DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

Measured field of view for Mavic Air: 59° video, 69° photo

Prismatic

Well-Known Member
Joined
Feb 26, 2018
Messages
587
Reactions
644
Age
70
Location
Northern Colorado
Summary: The camera on the Mavic Air is advertised with an 85° field-of-view. A careful test shows something very different!
  • When shooting video, the measured FoV is about 59° (actually, more like 58°, see below).
  • When shooting still photos the measured FoV is 69°
---------------------------------------------------------------------------------
I was puzzled to find that shots I'd carefully composed offline (using Virtual Litchi Mission) were not framed as I'd expected in the field. It became apparent that the advertised 85° field-of-view is not accurate. So I ran an experiment to measure the true field of view for the Mavic Air. Some walking and simple trigonometry gave the answer. Was I surprised! ?

Free to wander the campus of a local high school, and using a 50' tape, I measured a long baseline segment of a pillared wall (see below). I made sure each end of the baseline was the outside edge of a pillar, to make framing easy. The baseline--red in the diagram--came out to 220'. From the center of that baseline, I measured a perpendicular line--yellow in the diagram--out to 200', with stakes at 100', 150', and 200'.

I started the MA (without props) so that I could see and capture images hand-held. I put it into video mode, and walked along the yellow line, facing the wall, until the ends of the 220' baseline--the outside edges of the end pillars--barely touched each edge of the frame. I marked that spot on the yellow line, and then switched to photo mode. With its wider FoV, I had to move toward the wall to get the two pillars at the edge of the frame. Then I measured the distance of each.

The sketch below is based on a perspective view (from Litchi Mission Hub), so the yellow line is somewhat foreshortened and the proportions don't look quite right. Sorry; anyway, the numbers are what matters.
  • The point on the yellow line where the outer edge of the end pillars spanned the frame in video mode (pink) was 195' from the center point of the red line.
    • That yields an FoV of very nearly 58.9° for MA video.
  • The corresponding point in photo mode (blue) was 160' from that center point.
    • That yields an FoV of almost exactly 69° for MA photos.

1586563136278.png

Here's the still image. Note the small yellow stake in foreground, with the tape measure attached to it and extending toward the camera; that stake is 150' from the wall:
1586564312255.png

And a frame of the video. (I'm actually a couple of feet too close, cutting off most the actual end pillars, so the true FoV is slightly less than I've calculated, closer to 58°). Here, note again the measuring tape that ends at that yellow stake. Neither the stake nor the tape were moved between images, but here--because of the narrower FoV in video--the stake is much farther away.
1586617043368.png

However DJI calculates FoV for the MA, it bears little relevance to reality. I'd be surprised if this is not true in varying degrees across the DJI lineup.
 
If going into the exif DJI gives away the real focal length ... it is 4,5mm. the 1/2,3" sensor is 6,3x4,7mm.

With this into this formula:
FOV = 2 arctan(SensorSize/2f)

where:

- f is the focal length
- sensor size is the Height or Width or Diagonal of the sensor (in mm)

Gives the FOV 69,98° for the MA's FC2103 camera.
 
If going into the exif DJI gives away the real focal length ... it is 4,5mm. the 1/2,3" sensor is 6,3x4,7mm.

With this into this formula:
FOV = 2 arctan(SensorSize/2f)

where:

- f is the focal length
- sensor size is the Height or Width or Diagonal of the sensor (in mm)

Gives the FOV 69,98° for the MA's FC2103 camera.
Assuming you used sensor width in the formula, that's close to my empirical result for still photos, certainly within my margin of error. Thanks!

It's unclear why so much of the sensor is left unused for recording video. While it seems like a waste, I'm sure the design engineers didn't make stupid choices. (I'm one of those crazy people who lend credence to actual expertise. ??)

What's annoying is DJI's misrepresentation on the matter.
 
Summary: The camera on the Mavic Air is advertised with an 85° field-of-view. A careful test shows something very different!
  • When shooting video, the measured FoV is about 59° (actually, more like 58°, see below).
  • When shooting still photos the measured FoV is 69°
---------------------------------------------------------------------------------
I was puzzled to find that shots I'd carefully composed offline (using Virtual Litchi Mission) were not framed as I'd expected in the field. It became apparent that the advertised 85° field-of-view is not accurate. So I ran an experiment to measure the true field of view for the Mavic Air. Some walking and simple trigonometry gave the answer. Was I surprised! ?

Free to wander the campus of a local high school, and using a 50' tape, I measured a long baseline segment of a pillared wall (see below). I made sure each end of the baseline was the outside edge of a pillar, to make framing easy. The baseline--red in the diagram--came out to 220'. From the center of that baseline, I measured a perpendicular line--yellow in the diagram--out to 200', with stakes at 100', 150', and 200'.

I started the MA (without props) so that I could see and capture images hand-held. I put it into video mode, and walked along the yellow line, facing the wall, until the ends of the 220' baseline--the outside edges of the end pillars--barely touched each edge of the frame. I marked that spot on the yellow line, and then switched to photo mode. With its wider FoV, I had to move toward the wall to get the two pillars at the edge of the frame. Then I measured the distance of each.

The sketch below is based on a perspective view (from Litchi Mission Hub), so the yellow line is somewhat foreshortened and the proportions don't look quite right. Sorry; anyway, the numbers are what matters.
  • The point on the yellow line where the outer edge of the end pillars spanned the frame in video mode (pink) was 195' from the center point of the red line.
    • That yields an FoV of very nearly 58.9° for MA video.
  • The corresponding point in photo mode (blue) was 160' from that center point.
    • That yields an FoV of almost exactly 69° for MA photos.

View attachment 98386

Here's the still image. Note the small yellow stake in foreground, with the tape measure attached to it and extending toward the camera; that stake is 150' from the wall:
View attachment 98389

And a frame of the video. (I'm actually a couple of feet too close, cutting off most the actual end pillars, so the true FoV is slightly less than I've calculated, closer to 58°). Here, note again the measuring tape that ends at that yellow stake. Neither the stake nor the tape were moved between images, but here--because of the narrower FoV in video--the stake is much farther away.
View attachment 98418

However DJI calculates FoV for the MA, it bears little relevance to reality. I'd be surprised if this is not true in varying degrees across the DJI lineup.

I think you may have misunderstood the FOV definition - it's diagonal field of view. Correcting for that gets approximately 83° for the photo mode.

If going into the exif DJI gives away the real focal length ... it is 4,5mm. the 1/2,3" sensor is 6,3x4,7mm.

With this into this formula:
FOV = 2 arctan(SensorSize/2f)

where:

- f is the focal length
- sensor size is the Height or Width or Diagonal of the sensor (in mm)

Gives the FOV 69,98° for the MA's FC2103 camera.

You have also used horizontal FOV, not diagonal.

2tan⁻¹(√(6.3² + 4.7²)/(2*4.5)) = 82.3°
 
Last edited:
I think you may have misunderstood the FOV definition - it's diagonal field of view. Correcting for that gets approximately 83° for the photo mode.



You have also used horizontal FOV, not diagonal.

2tan⁻¹(√(6.3² + 4.7²)/(2*4.5)) = 82.3°
Yeah I know ... it was intentional (but perhaps unclear) & was relating to Prismatic's horizontal measurement. That's why I wrote "is the Height or Width or Diagonal "
 
  • Like
Reactions: sar104
I think you may have misunderstood the FOV definition - it's diagonal field of view. Correcting for that gets approximately 83° for the photo mode.

You have also used horizontal FOV, not diagonal.

2tan⁻¹(√(6.3² + 4.7²)/(2*4.5)) = 82.3°
Well, that is news. Thanks ?. I still don't understand why it's different for video.
 
Depends on the aspect ratio. 16:9 uses less of the sensor and has a smaller diagonal field of view than 4:3.
I understand (despite my initial error in the definition of FoV). This all came from my frustration with the difference between my "virtual framing" (using VLM) and the actual framing from a flight. And in that, at least, I'm right.

Given what I've learned here, I should specify a FoV of 74.5° in Virtual Litchi Mission, not 85°. While the specification may be accurate for the camera, in video mode the smaller effective sensor reduces the FoV by ~8°. That makes quite a visual difference!

In free flight, you can frame images to your liking on-the-fly (so to speak). You don't have that luxury with a pre-planned mission. It's a real disappointment to discover that (say) the horizon--which you wanted just at the top of the frame--is not visible at all.

Thanks for the help!
 
I understand (despite my initial error in the definition of FoV). This all came from my frustration with the difference between my "virtual framing" (using VLM) and the actual framing from a flight. And in that, at least, I'm right.

Given what I've learned here, I should specify a FoV of 74.5° in Virtual Litchi Mission, not 85°. While the specification may be accurate for the camera, in video mode the smaller effective sensor reduces the FoV by ~8°. That makes quite a visual difference!

In free flight, you can frame images to your liking on-the-fly (so to speak). You don't have that luxury with a pre-planned mission. It's a real disappointment to discover that (say) the horizon--which you wanted just at the top of the frame--is not visible at all.

Thanks for the help!
I also have a Field of View difficulty in relation to GoogleEarth and Virtual Litchi Mission. This thread (Yet another Virtual Litchi Mission (Chrome/Firefox Extension)) explores the issue a little, with a few sample pictures.

I find that if I change the aspect ratio of my Google Earth Pro window (from, say, 4:3 to 16:9 or 1:1), Google Earth will always show a full window, irrespective of what setting I make for the virtual drone. Therefore, predicting the Field of View is far from a precise science...

Exaggerating the problem...
Below - same virtual film frame, different window ratios, different FoV?
Contrasting VLM Shots in Google Earth - Tall.jpg. Contrasting VLM Shots in Google Earth - Wide.jpg
 
  • Like
Reactions: Prismatic
I also have a Field of View difficulty in relation to GoogleEarth and Virtual Litchi Mission. This thread (Yet another Virtual Litchi Mission (Chrome/Firefox Extension)) explores the issue a little, with a few sample pictures.

I find that if I change the aspect ratio of my Google Earth Pro window (from, say, 4:3 to 16:9 or 1:1), Google Earth will always show a full window, irrespective of what setting I make for the virtual drone. Therefore, predicting the Field of View is far from a precise science...

Exaggerating the problem...
Below - same virtual film frame, different window ratios, different FoV?
View attachment 98894. View attachment 98895
The FoV (or AoV) is measured on the diagonal of the frame. As I noted, it's 74.5 degrees for the Mavic Air when shooting video, which is of course 16:9. If I set VLM FoV to 74.5, and set the GE viewport to any 16:9 size, the simulation is very close to the actual shot.

If the GE viewport is not 16:9, then yeah, the FoV is whatever, but who cares?
 
Last edited:
Yes agree the FOV is 68 or so degrees wide. Spec says 85 degrees.

I am helping some students to "map" some ground, and I can really only simply scale the photos by using the photo width as a dimension. It will only work on flat ground.

I did a measurement over 1 meter in the kitchen, and found the FOV to be 1.36 metres. In the field I did a rough measurement from a 30 metre height, and came up with about a 40 metre FOV. Not a precise measurement. I will be doing a more precise measurement before the project starts...

OK perhaps the FOV is diagonal, why I don't know why that dimension is chosen, but hey ho, we agree on the 68+ Horizontal FOV
 
Cool project for your students!

OP here: This thread led me to further study the question. I then posted a comprehensive explanation of how you can calculate the FoV for any camera, using the Mavic Air as an example. You'll find it is very different when shooting video as compared to still photos.

If you're using the original Mavic Air for this project, you can use the formula in that post to show that--for still photos--the drone sees a 70° field on the wide dimension. For video, the wide dimension covers 67°.

By convention, the FoV of any camera is specified on the diagonal. TV screen sizes are also specified on the diagonal.
 
  • Like
Reactions: MJ224 and Facherty
Lycus Tech Mavic Air 3 Case

DJI Drone Deals

New Threads

Forum statistics

Threads
130,592
Messages
1,554,169
Members
159,596
Latest member
da4o98