My Tips on Getting Jello-Free Video wtih GoPro 3

jp3556 said:
Exactly. So your question is, which provides better results - faster shutter speed or slower shutter speed, either being achieved by using ND (to slow it down) or frame rate (to hopefully speed it up). My very limited tests seem to conflict with the others here so I'd assume using ND to slow down the shutter is the way to go.

Rolling shutter due to CMOS chips: http://www.diyphotography.net/everythin ... ng-shutter

Can I see you test? How can we not have a definite answer to this? Higher or lower shutter speed is better for rolling shutter?
 
justsomeguy said:
yalag said:
Can I see you test? How can we not have a definite answer to this? Higher or lower shutter speed is better for rolling shutter?

I get that you're pretty lazy but I'm not understanding why you are incapable of doing your own tests. You have a Phantom and a gopro. Get to work and figure it out on your own already.

Hahaha.

@yalag, since you can't change the shutter speed on the GoPro anyway, it's not really worth debating - the results are all that matter. So let's just say that a good place to start would be to use ProTune at 60 fps and with an ND filter when shooting in bright lighting conditions which apparently helps with 'prop flutter'.

I'll try to post some tests once I get my gimbal.
 
jp3556 said:
Hahaha.

@yalag, since you can't change the shutter speed on the GoPro anyway, it's not really worth debating - the results are all that matter. So let's just say that a good place to start would be to use ProTune at 60 fps and with an ND filter when shooting in bright lighting conditions which apparently helps with 'prop flutter'.

I'll try to post some tests once I get my gimbal.

Cool thanks jp! I have that troll blocked, maybe you should too..
 
justsomeguy said:
jp3556 said:
..since you can't change the shutter speed on the GoPro anyway, it's not really worth debating - the results are all that matter.

You're missing the point, again. You can change the shutter speed by shooting with the same settings in different light conditions.

I suppose I should have said that "in the same lighting conditions, you can't change the shutter speed on the GoPro anyway". The fact that it changes under varying lighting conditions is a given as that's the way it adjusts exposure.

My tests in the same conditions seemed to produce better results at 60 fps than 30 fps, which I assumed (evidently incorrectly) was because it was forcing it to use a faster shutter speed, which is one way you maybe *could* have some manual control over exposure. But another way for sure is to use an ND filter and force it to go the other way, as you pointed out.

DSLR's using CMOS chips I'm pretty sure tend to be prone to more rolling shutter effect at slower frame rates (could be wrong though) and you'd typically be shooting at around double a frame rate's shutter speed to achieve natural motion blur. Having said that, lower shutter speeds may help mask (but not eliminate) rolling shutter simply because of the added motion blur. To be honest, it doesn't make sense to me since using a higher shutter speed should reduce the amount of time for vibration and therefore the rolling shutter to affect each frame, but I submit on this point - the results seem to show otherwise.

justsomeguy said:
Re: ND filters eliminating prop flutter, that's only because they act as a lens hood preventing the prop shadows from hitting the lens. A piece of tape stuck on the top of a bare lens will accomplish the same thing.

That's true, as that video points out as well that you mentioned below. Prop flutter really is just from shadows (since you probably can't actually see the props in the shot).

Thanks for adding to the thread! Was merely hoping to build some 'best practices', since I had a hard time finding this information when I first started a few months back.
 
The ND filter is how you control the shutter speed, the more light the darker the filter you need to reduce shutter speed and therefore also reduce jello effect. With no filter you have no control over the shutter speeds and in very bright light will get excessive jello if you have any vibrations and also the video will have a strobe type effect.

Prop flutter is quite simple to see that its caused by the shadows off the props, use a ND filter and the props get blurred to the point they are almost not visible, no nd filter and bright sun the props will be seen clearly with very high shutter speeds as full blades but with rolling shutter distortion. When you see a full clear prop it will also cast a shadow that is exactly the same as the prop unsurprisingly. If the shutter speed is slowed to the point that the prop is blurred completely then its shadow will also do the same and so eliminate prop flutter.

I dont know if the faster frame rates also increase the sensor rolling shutter scan rate and this could be possible as faster frame rates show less jello in the same light conditions as 30p or less in general. This kind of goes against the ND filter as a slower shutter speed will blur the micro vibrations shown as jello. But with the faster frame rate it appears the scan speed is quicker but an ND filter will still help even at faster frame rates. Though shooting at faster rates you cant go too dark a filter as this may get it to a point where it doesn't get enough light at 1/120 shutter speed in 60p so the ISO and gain will be upped creating more noise in the image.

Best way to avoid jello is balance your entire setup and isolate vibrations getting to the camera as well as you can, then also use a ND filter that can get your shutter speed to double the frame rate but not so dark that the camera needs to up gain and ISO to avoid underexposure, with these you can shoot in any frame rate and lighting condition pretty much jello free but you need to understand the effects of faster frame rates on min shutter speeds and so will need lighter filters at higher frame rates and in different light conditions.

Here is one of my videos comparing using a ND filter and no filter.

[youtube]http://www.youtube.com/watch?v=E34obz9qcTw[/youtube]

This was shot in 2 consecutive flights but light levels were changing constantly as the sun went in and out of clouds. I did shoot some side by side tests but had some bad vibrations on my mounts so the jello was worse then using a single camera. I have improved my 3d setup since and will redo another side by side ND vs no ND when we get a bright day.

Here notice the props without ND filter, see how this creates prop flutter when flying back towards the sun yet with the filter the props are blurred and so the flutter isn't really visible.

Here is one of my side by side Filter vs no filter videos ungraded in cam raw protune 1440p30 on both sides. The mount was shakey and this caused some jello on both sides. If your mount has wobble in it this is the first point of jello so balancing everything and using a filter with lots of wobble in the mount will either cause excessive blurring or bad jello. The filters only work if you have just tiny high frequency vibrations if its vibrating like crazy nothing is going to help reduce the jello. ;)

[youtube]http://www.youtube.com/watch?v=mMZ-UvIEEto[/youtube]

Here is a comparison between a polar pro ND3 and tiffen ND.9 filter, raw and ungraded both are 3 stop ND filters and it shows how crappy the optics are in the polar pro filters here I think. Including the weird reflection these filters get whenever the sun is very bright which is kind of when you need the filters the most so I wouldn't use the Polar Pro filters other then on a few test flights.

[youtube]http://www.youtube.com/watch?v=D_K9KTCkNqs[/youtube]

The filters will slightly soften the image as they reduce light and also add some motion blurring, I much prefer smooth video with some motion blur compared to the stuttery high shutter speed motion in very bright light. Stills will be better without a filter but for video to play smoothly some motion blurring between frames is preferred. Though filters will soften the image somewhat poor filters will be much worse then better quality ones.
 

Members online

Forum statistics

Threads
143,066
Messages
1,467,354
Members
104,933
Latest member
mactechnic