Metadata extraction from samsung gear 360

[giuli]

Hi every one!
I'm trying to extract metadata information like gps position, speed and acceleration from a video taken by a samsung gear 360, but i can't. Does some of you know if it is possible or not?
Thank you very much!
Giulia

[StarGeek]

Can you provide a sample file?

[giuli]

Yes, because of the video is large like 137 MB can I send it to you via email or something like that?

[Phil Harvey]

Could you upload it to a google drive or some other file hosting service, and post the URL here?

- Phil

[giuli]

Yes, already on google drive, this is the link: https://drive.google.com/open?id=15d3Tdabwq0nl4PVqA6v4BTc4Euj37r47

Thank you very much

[Phil Harvey]

Well, there isn't a metadata track like many cameras write.  However, there are some unknown UserData tags, one of them quite large (vrot), but to me it looks like this just contains accelerometer data, not GPS positions.  Do you have any software that will display GPS from this video?  If so, we stand a chance, but at the moment I can't see where it could be hiding.

Here is the unknown UserData.  Any ideas about any of it?:  (specifically, is "CFN0017CDEE06A3" the camera serial number?)

Code Select Expand

  | + [UserData directory]
  | | Unknown_vrot = .h................................................................[snip]
  | | - Tag 'vrot' (45512 bytes):
  | |  891460a: 00 00 00 00 00 00 07 68 00 00 00 00 00 00 00 0a [.......h........]
  | |  891461a: 00 00 00 1a 00 00 00 0a 00 00 00 1c 00 00 00 0a [................]
  | |  891462a: 00 00 00 00 00 00 00 0a 00 00 00 19 00 00 00 0a [................]
  | |  891463a: 00 00 00 1d 00 00 00 0a 00 00 00 00 00 00 00 0a [................]
  | |  891464a: 00 00 00 18 00 00 00 0a 00 00 00 1d 00 00 00 0a [................]
  | |     [snip 45432 bytes]
  | | Unknown_m360 = .
  | | - Tag 'm360' (8 bytes):
  | |  891f7da: 00 00 00 00 00 00 00 01                         [........]
  | | Unknown_opax = ...........~.:...@...L.z..6d..&....G......d}....`.x.....x..b..Fg..[snip]
  | | - Tag 'opax' (164 bytes):
  | |  891f7ea: 00 00 00 01 93 0d 00 00 d7 fc ff ff 93 ff ff ff [................]
  | |  891f7fa: 7e 14 00 00 3a ca f9 ff 40 e8 f8 ff 4c 00 04 00 [~...:...@...L...]
  | |  891f80a: 7a f4 03 00 36 64 9e 0b 26 a7 f2 ff e2 47 fb ff [z...6d..&....G..]
  | |  891f81a: 08 d7 0e 00 ca 64 7d 01 b7 ec 03 00 60 03 00 00 [.....d}.....`...]
  | |  891f82a: 78 a9 00 00 ea 9e ff ff 78 0f 00 00 1f 62 01 00 [x.......x....b..]
  | |     [snip 84 bytes]
  | | Unknown_opai = .%.).CFN0017CDEE06A3
  | | - Tag 'opai' (32 bytes):
  | |  891f896: 00 00 00 00 01 00 00 00 25 05 00 00 29 07 00 00 [........%...)...]
  | |  891f8a6: 43 46 4e 30 30 31 37 43 44 45 45 30 36 41 33 00 [CFN0017CDEE06A3.]
  | | Unknown_intv =
  | | - Tag 'intv' (16 bytes):
  | |  891f8be: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 [................]

- Phil

[Phil Harvey]

The 'vrot' data consists of 3 columns of rational64s numbers.  The first (series1, green) wraps at 360 (plotted below with 360 subtracted from values over 180), so it is likely some angle in degrees.  The second (series2, blue) is always zero.  The third (series3, red) goes positive and negative.  Here they are all plotted.  I don't have any ideas yet what they could be:



- Phil

[giuli]

I don't know if that one if "CFN0017CDEE06A3" is the camera serial number because the camera is not mine, I've only the videos.
I think that it has the gyro, but not the gps. I'm not sure about accelerometer. I don't know any software which shown the data. The only thing, which maybe could be useful, is that there is a software to stitch the video, so maybe all that data are used only for that aim.

Giulia

[Phil Harvey]

Doing more research into this, it seems that the plotted values may be yaw (blue), pitch (red) and roll (green) angles, although looking at the video this doesn't make a lot of sense to me.

- Phil

Edit:  Looking more closely, I can see this now.  The angles are the orientation of the camera from the accelerometer.  With the way the camera was oriented for this video, positive roll (green) is acceleration to the right, and positive pitch (red) is acceleration backward (ie. braking the car).  I'll decode this in an AccelerometerData tag in the next ExifTool release.

[giuli]

Thank you very much!!
Giulia

[giuli]

Hi Phil!
I'm trying to extract the accelerometer data with this command
exiftool -ee -n -p "$AccelerometerData" gear360.MP4
as I've done for other cameras, but it gives in output this string only

Code Select Expand

Binary data 20645 bytes

Do I have to change something in the command?
Thank you

[Phil Harvey]

Try this:

exiftool -b -accelerometerdata FILE

- Phil

[giuli]

Thank you, the command works. The output is like this

Code Select Expand


0 2.6 2.8 0 2.5 2.9 0 2.4 2.9 0 2.3 2.9 0 2.3 2.8 0 2.2 2.8 0 2.2 2.7 0 2.1 2.7 0 2.1 2.6 0 2 2.6 0 2 2.6 0 2 2.6 0 1.9 2.6 0 1.9 2.7 0 2 2.7 0 1.9 2.7 0 1.9 2.7 0 1.9 2.7 0 1.9 2.6 0 2 2.6 0 2 2.6 0 2 2.7 0 2.1 2.7 0 2.2 2.6 0 2.3 2.6 0 2.4 2.6 0 2.6 2.5 0 2.8 2.4 0 3 2.4 0 3.2 2.4 0 3.4 2.4 0 3.6 2.3 0 3.8 2.3 0 4 2.3 0 4.3 2.1 0 4.5 2 0 4.7 1.8 0 4.9 1.8 0 5.1 1.8 0 5.3 1.9 0 5.5 1.9 0 5.6 1.9 0 5.7 1.8 0 5.8 1.7 0 5.9 1.6 0 5.9 1.6 0 5.9 1.5 0 5.9 1.5 0 5.8 1.5 0 5.7 1.5 0 5.6 1.5 0 5.4 1.4 0 5.2 1.4 0 5 1.3 0 4.8 1.2 0 4.6 1.1 0 4.4 1.1 0 4.2 1 0 4 1 0 3.7 1 0 3.4 1 0 3.2 1 0 3 0.8 0 2.9 0.6 0 2.8 0.3 0 2.7 0.2 0 2.6 0.1 0 2.5 0.2 0 2.4 0.2 0 2.3 0.3 0 2.2 0.3 0 2 0.3 0 1.8 0.2 0 1.7 0.1 0 1.5 0 0 1.4 359.9 0 1.2 359.8 0 1 359.7 0 0.9 359.5 0 0.7 359.4 0 0.6 359.3 0 0.4 359.3 0 0.3 359.2 0 0.1 359.2 0 0 359.1 0 -0.1 359.1 0 -0.3 359 0 -0.5 359 0 -0.8 358.9 0 -1.1 358.8 0 -1.4 358.7 0 -1.9 358.6 0 -2.3 358.4 0 -2.8 358.3 0 -3.4 358.2 0 -4 358.1 0 -4.7 357.9 0 -5.3 357.7 0 -6 357.4 0 -6.6 356.9 0 -7.2 356.5 0 -7.8 356 0 -8.2 355.5 0 -8.6 355.1 0 -9 354.9 0 -9.4 354.8 0 -9.8 354.6 0 -10.2 354.4 0 -10.5 354.2 0 -10.7 353.8 0 -10.8 353.3 0 -10.8 352.8 0 -10.8 352.3 0 -10.8 351.9 0 -10.9 351.6 0 -11 351.3 0 -11.1 351 0 -11.2 350.7 0 -11.2 350.5 0 -11.2 350.2 0 -11.1 349.9 0 -11.1 349.7 0 -11.1 349.4 0 -11.1 349.2 0 -11.1 349 0 -11.1 348.8 0 -11 348.6 0 -10.9 348.4 0 -10.8 348.2 0 -10.7 347.9 0 -10.7 347.6 0 -10.6 347.3 0 -10.5 347 0 -10.5 346.7 0 -10.4 346.4 0 -10.4 346.2 0 -10.5 345.9 0 -10.5 345.8 0 -10.4 345.7 0 -10.4 345.6 0 -10.3 345.6 0 -10.3 345.5 0 -10.4 345.4 0 -10.4 345.4 0 -10.5 345.3 0 -10.5 345.4 0 -10.6 345.6 0 -10.6 345.7 0 -10.6 345.9 0 -10.7 346.1 0 -10.7 346.4 0 -10.8 346.7 0 -10.9 347 0 -11 347.4 0 -11 347.7 0 -11 348.1 0 -11 348.4 0 -11 348.8 0 -11 34


Are data saved as acc_x1 acc_y1 acc_z1 acc_x2 acc_y2 acc_z2  and so on or in another way?

[Phil Harvey]

The measurements are yaw, pitch and roll rotations in degrees (which is rather odd).  To convert to acceleration in g, I presume you would take the sine of these angles.  Note that the yaw is always zero.

- Phil

[giuli]

Perfect,
thank you very much!

Giulia

[Phil Harvey]

Hi Giulia,

Thinking about this a bit harder (and  drawing some diagrams this time), you would take the tangent (tan), not the sine (sin), to convert to acceleration in g.

- Phil