GH4 Magic Mini-Box That Stabilizes Your Footage in Post... :-)

TheDingo said:
My guess is that the "magic" box is recording three dimensional movement data along with your camera's audio track, so that the audio track will be used to sync the 3D info with your video, and then the 3D stabilization software will automatically manipulate your video image to get rid of all the shaky camera movement. This also means that there will be a little bit of cropping to the video image in order to stabilize your subject, which is why they recommend shooting with a 4K camera like the GH4, and probably shooting with loose framing so you won't lose anything important in the final image.

I think it's a very cool idea provided it's reliable and doesn't take a ton of processing in post to produce the final video.
Click to expand...


What is the advantage over crop image stabilization in post?
Great idea but a "steadycam", while bulky doesn't crop & reduce quality.
 
ralfmouth said:
What is the advantage over crop image stabilization in post?
Great idea but a "steadycam", while bulky doesn't crop & reduce quality.
Click to expand...

Great question

I am only guessing but I am betting there is some sort of motion sensor in the box which then aids the software in making it work more efficiently and effectively. This would explain why the box needs to be attached to the camera. An enterprising Camera company should buy the technology and just integrate into their cameras.
 
or or or.... the GH4, which already has a tilt detector could output metadata or something in addition to the footage
 
This looks incredible, if it works anywhere near as well as they want to indicate.

Even if it's not perfect or as a good as a steadicam, for $300 and THAT LIGHT AND PORTABLE - You can pack this in a backpack and achieve these sorts of shots in all kinds of situations without fatique. Even if this just acts as a sort of (very) premium version of adobe warp stabilizer without the weird distortions, etc... that IS a game changer. Warp Stabilzer just can't handle running footage, etc. Pulling off quasi dolly, jib, and slider motions and walking/tracking shots with even 70% of the results of a steadicam would be killer. Especially with the room to spare that 4K offers. And espcially if you could run it through their stabilization reconstruction, render, and then run again through Adobe Warp Stabilizer if needed...

Count me as highly interested, a little skeptical, and incredibly hopeful
 
Thinking about it now there really are no downsides besides the extra step in post which if it's simple enough I'd really be into this.

I'm shooting at 4K on the GH4 so cropping doesn't matter to me, the real advantage for me is to keep it light and simple getting smooth shots handheld without extra hardware while still being able to rack focus which I cannot do on a steadicam or will be tough and more expensive to do on a brushless gimbal system.
 
I want to make a point about those that will say, "this doesn't compete with a steadycam or a shoulder rig". In case you didn't watch the video, one of the clips is a POV shot on a bike. The steadicam and shoulder rig is absolutely worthless in those situations. Before the steadycam there was what? And what was before that?

This device isn't saying, "Go out and throw your camera around and it will all be smooth with our tech". You can still use this device on a steadycam, a gimbal, a car mount, helmet cam, all sorts of things to "record" the camera movement. You can then "choose" to use than info or not. Tell me. How many of you have used post stabilization? I would assume all of us.

So since I believe everyone raised their hand to the question of how many of us have used post stabilization here is how I see this unit working for me (if I indeed see that it works to my expectation). I would rig this on the camera at all times. Then, for shots that do in fact "choose" to stabilize I would use this devices data over some built in NLE stabilization. Not only would it probably speed up the process because the NLE doesn't have to analyze the footage, it would result in better post stabilization. Think about it. How many times have you used an NLE's stabilizer and found that the action in the frame throws off the analyzed result? Be it a person walling through frame or a subject moving towards camera. I would think this device reads the accelerometers built in and writes the data for that shot. It won't care if a tree branch enters your frame which confuses the NLE analysis.

Why worry? Its just a tool. Lets see how effective it is. Just another way and tool to open up the possibilities of getting that shot.
 
themac said:
Thinking about it now there really are no downsides besides the extra step in post which if it's simple enough I'd really be into this.

I'm shooting at 4K on the GH4 so cropping doesn't matter to me, the real advantage for me is to keep it light and simple getting smooth shots handheld without extra hardware while still being able to rack focus which I cannot do on a steadicam or will be tough and more expensive to do on a brushless gimbal system.
Click to expand...

Why do you call it an extra step? Isn't checking "stabilize" in your NLE an extra step? In fact I would call this device one less step because you are analyzing your footage while shooting. All they need to do is add a plug-in for the NLE for your choice and I would think easy as pie.
 
Remember, the Instagram app Hyperlapse already does this, so that's your tech demo right there. It won't work with the GH4 on its own because cameras don't have gyroscopes, only a simple level mechanism.
 
dustylense said:
Why do you call it an extra step? Isn't checking "stabilize" in your NLE an extra step? In fact I would call this device one less step because you are analyzing your footage while shooting. All they need to do is add a plug-in for the NLE for your choice and I would think easy as pie.
Click to expand...

Thinking about it now it wouldn't matter as there are always extra steps involved no matter what. Even if I were to use a glidecam or gimbal the extra step is balancing the thing or setting up the thing plus dealing with the extra weight and logistics to carry that extra hardware.

I wish the camera makers would incorporate this into their cameras, built in gyro sensor and have the processor do the stabilization work in camera when I want to, that would be amazing.
 
Yes also wondering if this isnt just the same as Warp stabilizer. If it is almost the same I will not buy it. The problem with software stabilization is that it cannot correct the micro jitter, when you mount your cam to a glidecam or gimbal. It does not have that microjitter video shake. As I have no intentions using a higher shutterspeed then 50.
 
Wouldn't it be nicer to discuss something for what it is and not for what it is not? Or what may be lacking? Espresso-module for instance? 300 $ and it looks as if it can do some marvellous things for not so much money. Easy decision: buy it or not.
 
Trouble is that you get motion blur in frames which is unrapairable (or you have to shoot at high jitter shutter)

I can sort of see some aplications however, for example stablising steadicam footage to make it very pefect
 
This device could also help with compositing....should I call the device a "live Camera Solver"? Could the data (keyframes) be used to help composite 2D and 3D elements into a scene? Many times camera solving in post can be tricky. Maybe this device could produce more accurate results? I'm all for that!
 
I'm interested. If the developers are reading, I would love to see video in the best and worst case scenarios.
 
dustylense said:
I want to make a point about those that will say, "this doesn't compete with a steadycam or a shoulder rig". In case you didn't watch the video, one of the clips is a POV shot on a bike. The steadicam and shoulder rig is absolutely worthless in those situations. Before the steadycam there was what? And what was before that?

This device isn't saying, "Go out and throw your camera around and it will all be smooth with our tech". You can still use this device on a steadycam, a gimbal, a car mount, helmet cam, all sorts of things to "record" the camera movement. You can then "choose" to use than info or not. Tell me. How many of you have used post stabilization? I would assume all of us.

So since I believe everyone raised their hand to the question of how many of us have used post stabilization here is how I see this unit working for me (if I indeed see that it works to my expectation). I would rig this on the camera at all times. Then, for shots that do in fact "choose" to stabilize I would use this devices data over some built in NLE stabilization. Not only would it probably speed up the process because the NLE doesn't have to analyze the footage, it would result in better post stabilization. Think about it. How many times have you used an NLE's stabilizer and found that the action in the frame throws off the analyzed result? Be it a person walling through frame or a subject moving towards camera. I would think this device reads the accelerometers built in and writes the data for that shot. It won't care if a tree branch enters your frame which confuses the NLE analysis.

Why worry? Its just a tool. Lets see how effective it is. Just another way and tool to open up the possibilities of getting that shot.
Click to expand...

^^^What THIS guy said....to the letter!
 
Dustylense is bang on.

Its not about replacing all other stabilization but adding another tool to get the shot you have in your mind which all film makers should be excited about and at the proposed price bracket us lower budget commerical and film guys should be really excited about this. Especially for stealing shots when needed.

I just shot some branded travel content and would have loved to have this device for the low profile airport/In transit shots. It opens up creative possibilities and it the creative that keeps me going in this industry.
 
You must log in or register to reply here.
Back
Top