A barometer in an altitude costume.
An altimeter does not measure altitude. It measures static pressure — the same line described in the pitot-static article — and converts it to metres using a fiction called the International Standard Atmosphere: sea level at 1013.25 hPa and 15 °C, cooling 6.5 °C per kilometre. Feed the ISA a pressure and it returns an altitude. Feed it the real atmosphere, which has never once matched the standard, and it returns a useful lie.
The raw conversion — pressure straight through the ISA with no correction — is called pressure altitude. It is wrong as geography, but it has a priceless property: two aircraft at the same pressure altitude are at the same physical level, whatever the weather is doing. Aviation runs on that property.
The knob: QNH, QFE, QNE.
The setting knob slides the whole scale to anchor the fiction somewhere useful. Set QNH — the local pressure reduced to sea level — and the altimeter reads altitude above mean sea level: airport elevations and mountain heights now mean something. Set QFE, the field pressure itself, and it reads height above that airfield: zero on the runway, a habit of winch sites and old training fleets. Set 1013.25 — the QNE convention — and you are back to pressure altitude, renamed a flight level: FL95 is 9500 ft of pure pressure, the common currency of everyone above the transition altitude.
One hectopascal on the knob moves the reading about eight metres. Mis-set QNH by 10 hPa — yesterday's value, or the airfield across the ridge — and every altitude on the panel is 80 metres of quiet fiction. The knob is a flight control; treat it like one.
Four answers, one glider.
Drag the temperature to −20 °C: the altimeter still politely indicates 1500 m while the glider is actually 120 m lower. Cold air is denser, the pressure levels squeeze together, and the altimeter — which only counts pressure — congratulates you on altitude you do not have. Cold and high terrain is the one combination where this diagram is a safety brief, not trivia.
GPS measures a different planet.
A GNSS receiver knows nothing about air. It measures geometry — your position relative to satellite orbits — and its natural altitude is height above the WGS84 ellipsoid, a mathematical egg fitted to the Earth. Sea level does not follow the egg: the real equipotential surface, the geoid, undulates up to a hundred metres around it as gravity varies. Your receiver carries a geoid table and subtracts the local offset — around 46 m in Slovenia — to report height above mean sea level.
So GPS altitude is honest geography: no QNH to mis-set, no temperature error, just a few metres of geometric noise — the vertical solution is the receiver's weakest axis, as the GNSS article explains. On a standard day the altimeter and the GPS roughly agree. On a cold low-pressure day they can disagree by two hundred metres, and now you know which one is describing rock and which one is describing air.
What shows where.
Reading the panel with all of this in mind: the altimeter shows QNH altitude, and is the instrument the airspace map and the controller mean. The flight computer's final glide runs on the same barometric altitude — corrected by the QNH you set — because the arrival at the far field is an atmospheric event. The GPS altitude on the info page is the surveyor's truth, best used as a sanity check on the other two. And the IGC file writes down both, so the argument can be settled after landing.
One habit ties it together: set QNH before every launch, from a source on the field, and glance once at the GPS altitude on the runway — if the two disagree by more than the geoid's share, something on the static side deserves attention before the air gets thin.