Aerial photography -- remote wing mount

[Eric Muehling]

Last weekend I mounted a Canon XH-A1 to an aircraft and went flight seeing. Please, no more FAA comments (re: external load permit).

There is a jerkyness problem with the video. Maybe someone can tell me the exact cause.

See this URL for a four-minute, small frame, edited version (Flash video).
The jerkyness is reduced since this is a small-screen version.
http://www.ericmuehling.com/video/aerial_shorts.html

See this full-size 10-second video that shows the image jerkyness problem.
Video compression is not causing the jerkyness (compression masks it a bit).
Can anyone tell me how to solve the jerkyness problem?
http://www.ericmuehling.com/video/bluffs_short.html

First, I read all of your comments before taking these aerials. Thank you.
All of your comments helped. Sometimes I had to choose between
different opinions, but they all helped.

POSITIVE RESULTS
Safe trip. Interesting footage. No dropped frames. No damage to my precious Canon XH-A1. The footage was generally well-exposed and in focus.

-- The camera's downward angle was perfect (1/8 sky, 7/8 ground) and exactly what I wanted. The horizon on the footage was level when the airplane was straight and level.
-- Shutter speed (Tv 1/100) seemed just right.
-- It drizzled once, but the rain drops on the haze filter dried without leaving a residue. Thanks to Alaska's clean mountain air.

ROOM TO IMPROVE
I'm dissatisfied with the jerkyness of the images. Not professional. It looks like the image 'jumps' (and when it jumps it blurs) about once every 1/2 second. Note: see the dead mosquito near end of the video on the haze filter. Even dead, its silhouette moves slightly. I think this is an OIS issue.

Auto exposure resulted in some extreme, fast changes when the
sun reflected off water, or when the amount of blue sky in the frame increased, or when glacier snow showed up in the image. The next time I'll set the response times (automatic exposure controls) to their slowest to avoid the camera attempting to rapidly hunt for best exposure levels.
I may even shoot in manual mode, although I hate having to ride exposure all the time.

PHYSICAL SETUP
-- Canon XH-A1 in a steel bracket. Canon ZR-2000 wired remote controller. -- Sony DVD player (as a monitor).
-- Foam rubber (1/4 inch) provided some insulation from vibration between camera/bracket.
-- Blue painters' masking tape covered all gaps and slots on the XH-A1 (to avoid air impacting the tape drive mechanism). The blue tape is easy to remove. The microphone was wrapped with a plastic baggie (to avoid air impact damage).
-- Covering everything (camera and bracket) ... duct tape.

XH-A1 SETTINGS
-- Tv (shutter at 1/100 second)
-- Manual focus (set at infinity most of the time)
-- Optical Image Stabilizer on
-- ND 1/6
-- Power Save option off
-- A1 filter

PROBLEMS
There appears to be a jerky frame about every 1/2 second.
Having a fuzzed-out frame is very consistent throughout the three-hour trip. I am guessing at the cause:
-- Engine/prop vibration. Note, the pilot and I experimented with different RPMs during flight and we minimized this issue. But it never went away. It's possible prop vibrational frequencies created harmonics in the wing, or in my camera bracket.
-- Optical image stabilization was constantly trying to correct the image motion (when no correction was needed). Proof of this is the mosquito*that hit the A1 haze filter near the end of the flight. The dead mosquito moves around as a result of the OIS.

WHAT I WISH I HAD DONE ...
Turn off Optical Image Stabilizer.

The Sony DVD player (my monitor) battery lasted three hours. The flight was slightly longer. I forgot to bring the 12-volt adapter that could have powered the DVD player.

-- Reminder to self: Do most of the low flying near the end of the trip. Mosquitos*fly low.
-- Next time: OIS off.
-- Next time: Manual exposure (on a sunny day the ground exposure doesn't change that much).

[Jack Walker]

Was the jerkiness possibly caused by the OIS?

[Bob Hart]

My bet would also be on the optical image stabilisation and that in combination with the higher frequency motion induced by airframe vibration and the highly textured subject in motion through almost the entire frame area, that it is unable to do its job properly.

There may also be a rolling shutter issue but it is near impossible to assess this on my computer on downloaded material.

The once per half-second period, (15-frame GOP??) suggests that the OIS over-corrections, if they are happening, are also, along with the movement of a higly detailed and textured subject through almost the entire frame, - possibly too much for the HDV codec to deal with and that resolution is being thrown out for maintaining frame rate and that even this is getting compromised by the data load.

My mext move would be, fit a wide-angle adaptor on front to slow the motion through the frame and of course OIS selected "off" all also to reduce data change between frames and if this doesn't work, bring a little more sky into the shot, just a little, to reduce the data change between frames.

I would also set the video gain to manual and 0db as this on auto, may also add some data load when sensor noise is amplified.

Looks good so keep trying.

Seeing your footage so far, I can well understand why you want this to work - magnificent.

[Jack Walker]

I would be curious to see what it looked like slowed down 50%. If you shot at 100, then 50 would be close to normal. It wouldn't help the jerkiness, I suppose, but I think it may help the overall vibration and movement.

[Winfried Dobbe]

In case you can't re-shoot the flight and have to use this footage, you might try the new smoothcam filter of Final Cut Pro 6 (if you edit on a mac). It works usually very well, although it takes some time to analyse the footage.
If you are going to use it, remember that it analyses the entire clip, disregarding the in an out point, so capture just a small clip to test.

[Bob Hart]

On reflection, bringing the shutter speed back to 1/50th or 1/60th sec might assist. Motion blur might soften the fastest moving portion of the image in the lower part of the image frame and the slower moving portions in the upper part of the image frame might be enabled to stay sharp.

I would not mind being in that aircraft with you to solve this thing. That country below looks magic.

I think your mount should not need any mods except for safety and endurance. If there is airframe movement as I expect there is, nothing short of a mount system which totally isolates the camera will make it any better.

The only extra thing I might be tempted to do would be to add more support for the camera very firmly in a compressed foam cushion surrounding the front end of the lens enclosure.

This support would need to be rigidly attached to the main mount and also be firmly cushioned against the camera body sides. This is intended to eliminate the possibilty of the camcorder lens enclosure flexing relative to the main section of the camera body and moving the image laterally on the sensor.

A wide angle lens on front of the camera if available as an accessory should help slow the movement down, assist sharpness and maybe help with any data overload of the codec.

To eliminate the HDV codec as an issue, try a test shooting MiniDV at SD if this option is available in your camera. If the problem is still there in MiniDV SD, then there is another goose to chase.

Good luck. Don't forget to try a polariser when flying over some water with fish and reefs in it. You'll need to rehearse that handheld in the cabin and perhaps make notes as to time, direction of travel.

[Michael Pace]

gentlemen, this has been a fantastic thread.

i had some relevant experiences a few weeks ago w/ a dissimilar camera (1-chip Optura 50) but my jockeying around with mount, shutter speed and OIS may be useful:
I had to mount the optura to a diesel Kubota ATV ( a farm vehicle with full cabin, roll bars, windshield etc-- a tiny 'jeep' essentially). I used the same mount i had used previously to mount my XL-1 to cars and motorcycles-- from my moto experience i upped the shutter speed to as high as 1/2000, and the harmonics of the damned diesel engine simply played hell with the camera whether OIS enabled or not. I did tests of all shutter speeds from 1/60 to the upper end, no difference. To clarify: the quite solid Optura footage never showed a physical/mechanical tape dropout (ie the transport mechanism never glitched) , but the chip/imager could not cope with the harmonics of the diesel engine-- i can correlate the vibration with engine RPM using the on-board soundtrack; you can clearly see the imager stuttering in various power bands of the Kubota's motor.

Again, I know the cams are different, and the OIS parameters/performance of an Optura vs an A1 are likely significantly different, but i can't help but think that resonance from engine is always going to be a factor in any set-up that is'nt mechanically optimized for vibe isolation. And the relative wind speeds you're trying to work in are going to certainly exacerbate things.

My dad flies an Aztec and we've talked at length about shooting possibilities from it-- it's not a Cub, of course, and his Aztec has a plexi'd camera port up in the front wheel well, which makes cam mounting just an internal mount issue.

For an external solution I'm wondering if the smallest underwater camera housing available-- mounted to the Cub's strut-- might make life easier on you. You've already sussed out the remote-operation/monitoring thing so maybe that's an alternative?

Just my observations that we cannot expect OIS or full-wide lens settings to replicate dedicated (and spendy) aero-mounts when such severe operating vibration/wind-resistance harmonics are going to be present

Good luck, mang!

MRP

[Larry Huntington]

That is good footage for the first go around Eric. I have had my A1 out the window of a Cessna 182 using an adapted tripod (we took the front passenger seat out and taped up the window to the wing.) Twice we went up to shoot and twice I got some very bizarre vibrations from the engine for most of the footage. I noticed that the vibrations correlate to the rpm of the engine. I noticed more blurriness when using OIS (with my experience). I also figured the vibrations were transfering from the floor to the tripod, which I could help fix by placing the tripod on some hi-tech foam. I shot all manual- 24p Shutter at 48 with no filters (hood off). Yes, 80-90 Mph winds are tough to handle, but there are times the plane can slow down enough for it to be tolerable for the camera and my arm. In your situation where water and bugs are around, protect the camera in some way without creating too much wind resistance. Also, make sure to take the hood off of the lens. That will lessen wind vibrations.
My guess on your trouble with vibration is that the camera needs to come in more on the wing. It might be less shaky that way. Even if the prop is in the way, you can experiment with shutter to avoid the prop. Better yet, could you mount the camera under the airplane?
For strait down shots, I was thinking you could modify the plane to have a small window in the floor of it. Is this possible without disaster? I thought that would be nice for those kinds of shots.
Keep up the good work! I think your first round is quite nice considering all of the hurdles involved.

[Larry Huntington]

Oh, and to help in post, I have had good results using the new smooth cam feature in FCP6. You may want to experiment with that. It will crop your image but sometimes it does quite well in bumpy situations.

[Eric Muehling]

Michael, GREAT idea about the underwater housing.
Some of my preparation time was spent taping the
openings and seems of the XH-A1 so high speed wind
would not impact the camera's internal tape drive.

I spoke to the pilot today. The next time we do this we
will spend the first 1/2 hour or so flying at various
engine RPMs as I observe image jitter. We'll find the
engine RPM that produces the least jitter and fly with
that (providing it's a speed that will keep us airborne).

[Nathan Quattrini]

Hows the progress? If it goes well before next april do you think you would be able to help me with this?
http://dvinfo.net/conf/showthread.php?t=103450
I am also using a canon A1 so footage would easily integrate with my film....thanks

[Eric Muehling]

I don't intend to fly again until next spring (2008).
So there won't be any more to report until then.

I have concluded that most of the vibration problem
was due to the engine rpm, and on this light airplane
the engine vibration transfered easily through the wing
strut to the camera. Air turbulence was also a factor.
The fabric wing on this small airplane was simply
too bouncy to mount the camera.

The next time I plan to mount the camera on the
airplane's floats in hopes that location vibrates less.

I'm also considering another mount. I'm looking for
8-inch diameter PVC pipe. I plan to cut it 2-foot long
and slice it twice length-wise.

Next, I'll wrap the XH-A1 in foam packing and
close the PVC clamshell around the camera, then
wrap the pipe with duct tape.

That will provide a solid shell around the camera
to mount to. This will encase the camera in more
vibration-dampening foam than before. As you can
see, I enjoy tinkering and experimenting.

[Dave Morris]

I also shoot aerial video, but mine is done from inside a Mooney at higher speeds. My initial reaction upon watching your clip was "hey that was shot from a helicopter", i.e. low-RPM engine vibration. But my next reaction was "hey that's wind vibration".

Vibration can really plague airplane owners. One of my friends is currently battling a problem where his compass vibrates so badly that the liquid inside it is frothing. Mechanics can't seem to fix it.

But you might strongly consider the possibility that it's not the engine creating the vibrations, but rather the wind. If your mount is not completely 100% rigid (and it is impossible to make it 100% rigid!) then the wind can cause a tiny amount of oscillation. In this case, the faster your airspeed, the higher the vibration frequency. The problem is that the engine speed is linked to the airspeed, so how do you know which problem to attack?

There is a trick. You might be able to differentiate wind effects from engine effects by having the pilot slow the plane down to a very slow speed, pull the nose up, and rev up the engine while keeping the speed very slow. This is typically called "flight at minimum controllable airspeed", and every pilot knows how to do it. It will change the whole dynamic so that the engine is running just as fast as normal, but the wind effect will be drastically reduced. If you find the vibration changes, then it may be the wind doing it. At the other end of the spectrum, you could have the pilot create a high airspeed, low RPM condition by pulling the power and lowering the nose to pick up speed. See what difference that makes to the vibrations.

I'm going to want to talk to you some more offline, as there aren't too many of us using the XH-A1 for aerial videography, and we might be able to share tips and experiences.

Dave Morris

[Chris Hurd]

Thanks for your post Dave, please discuss this topic online right here, for everyone's benefit -- much appreciated,

[Bill Pryor]

I've done some aerials over the years, both from small and large helicopters and small planes.The most important thing I found is the pilot. If you get a pilot who is experienced at aerial photography flying, you're halfway home already. One time I was having difficulty from a helicopter with the wind, and the pilot went way upwind and tilted over and sort of drifted back over the site, and it was nice and smooth, moving with the wind at a perfect speed.

My single most important piece of equipment for aerials is...a barf bag.