Vibrations and Jello effect causes and cures.

by TommyGunn | May 14, 2013 | (36) Posted in How To

I have been exploring ways to make my videos smoother using my Gopro.
One of the thing that plague fpv flyers is vibration and what we call the “Jello effect” which is really In the correct term it is called Rolling shutter.

What is rolling shutter and what causes it?
NTSC video is normally 29.97 frame per second our uk and over seas friends use PAL which is 25 frames per second.
Each fame that is captured is not capture in one shot.. But captured by scanning very rapidly ether up and down or left to right.
ok now remember were going to talk about one frame of the 29.97 frames in that one second.

here is how rolling frame shutter is cause...

RectifyingRollingShutter

In that one frame we just talked about...when you have a fast-moving object... the camera is scanning left to right going from the top to the bottom very quickly.

While is doing that the scene is moving so while is is scanning the top part of the video halfway down the scan the scene has moved a little.

Rolling_shutter_01So for the second part of the video it is starting to scan " moved part of the image and it is trying to glue those 2 parts together.

think about cutting a picture in half.. then slightly misaligned it and taping it together.

Now do that with 29.97 frames you gonna have a wobbly image.

VM_MakingMovies2_05

The cameras that suffer from this are all SLR's and Cmos cameras they all work using the scanning system.

CCD cameras do no have rolling shutter issues they take an entire picture in one shot instead of using the scanning method. There is not much in the way of stopping the "Jello effect" But you can calm it down depending on what camera you are using In my case I have the go pro 2 and a go pro 3.
With the go pro 2 setting it to 720p and 60 frames per second it will cut the jello down.

with the gopro 3 you can have higher resolution by setting it at 1080p with 60fps  or at 720p at 120 frames per second will probably give you the best results.

the faster it can capture the less jello you will have.

Vibration can add to this problem too But it is not the total cause of the "jello effect". The cause is fasting moving objects that the camera can not keep up with.

blades-1282772892

Vibration will cause shaky video, smearing blurriness. they are different way to test how much vibration the model is causing.

there tools like the IPhone ivibrometer that measure vibrations on the x, y and z-axis,

330500512_screen0360x480

 IVibrometer

you can use lasers and bounds then off you model and see how much the little red light moves.
You can also use strobe lights to stop motion and see what is off this is called stroboscope method by adjusting the speed of the strobe you can slow down the propellers and see what is off and also the same with the shaft.

 

 

 

Today I am going to try some different and something I have not seen anyone try yet to test the smoothness of the image....your eyes.

Yep I said it your are probably thinking i do that all the time when I fly I check my videos to see what is better! you knuckle head!

Well I am going to do it in a more controlled environment and not flying around the yard which would be impossible to duplicate and compare!

I will be doing this indoors and using an eye chart!

eye_chart

 with the eye chart I will be able to clearly see( no pun intended) what is really visually acceptable with this method.

The Parts I will be dampening the vibrations will be on:

The spacer's and frame.
Camera base.
Propellers.

The source of vibrations will always come from the motors.. So in my thinking I want to slow the vibration down at different point so by the time it gets to the camera there is hardly any at all.

First almost all frames have the arms attached to the frame like a "sandwich"

Point 1 propeller

Point 2 the motor mount

point 3 the arm

point 4 the frame

point 5 camera base

point 6 camera mount

Point 1

There many tutorials on balancing propellers online so I will not go in to details about it now I use the scotch tape method it take about 2 to 3 minutes for each propeller.

 I know it is a boring and a pain in the butt to balance each prop every time you fly. But who said you need perfectly balanced props every time you fly?  I suggest keep about 8 pairs balanced and used those for when you are filming if you are just flying for fun who cares if it vibrates a little use the unbalanced ones when you are just fun flying but if its really unbalance i sugguest use the balance props.

IMG_7870

Here is what I did at point 2 between the motor and the frame arm I used a thin piece of rubber dense foam rubber.

IMG_8009

 I cut this out to fit between the motor mount and the frame arm. Doing this Muffles the sharp vibrations slightly coming from the motors. Also use lock tight on the screws here, the screws here can become loose.

IMG_8013

 point 3 and 4

Spacer between the frame and arms.I had some heavy thin rubber tubing which I cut a tiny bit longer the spacer itself so it would overlap on both ends.

IMG_7875

 

By doing this when the spacer are between the frame and screwed down it the rubber tubing will be squashed between the frame enough to insulation the spacer and frame vibrations that would have traveled from arms to the frame.

IMG_7853

 

point 5

 At this point i added a 1/4 inch neoprene rubber pad to the camera base this should knock out any harsh vibration left and only adds about 1 oz to the frame..

IMG_7856

 

point 6. the camera base.

Just to really get an almost vibration free on top of the neoprene I made a camera mount. This was made out of sound proofing foam which is very dense which is a great insulation for sound and vibration not like a sponge which can be too bouncy.

Using the second layer adds a nice "cushion"

Slot to fit in back of the frame.

Ruberr bands and hair ties. Hair ties are much stronger than rubber band so camera will not go flying off in case of a crash.

Hair tie looped to back of base.

The Test

Frame: Hoverthings ht-fpv flip
Controller: DJI Nasa M
ECS: DJI opto 30A ECS
Motor: Readytofly.com standard ht-fpv motors.
Props: Gemfan 10.5 plastic props
Camera: gopro 2 set at 720 60fps.

I had already had placed the spacer and the mounts and the neoprene before writing this article I do have one video but it was not in a control environment.

I used 3 types of eye charts and some text for the test.

If i can read red text and see the lines and symbols clearly without shaking i am happy with results then.

Video was shot using a gopro 2 using medium wide and 720p I did not slow down the video or use any smoothing additives.

i was getting a lot of prop wash and ground effect so it was hard to keep the quad steady.

You will see how my quad used to shake and vibrate just look at the letters on the big sign on the before.

And you can see the difference of the after video. I did this video after i did all the dampening so i did have a "before video" using the eye chart.

Also look at the after video anf the hand held video and compare the two let me know what you think how close they look.

Here is the video

Hope this helps you guys if i missed something please add it the comments and I will updated this post.

I have also started a blog with all my projects and stuff I have done with lots of videos.

it is at http://www.tommy-gunn.com

Thanks for reading!

Tommy Gunn

 added notes:

A couple of things to remember  use lock tight on the motor mounting screws even if slight loose they will cause the motor to be unbalance and possibly come off. don't get lazy always check mounts before and after flight to check if they loose with hex driver or what ever you use to mount the motors.

Check the prop adapter cones before every flight make sure they are tight you don't want them flying off in the middle of a flight. ( ask me how i know this..)

if you have a crash check to see if any of the motor shafts are bent using the stroboscopic method I will do a video on how to do this soon. i always have spare prop adapters they cost between 1.25 -2.00 cheap enough to have few to keep in your repair kit.

 

COMMENTS

RPflyer on May 18, 2013
Very good !
Log In to reply
FlyingMonkey on May 18, 2013
Very impressive, especially since the "After" video was shot indoors, low light and florescent lighting at that, by the looks of things.
Log In to reply
TommyGunn on May 18, 2013
Yea the prop wash was making it bounce off the walls and floors.
Log In to reply
windfou on May 18, 2013
Tommy the best to reduce vibration is stainless steel cable suspended and you can buy it at http://www.ebay.com/itm/WHITE-DJI-Phantom-GoPro-video-vibration-isolator-cut-jello-multicopter-/261198086765?pt=US_Radio_Control_Control_Line&hash=item3cd09e826d
Log In to reply
Cyberdactyl on May 19, 2013
One thing to also remember. . . every geometry is different, even on a similar frame, thus requiring a slightly different method and/or medium to nullify the higher mechanical frequency vibration.

But here's another approach to find. . . and real time test, for jello. And unlike the novelty of using laser bounce, this provides instant video feedback.

I have the HoryzonHD Foxtec V3. It has the camera separate from the processor, connected by a 300mm stainless steel ribbon covering the video wires. I can also, real-time, video-out to a larger screen on my laptop.

I rubberband or tape my camera to the boom while balancing my motors. It is an amazing indicator for even the slightest vibration. (a prop balancer is plenty good enough for the props)

One caveat, I would suggest you don't connect the camera, if you know the motor is extremely unbalanced, as it might damage the sensor.

I've thought a lot about putting together a video studying different jello sinusoidal wave paths and how that correlates to outrunner bell vibration, a bent shaft, bad bearings...allowing precession and nutation.

Correlating those patterns to the vibration host would make a great thesis. The huge catch is those "jello waves" would be for a highly specific geometry, so it would be WAY too complex a study for anything other than one condition.
Log In to reply
TommyGunn on May 19, 2013
Glad you mention bent shaft.. I will add this to the article.
couple of things to remember use lock tight on the motor mounting screws even if slight loose they will cause the motor to be unbalance and possibly come off. don't get lazy always check mounts before and after flight to check if they loose with hex driver or what ever you use to mount the motors.

Check the prop adapter cones before every flight make sure they are tight you don't want them flying off in the middle of a flight. ( ask me how i know this..)

if you have a crash check to see if any of the motor shafts are bent bent using the stroboscopic method I will do a video on how to do this soon. i always have spare prop adapters they cost between 1.25 -2.00 cheap enough to have few to keep in your repair kit.
Log In to reply
TommyGunn on May 19, 2013
windfou:
the steel cable is interesting but it seem to only work "by hanging" is that correct? and adding like 3 addition inches. which means you will need longer landing gears.I just prefer my cams on top close to to the center of the quad. Ycopter has similar version which is a lower profile and uses plastic tubing which i would thing is less prone to vibration than metal but I could be wrong.
I think I have eliminated all my vibration look at after and hand held video hard to tell the difference with the sound off.
Log In to reply
Cyberdactyl on May 19, 2013
Another technique I have NOT tried, so this is pure conjecture, is to use the battery as an inertia vibration ABSORPTION mass. If anyone has tried to record video from an extremely small camera, such as the 808 #11 or #16 or the Foxtech V2 or V3, will be aware of the difficulty they present in having so little mass.

Having the camera and battery on the same vibration mount (which does the isolation), in theory, should utilize the high available mass (which does the absorption) for higher mechanical frequencies.

I get a good laugh whenever I read of someone distressing their relatively large GoPro is extremely sensitive to vibration.
Log In to reply
TommyGunn on May 19, 2013
Indeed Cyberdactyl light cameras do have a bit of harder time with vibrations.
yes you idea sounds good on concern depending the size of the battery and where you place battery and cam is CG.
i would try a block of sound proof foam then cut a slit in it to have a tight fit for the camera.
then pc board and glue 4 memory ear plugs side ways to it then glue any another pc board on top of that like a sandwich which would be about 1/2 inch. then on top of then you will need a little weight i would say like 3 aa battery's. you could put them in one of those black battery cases then glue the foam on top of that.
The whole rig would be about 3-4 inch tall and i think that would work and also protect you cam too.

I have my battery on the underside of my quad and i use a couple of strips of neoprene and velcro
and a Velcro strap .
i have velcro on all my battery's for a couple of reasons
1 safety no flying batterys
2 easier to keep battery cg in place
3 with neoprene and velcro no change of vibration.

Log In to reply
Cyberdactyl on May 20, 2013
David of Flitetest had the solution a while ago with his tricopter design putting the camera and battery on the same platform suspended by arched wire.
Log In to reply
TommyGunn on May 21, 2013
Yes David's was a very good solution slowing down the vibrations thru a thin piece of wire.
I thought you meant putting the camera on top of the battery my mistake.

Log In to reply
TommyGunn on May 21, 2013
Addendum:
for some reason I had forgotten to insert Images for the last part of the camera mounting part 6 I have updated it now. Sorry about that,
Log In to reply
pramodgupta on October 18, 2013
Nice
Log In to reply
ChrisJ on April 12, 2014
I'm pretty sure vibration does not cause the 'jello' effect, though it can exaggerate it. Even a digital video camera on a solid tripod stood on concrete will have 'jello' shutter when recording any moving subject unless you spend a fortune on a camera that has a very fast processor to reduce the scan time (the time the processor requires to serially output the frame data to the A/D converter and save it) to 1/1000ths of a second.

The scan time for most of the video cameras which are light enough to carry on a model (including the overpriced GoPro) is around a rather pedestrian 1/16th of a second. Look closely at every frame in a video editor and you will see several identical frames if you record at 30fps (usually 2), and twice as many if you record at 60fps.

Another partial option to reducing 'jello' shutter is the opposite which is to force the camera to record with slow shutter speeds by using an ND filter which reduces the shutter speed as the camera reacts to the lower light level, which records fast movement such as a spinning prop as a blurred disc, much as our eyes see it.

Chris
Log In to reply
TommyGunn on April 12, 2014
Chris I don't think I said in my article that vibrations caused the rolling shutter effect " jello"
I said "Vibration can add to this problem" making videos blurry.

I think I went over the gopro 60fps 120fps

ND filters thanks for bringing that up yes nod filters are good for that even thought it will darken the video some what. At the time if the writing there was no real ND filters for our small cameras. But I think they make them for gopro now.

Thanks Chris for the reminder for ND filters!

Tommy
Log In to reply
TommyGunn on April 12, 2014
Oops... Auto spell check messed up nod=ND
Log In to reply

You need to log-in to comment on articles.


Vibrations and Jello effect causes and cures.