Welcome Inspire Pilots!
Join our free DJI Inspire community today!
Sign up

Jello in high cross winds

Joined
Mar 20, 2014
Messages
3,460
Reaction score
1,674
Location
Boston, MA
It's become apparent to me that

1) The Inspire does a significant amount of software-based stabilization, needed by the

2) Small footprint and poorly supported gimbal that shakes significantly on windier days and

3) Cross winds overwhelm the software based stabilization capability,

4) Resulting in jello

This is a design flaw of the gimbal in my opinion. It is poorly supported with the small attachment face plate surface (unlike the Phantom H3-3D, with its much wider attachment surface), resulting in easy shaking/swaying in moderate cross winds.

The only remedy I've found frankly is to move the I1 to an angle that minimizes the winds buffeting effect on the gimbal, so the flat lens surface is not at an angle creating the most windward surface.

No commentary needed or requested, just an engineering observation and some disappointment as I've had to ditch some otherwise gorgeous footage due to wind-induced jello.
 
Last edited:
Joined
Jan 8, 2015
Messages
399
Reaction score
130
Location
Chicago
Website
vimeo.com
Hmm. Ive always wondered why videos showing the craft, the cameras shook so much and the videos posted didn't.
It is a thin footprint thinking about it. Do they sell different density rubber boots I wonder.. A slightly stiffer joint might be the job

So far, I've been pretty happy. Heck, it's usually me that buggers the shot
 
Joined
Mar 20, 2014
Messages
3,460
Reaction score
1,674
Location
Boston, MA
The stabilization SW does an admirable job but it has its limits. Really severe shaking is beyond its ability to fix the image.

Come to think of it, I wouldn't be surprised if the gimbal shaking due to its shape and mount might have actually been the rumored production line delay as the stabilization SW got added to compensate. But that's just speculation. It is clear to me the shaking gimbal can exceed the stabilizer SW's ability to correct the image when the shaking becomes severe and that's the point where you see artifacts in the resulting video.

For a demo of the stabilizer software in action doing its thing under normal circumstances, see it in action and the side by side results:
 
Last edited:
Joined
Jan 8, 2015
Messages
399
Reaction score
130
Location
Chicago
Website
vimeo.com
That is wild. So all that digital stabilization has to come at a price as goes image quality. Just seems too loose to be able to do what it does. Thanks for posting the video, very interesting
 
Joined
Feb 3, 2015
Messages
1,882
Reaction score
733
I am most definitely sure that NO digital stabilization is being done. Given the lens characteristics it would be very apparent, but is not.
Actually the video posted above (when the I1 cam and gopro film each other) leave no doubt about it.

Small linear moevements like that are unnoticeable as long as the subject is farther than a couple of meters.
Jello is caused by small high frequency movements, and the video shows that the rubber boots make a good job at only letting those with much lower frequency to go through.
 
  • Like
Reactions: ted35
Joined
Mar 20, 2014
Messages
3,460
Reaction score
1,674
Location
Boston, MA
I am most definitely sure that NO digital stabilization is being done. Given the lens characteristics it would be very apparent, but is not.
Actually the video posted above (when the I1 cam and gopro film each other) leave no doubt about it.

Small linear moevements like that are unnoticeable as long as the subject is farther than a couple of meters.
Jello is caused by small high frequency movements, and the video shows that the rubber boots make a good job at only letting those with much lower frequency to go through.
Skip forward in the video where the Inspire is filming the GoPro in Free mode. About 1 foot or less away.
 
Joined
Feb 3, 2015
Messages
1,882
Reaction score
733
That's exactly what I was talking about, the lateral movements recorded by both cameras match perfectly.

If there was digital stabilisation applied in the I1 camera the GoPro would appear to be moving less in the I1's image, the movement would be smoothed and would lag a bit behind.

And the I1's lens isn't perfect and is a little soft on the outside (for me, some have soft patches closer to the center...), if there was digital stabilization these zones of different sharpness would move around the image, but they don't.
 
  • Like
Reactions: ted35
Joined
Mar 20, 2014
Messages
3,460
Reaction score
1,674
Location
Boston, MA
That's exactly what I was talking about, the lateral movements recorded by both cameras match perfectly.

If there was digital stabilisation applied in the I1 camera the GoPro would appear to be moving less in the I1's image, the movement would be smoothed and would lag a bit behind.

And the I1's lens isn't perfect and is a little soft on the outside (for me, some have soft patches closer to the center...), if there was digital stabilization these zones of different sharpness would move around the image, but they don't.

You must be looking at a different video :).

The movement is there but vastly minimized by stabilization. It's ok to have stabilization on board. The GoPro has it, other cameras have it, it's fine.

What's not fine is the the shake and wobble of the I1 camera / gimbal combo. Under normal conditions (this video) it's not a big deal.

Under windy conditions with some angled flight vs wind direction, flight wind speed plus air wind speed can produce violent shaking that cannot be overcome by the onboard stabilization. Artifacts in the video result and can ruin otherwise good and expensive to shoot footage.

The gimbal needs support. It's possible the stiffer dampeners shipped with the I1 can help, as is simply checking footage on site and being prepared to do additional takes from different angles.

Either way, it's an issue that you need to be aware of in windy conditions, and the gimbal design is the main culprit in my opinion.
 

The Editor

Moderator
Staff member
Joined
Aug 7, 2013
Messages
9,870
Reaction score
5,186
Additionally, if the I1 camera was optically stabilised you would be able to see the lens moving on both axis,which it doesn't..... I don't mean on this video, I mean in general)
If it was electronically stabilised the sensor would need to be larger than 4096x2160 to allow for additional 'float room'
Don't forget the Zenmuse camp has a correction accuracy of +/- 0.03 degree which is pretty astounding really
 
  • Like
Reactions: Kilrah
Joined
Feb 3, 2015
Messages
1,882
Reaction score
733
The movement is there but vastly minimized by stabilization.
By the gimbal's stabilization, yes... but there is no additional stabilization besides the mechanical one.

The GoPro has it
Nope it doesn't :) (and it's one of the long-standing complaints about it)

other cameras have it
Yep.

Under windy conditions with some angled flight vs wind direction, flight wind speed plus air wind speed can produce violent shaking that cannot be overcome by the onboard stabilization. Artifacts in the video result and can ruin otherwise good and expensive to shoot footage.
You might need to adjust your expectations a bit... take any other system out there and it will actually either not even be as clean (including heavy, expensive rigs like S900/GH4/Zenmuse), or it will take days of tweaking to jsut match that, and much more to get better.

Sure it won't be perfect in all conditions, but that's life. It already works **** well.
 

The Editor

Moderator
Staff member
Joined
Aug 7, 2013
Messages
9,870
Reaction score
5,186
I believe I read somewhere (but can't remember where) that the supplied stiffer dampers are actually meant for windy conditions or high speed flight.

Why not give them a try and see if things are improved.
 
Joined
Mar 20, 2014
Messages
3,460
Reaction score
1,674
Location
Boston, MA
By the gimbal's stabilization, yes... but there is no additional stabilization besides the mechanical one.


Nope it doesn't :) (and it's one of the long-standing complaints about it)


Yep.


You might need to adjust your expectations a bit... take any other system out there and it will actually either not even be as clean (including heavy, expensive rigs like S900/GH4/Zenmuse), or it will take days of tweaking to jsut match that, and much more to get better.

Sure it won't be perfect in all conditions, but that's life. It already works **** well.
My bad: I thought GoPro had stabilization. Heck even your phone has it these days.

It does work well and nobody said it would work perfectly in all situations. It's an observation and comment on the gimbal design.

I remain convinced the I1 camera does stabilization in addition to the gimbal-provided stabilization and I guess you and I will have to "agree to disagree" on that point, since nobody from the DJI camera engineering team will confirm or deny, but it's immaterial to the key point I'm making: the gimbal design is unstable in windy conditions and under certain conditions can result in severe artifacts that can render video unusable and unfixable.

I love the Inspire. But it's got some room for improvement. My hope is the mythical camera upgrade, if and when it comes, will sport not only a massive MFT 4K sensor, less compression and amazing picture quality but also some gimbal design improvements to help stabilize the system further.
 
Joined
Feb 3, 2015
Messages
1,882
Reaction score
733
My hope is the mythical camera upgrade
I'll believe in it when I see it. My own take on the whole "OMG they joined the MFT group!!!11!ONE" is that DJI are indeed going to design their own MFT interchangeable lens camera, but it will be for next year's refresh of the S series so they also get DJI-made cameras like all their other product ranges. Not the Inspire as that will be too heavy for the mounting system and would cause too much drop in flight time, and not before next year as there's a pesky thing called development to be done between getting the specs and selling something and that takes a bit more than 6 months.

But who knows, hopefully I'm wrong...

Regarding mounting, too rigid and you get jello, too soft and you get too much shake in the airflow. No miracle, you need to choose a compromise there.
BUT wind doesn't typically cause jello, so it would be interesting to see some footage you mention to see what you're talking about.
 
Last edited:
Joined
Mar 20, 2014
Messages
3,460
Reaction score
1,674
Location
Boston, MA
I'll believe in it when I see it. My own take on the whole "OMG they joined the MFT group!!!11!ONE" is that DJI are indeed going to design their own MFT interchangeable lens camera, but it will be for next year's refresh of the S series so they also get DJI-made cameras like all their other product ranges. Not the Inspire as that will be too heavy for the mounting system and would cause too much drop in flight time, and not before next year as there's a pesky thing called development to be done between getting the specs and selling something and that doesn't take 3 months.

But who knows, hopefully I'm wrong...
I hope you're wrong, yeah. I think if they built a fixed-glass MFT camera from scratch it could easily go on the Inspire without a lot of extra weight. As Eric Cheng stated, most of the weight of the DSLRs are unnecessary for flying. Strip it down to a custom built camera with no body and I'd venture it wouldn't weigh more than a few hundred extra g vs the I1 camera.

We'll see. Maybe they'll surprise us at NAB.
 

New Threads

Members online

Forum statistics

Threads
19,381
Messages
190,116
Members
28,738
Latest member
oldcomputech