So.... When I was a young engineer, I wrote a specification for a pressure transmitter, using a specific model number that I thought would work. My more senior colleague George told me
"It won't work"
"Why not George? The range is correct!"
"Won't work"
"But the material of construction is consistent with application."
"Won't work"
After several more rounds of me giving George reasons why it should work, and him telling me it wouldn't, I finally said
"George, you're going to have to give me a reason rather than a denial"
"Jim, I could go into great depth on why it won't work. Suffice to say that I invented that type of transmitter, I hold the patent (several actually) on it, and I have my reasons for saying 'it won't work'"
"Uhhhhh. Ok. I'll find another transmitter"
(He did later explain)
In any case, pressure transmitters generally have an error spec that is a "percent reading plus percent full scale". So if you have a 1.5% of reading and 1.5% of full scale, at 95 of full scale you have about 2.1% (if my log mean calc is correct) or so (its more complex, but the approximation is good enough for an explanation anyway. If you get down a reading that is only 16% of full scale you have an uncertainty of about 1.6%, or about 10% of the reading. That's the situation at about 50,000 feet (15000m or so).
So the question depends upon just how high that you are measuring altitude. At some altitude, GPS may be more accurate. At
http://www.gpsinformation.net/main/... on this,1.5 x Horizontal error specification.
they say that "standard consumer GPS receivers should consider +/-23meters (75ft) with a DOP [Degree of Precision] of 1 for 95% confidence. " Barometric estimates seem to be about 25 feet at low altitude going up to 100 feet at higher altitude (but the altitudes weren't specfied - sigh). In any case, if you are lower than 15K feet, I would think barometric would be fine.
That said, I'm hoping to qualify for L1 next week and my Mad Cow Super DX3 (4") diameter, with a blue tube avbay will be motor deployment and will contain... an Eggtimer ION monofunction altimeter. So if someone has specific actual experience, it probably beats my general theory.
From an engineering standpoint, ideally you'd mix (fuse) your estimate from both signals, varying the barometric altitiude estimate less and the GPS more as your altitude increases. There's a paper on that
here.
Using a Kalman filter theoretically gives you the best estimate from either source, and you can also combine the inputs. Not sure that I've seen a filter that variably weights the two sources as the altitude changes.