I've always been surprised by the complete lack of information on How Wide the drone's camera actually captures in different shooting modes - DJI releases the specifications of the lens (in the case of the Avata 2 FOV 155⁰ or 12mm) and that's it - ever since the Mavic 2 I was disappointed by how much tighter the shot actually was.
So I decided since there was no information, to test it myself ;)
Deeply SCIENTIFICALLY , I put the Avatar 2 at  40 cm from the wardrobe - I stretched the tape measure supposedly in the middle of the frame, it turned out that everything was not visible in the goggles either, and therefore in the first frame, the beginning of the roulette is in the frame, I extrapolated a little ;)
So, I shoot in 4K 60 frames per second 16:9 (D-Log shouldn't matter, does anyone know ;) )
And I started - UltraWide without stabilization, UltraWide with stabilization, then Wide without and with, and finally Normal without, with, and with horizonSteady, After which I started  remembering geometry to calculate the angle (FOV - Field of View) and I measured it horizontally and diagonally (because I don't know what the spec is, I guess diagonal at 4:3?)
So here are the results

That's great, can you do the same for 4:3? I've always wanted to, but have been too lazy

(And yeah, sensor is 4:3 native so the biggest fov they mention in specs should be ultra wide without stabilisation at 4:3.)

fichek Posted at 4-26 06:17
That's great, can you do the same for 4:3? I've always wanted to, but have been too lazy

(And yeah, sensor is 4:3 native so the biggest fov they mention in specs should be ultra wide without stabilisation at 4:3.)

I tried 4:3 and did not like it so no go for this But it's easy to calculate as 4:3 keeps horizontal the same (what I measured but just add vertical pixels) so if I am not mistaken diagonal FOV of 4:3 UW no s should be

137.36

An interesting question coming out of this is how on earth they film in 4K (8MP) in Normal HorizonSteady which is using under 1/4-th of a matrix according to my calculations 2.3 MP???

Well now I know why there was no such information - nobody besides me care

alex_markov Posted at 4-26 06:49
An interesting question coming out of this is how on earth they film in 4K (8MP) in Normal HorizonSteady which is using under 1/4-th of a matrix according to my calculations 2.3 MP???

It's a 4000x3000px (12MP) sensor so there's a bit of wiggle room, but yeah 4K HorizonSteady probably must stretch pixels a lot.

alex_markov Posted at 4-26 06:38
I tried 4:3 and did not like it so no go for this  But it's easy to calculate as 4:3 keeps horizontal the same (what I measured but just add vertical pixels) so if I am not mistaken diagonal FOV of 4:3 UW no s should be

I'm not sure that's exactly how it works since the lens isn't anamorphic and the wider you go, the more it bends so the more diagonal fov stretches - that's probably how they get to 155.

fichek Posted at 4-26 22:52
I'm not sure that's exactly how it works since the lens isn't anamorphic and the wider you go, the more it bends so the more diagonal fov stretches - that's probably how they get to 155.

I assume they reported the Lense FOV but the lens matrix recording is a but less

Good job Alex. Very interesting to see all the variations.

fichek Posted at 4-26 06:17
That's great, can you do the same for 4:3? I've always wanted to, but have been too lazy

(And yeah, sensor is 4:3 native so the biggest fov they mention in specs should be ultra wide without stabilisation at 4:3.)

Yeah nicely done
Thats why i always record in 4:3 and stabilize in post.
I wish there was an option to correct the lens distortion while flying. Especially at 4:3. Or I just have to get used to the distortion^^.