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

Gimbal/camera shake?

It does not appear to be Gimbal or Camera shake. Here is why.

Put a piece of paper over the center of the video you linked while it plays. The top and bottom of the video is pretty much rock solid. If it were gimbal/camera shake you would see the entire image being affected.

This appears much more related to the video being processed as it is "fed" to the onboard SD card. Its a guesstimate that the SD card speed is unable to keep up with the raw data that is being transferred to it. There by causing a momentary lag that causes this . Try with the DJI provided SD card against the SD card you used for this video. I am assuming you used a different card besides the DJI provided SD card for this video.

Class 10 cards are strongly recommended to support the plentiful amount of data that needs to be written to the SD card.
 
  • Like
Reactions: borislip
It does not appear to be Gimbal or Camera shake. Here is why.

Put a piece of paper over the center of the video you linked while it plays. The top and bottom of the video is pretty much rock solid. If it were gimbal/camera shake you would see the entire image being affected.

This appears much more related to the video being processed as it is "fed" to the onboard SD card. Its a guesstimate that the SD card speed is unable to keep up with the raw data that is being transferred to it. There by causing a momentary lag that causes this . Try with the DJI provided SD card against the SD card you used for this video. I am assuming you used a different card besides the DJI provided SD card for this video.

Class 10 cards are strongly recommended to support the plentiful amount of data that needs to be written to the SD card.
You've got a point here. I am using a class 10 64gb sandisk card, will try with the one that came with the mavic, if i am, indeed, no longer able to reproduce, i'll buy a higher speed card.

Thanks.
 
That would be a good comparison.
If it persists with the DJI included card and you can reproduce the issue then DJI should absolutely warranty this unit as there is something involving the codec, compression, and /or computational power / ability of your specific Mavic Pro.

Make sure to compare apples to apples and oranges to oranges.
Write down the Video FPS and Resolution using each of the different cards.

It is an unfair comparison if you use different Frames Per Second and video resolutions when testing.
 
Didn't have a chance to try and reproduce it with the original SD card (although i am almost sure one of the videos this happens on actually comes from that original Mavic SD card, but almost doesn't mean 100% sure).

I did have a chance to look at it frame by frame, and, sadly, it doesn't looks like something related to the SD not keeping up. Take a look below. There are just 2 bad frames. And it look like some kind of interference, centered around some specific scan line, getting everything on that line out of focus and shifted right. The further away from that line, the less the effect is. On the next bad frame the line moves down to another position, the distortion seems centered somewhere either at the bottom of the frame or just below it now. On the following frame there is no more interference at all.

Sadly, i can't imagine it being any kind of write speed/buffer/compression issue. My best guess here is EMI. And as far as i remember, people did happen to spot EMI issues with other part of Mavic before, the compass. I remember reading about someone trying to minimize it by turning it's wires into twisted pair, too. Could it be that something on the way from the camera is way more susceptible to EMI than it should be? Could the sensor itself pick up an EMI induced by the gimbal motor coils, or the main motors, or anything on the main board, either from the air, or over the circuitry, as a result of poor board design?

I have to say, this does coincide with flying in sport mode and in high wind, so the load on pretty much everything is the highest, including main motors, gimbal motors, as well as pretty much every computational core on board (CPUs, MCUs, DSPs whatever is in there). Even the fan inside is probably stressed taking in gusts of wind.

Reviewing the videos again, i can't indeed find a video recorded in relatively calm condition and not in sport mode, that has this effect on it, but i didn't do much with the Mavic in calm weather yet.

Note: posting as food for though, will still try to reproduce on the original SD card, but my gut tells me i'll be able to do so :-(

DJI_0011_1488582893824.MP4 - 00000.png DJI_0011_1488582893824.MP4 - 00001.png DJI_0011_1488582893824.MP4 - 00002.png DJI_0011_1488582893824.MP4 - 00003.png

TBH, if i am, indeed, correct about this being an EMI issue, which only shows up when the machine is under high stress, the only solution may be not to take videos unders such conditions :-(
 
One more example, frame by frame. Found a longer one than just 2 frames this time. This one does look like camera vibrating, IMO. Kinda looks like camera footage on a Syma X5C with chipped propellers (that make it vibrate), with an addition of focus here is also being shaky.

DJI_0015.MP4 - 00010.png DJI_0015.MP4 - 00011.png DJI_0015.MP4 - 00012.png DJI_0015.MP4 - 00013.png DJI_0015.MP4 - 00014.png DJI_0015.MP4 - 00015.png DJI_0015.MP4 - 00016.png DJI_0015.MP4 - 00017.png DJI_0015.MP4 - 00018.png DJI_0015.MP4 - 00019.png

PS: I don't like it. I don't like it at all. And i like the idea of RMA-ing even less :(
 
Just tried to reproduce the issue indoors, moving the Mavic around, taking off and jerking it mid flight to random sides, quite violently, tapping it mid flight, lightly tapping the gimbal mid flight, shaking it quite violently by hand (not mid flight, obviously), and couldn't reproduce this "few frames glitch" once. Thinking "hmm, maybe its SD card after all, the original card is now in there", i've repeated the torture session with the very card i had this issue on. Nope. Not reproduced indoors. Period. If i try any more violently than what i already did, i'll either manually break it or crash it against some really expensive stationary object, such as TV panel, causing major damages.

And the main thing i can't simulate indoors is flying full speed in a high wind.
 
I am wondering if this is related to the over stressing as well. It looks to me like its related to throughput.
It could also be focus "hunting" for a millisecond and then regaining focus. You could switch to manual focus to try and isolate that and fly in Sports Mode to stress the Mavic and see if the issue is still presented . Using the SD card that you used when this video artifact occurred.
 
I am wondering if this is related to the over stressing as well. It looks to me like its related to throughput.
It could also be focus "hunting" for a millisecond and then regaining focus. You could switch to manual focus to try and isolate that and fly in Sports Mode to stress the Mavic and see if the issue is still presented . Using the SD card that you used when this video artifact occurred.
I have AFC disabled anyway, i just tap once, so it shouldn't have been doing any focus hunting. Also, focusing takes way, way more than 1..2 frames or even 10 frames, you can easily see how it focuses when you tap. It also doesn't explain the sinusoidal like appearance of it. A lense shaking inside the camera, riding its focusing mechanism (whats is it? A simple coil?), could easily do this, though. Maybe sports mode along with the strong wind really are that bad for the Mavic. Kinda disappointing. You need sports mode if you want to produce visble movement while 400 feet AGL, inability to film in it would suck :-(

I'll do the experiments sooner or later, cant really do it right now.

Not reproducible on demand at home/lab + RMA, how does that combination sound... :-(
 
People seem to be discussing the very same thing on another long-аss thread :-(

Definitely not an SD issue. Reproduced with the original SD (in strong wind, flying in sport mode again).
 

DJI Drone Deals

New Threads

Forum statistics

Threads
130,974
Messages
1,558,481
Members
159,964
Latest member
swigmofa