, beginning with NDBs in the 90s.

Vertical positioning accuracy in instrument approaches requires extreme accuracy; a +/- 200 ft altitude error on one can kill you. GPS-based instrument approaches have been FAA-approved and published for decades. I personally owned an aircraft with a GPS installed which was legal for instrument approaches.

Don't confuse altitude with elevation. GPS altitudes are dead-on accurate as it gets.

 

[Racklefratz]

I'm not disagreeing with the accuracy of GPS altitude, although I find it rarely to be rarely more than 30 ft out. My point was rather that relative barometric altitude is pretty good over short timescales but, as atmospheric pressure changes with weather, the absolute barometric altitude varies systematically and substantially - easily by hundreds of feet.

Which is exactly why, in real-world aviation, pilots are required to obtain the current LOCAL barometer setting prior to attempting an instrument approach to the airport, and also why ATC provides the local altimeter setting as part of the initial contact with any pilot.

 

[sar104]

Does DI actually manufacture the camera or do they buy it from another company as a private label? Every point-n-shoot camera today has a GPS receiver inside it, and that's all the altitude information it has. Why would the camera that DI buys be different?

BTW, IFR approach plates for manned aircraft require a barometric altimeter. GPS altitude is not permitted for IFR flight because altitude is notoriously inaccurate.

These are not point and shoot cameras. I don't know whether DJI make the cameras themselves or have them made to their own design. Either way, I'm quite sure that there is no GPS receiver in the DJI cameras - those data are inserted by the aircraft from the onboard GPS.

 

[SteveMann]

The dirty little secret is that GPS ALTITUDES are extremely accurate, so much so that, in real aviation, the FAA has been phasing out ground-based navigation facilities in favor of GPS navigation for decades, beginning with NDBs in the 90s.

All IFR approaches require a barometric altimeter because GPS altitude is notoriously inaccurate. The error is a mathematical fact and no amount of fiddling of the words will change that. The FAA is slowly decommissioning the VOR system for position information; the VOR's never had any altitude information.

 

[sar104]

All IFR approaches require a barometric altimeter because GPS altitude is notoriously inaccurate. The error is a mathematical fact and no amount of fiddling of the words will change that. The FAA is slowly decommissioning the VOR system for position information; the VOR's never had any altitude information.

Here are some data from the FAA's most recent (January 2017) report on civil GPS performance, rather than throwing around qualitative adjectives. Consistent with (most) conventional wisdom on this system, vertical error is approximately twice horizontal error, with a 95% vertical error of less than 4 m.

That's for high-performance civilian GPS with unobstructed sky view, as is typical in avionics installations, but the factor of two should be fairly consistent for all modern receivers.

 

[AlanTheBeast]

All IFR approaches require a barometric altimeter because GPS altitude is notoriously inaccurate. The error is a mathematical fact and no amount of fiddling of the words will change that. The FAA is slowly decommissioning the VOR system for position information; the VOR's never had any altitude information.

All IFR approaches require baro alt regardless of other navigation equipment. Indeed baro alt is required, period to fly any manned aircraft. Baro alt is a primary source and often self-powered and thus very reliable as is its repeatability. (could go into the MEL's here for various approaches ...)

As to GPS "error" It's mainly because the datums are different. Baro is more accurate over the short term - but only if you know QNH (for the purpose of an approach). Some non-tower approaches use other airport's QNH and thus, the DH has to be higher in most cases.

CAT I WAAS based approaches are accurate enough for a 200' DH (requires ±3.8 m horiz and vertical); . But the approach has to be programmed correctly (in the RNAV system) to account for geoid and ellipsoid differences. Actual performance has been about about 5X better than requirements for most of N.A. (And you still need baro alt)

On open ground (or in open air) GPS altitude error will be about 2X the horizontal error with respect to the GPS ellipsoid. That may or may not match your map (or an airport elevation) if they are not on the ellipsoid but probably referenced to the geoid.

 

[AlanTheBeast]

Here are some data from the FAA's most recent (January 2017) report on civil GPS performance, rather than throwing around qualitative adjectives. Consistent with (most) conventional wisdom on this system, vertical error is approximately twice horizontal error, with a 95% vertical error of less than 4 m.

That's for high-performance civilian GPS with unobstructed sky view, as is typical in avionics installations, but the factor of two should be fairly consistent for all modern receivers.
View attachment 81293

Referenced to the WGS-84 ellipsoid. Sure. Referenced to a given non-precision or precision approach: not necessarily so.

 

[sar104]

Referenced to the WGS-84 ellipsoid. Sure. Referenced to a given non-precision or precision approach. Not necessarily so.

Oh agreed. I'm not arguing about precision approaches - that was not the subject of this thread, at least until recently - I was just addressing the question of random, as opposed to systematic, error in GPS altitude.

 

[AlanTheBeast]

Oh agreed. I'm not arguing about precision approaches - that was not the subject of this thread, at least until recently - I was just addressing the question of random, as opposed to systematic, error in GPS altitude.

Yeah - sorry about that - I was including @SteveMann in the discussion.