Tau Labs Camera Stabilization
- Thread starter Gary Mortimer
- Start date
Let's hope there are no casualties I'm intrigued by this advice. I'd expect the opposite to be true. I get the thing about Newton's Law and flinging a heavy camera around, but most of that should be controllable with good camera CG balancing and sturdy gimbal construction, no? In techie terms the camera itself, when rotating, should have the lowest possible Moment of Interia, right? After that, I'd have thought the objective would be to drive the thing as fast as possible - you're already "behind the game" by the amount of time it took to compute the aircraft's attitude and issue the appropriate commands to the servos.
You have to start with the model and work backwards. My camera system is a stand alone device, it does not use the FC outputs. First you need to understand why camera movements above 5 degrees per second will produce blurred pixels which do not respond very well to post stab. software, if you have to rescue poorly shot footage. A pro camera system works on an inner and outer axis principle where the inner axis is caged to stay within less than twenty deg/sec so that very high sensitivity gyros can control it. Such gyros would saturate and lock up the system if you start trying to fly like an idiot. My camera ship will not fly in acro mode. It can fly fast but all rates of movement in pitch yaw and roll are very slow. The models FC gyros have a smaller Full Scale Range and are about 25 times more sensitive than the ones in your average model. Thus, the inputs have to be limited so that you simply cant over control it. Otherwise it would crash. That is not to say that the gyro response is also slow. It is in the order of 50mv/deg/sec as opposed to say 2 mv/deg/sec from something like an AD 610 fitted to an MK. Deadband and noise are in a different world.
In our model pro mount we use the model itself to create that outer axis. So the model is the first stage of keeping things slow. It gets much more complex but I have to start with the basics. Now, you may not be interested in a pro system such as mine but the principle is still very similar, if you plan to use the FC outputs to control your camera. You need to restrict the turn rates of the model, kill the vibration by creating mass on the camera base plate, Keep all camera movements very slow. Then if you use a decent image stabilised camera it has a chance of working like my inner axis and you should be able to get pro level quality at wide angles of view.
In our model pro mount we use the model itself to create that outer axis. So the model is the first stage of keeping things slow. It gets much more complex but I have to start with the basics. Now, you may not be interested in a pro system such as mine but the principle is still very similar, if you plan to use the FC outputs to control your camera. You need to restrict the turn rates of the model, kill the vibration by creating mass on the camera base plate, Keep all camera movements very slow. Then if you use a decent image stabilised camera it has a chance of working like my inner axis and you should be able to get pro level quality at wide angles of view.
Last edited by a moderator:
peabody124
Member
Regarding the "being behind the game" bit - one thing that can be done with OP Pro (maybe not as easily with CC) is actually use the control signals in the gimbal control, so you are actually predicting the change.
Also it's quite easy to separate the maximum roll rates in our system from the control gains them self, so it can be tightly locked but not roll very quickly overall.
We've also discussed routing a camera trigger (for still shots) via the board so that when you take a photo for 100 ms before and after the aircraft comes out of stabilization mode and goes into a high gain zero rate mode. (This will be a way off, more fun with ESCs first)
Also it's quite easy to separate the maximum roll rates in our system from the control gains them self, so it can be tightly locked but not roll very quickly overall.
We've also discussed routing a camera trigger (for still shots) via the board so that when you take a photo for 100 ms before and after the aircraft comes out of stabilization mode and goes into a high gain zero rate mode. (This will be a way off, more fun with ESCs first)
There certainly is a lot of work going on to improve electronic ACTIVE stabilisation, but a lot of work needs to be done with PASSIVE aerodynamic stabilisation. In order for the gyros to make fast corrections we need higher prop efficiency. Currently most FC's that I have tested need a little inefficiency (smaller dia.) to mask the deadband. Holger has also started working on this with his wedges to remove mechanical deadband in the yaw. Someone should start working on auto switchable gyros to get the best of both in terms of FSR. Fibre Optic gyros already deliver by far the best solution but the cost and being Mil. spec. make them prohibitive at this moment in time.
Last edited by a moderator:
I put this in the wrong place!
http://vimeo.com/26002335
On another topic anyone tried these G6 lipos? they look pretty good.
http://vimeo.com/26002335
On another topic anyone tried these G6 lipos? they look pretty good.
ZAxis
Member
There is an alternative to electronic and passive stabilisation ...
see http://www.youtube.com/watch?v=adlgpovEv7g
see http://www.youtube.com/watch?v=adlgpovEv7g
Droider
Drone Enthusiast
Maybe I'm a super chicken!.. I have tried to mimic the chicken head while filming with head cams for years. This short movie was taken down a really steep track in Scotland. the second half when I can get some speed up gets smoother filming but the ground is just as punishing . if you watch the top pro downhill mountain bikers there upper body and especially the head seems like its floating on air..
http://www.youtube.com/watch?v=krnKtD0l5RU&feature=fvsr
how this can be moved to this science is for people who can think out of the box!
http://www.youtube.com/watch?v=krnKtD0l5RU&feature=fvsr
how this can be moved to this science is for people who can think out of the box!
Last edited by a moderator:
Gary Mortimer
Member
ZAxis, one million chicken points that was great.
Bartman
Welcome to MultiRotorForums.com!!
ZAxis
Member
Emowillcox
Member
Emowillcox
Member
I am curious if anyone has tried putting one of the flight stabilizers on a camera mount. I used one on my helicopter to help keep it level with the horizon when shooting so just seems like the same system might work well on a camera mount.