meta data meta physics?

[JD Holloway]

I was thinking about Cookes meta data lenses...

Does anybody think a third party will develop a 3D 3-axis accelerometer with a simple output that will plug into RED to give data for camera motion, recorded on frame by frame basis?

Just a thought.

J.

[Darren Orange]

at some point sure....right now i find that for tracking purposes boujou 4 works really well...

[Mark L. Pederson]

check out PF TRACK 4 at NAB and keep your eye on ImagineerSystems

[Jag Sandhu]

some more info on PFTrack software synced to 3D camera data on a frame by frame basis for postproduction tracking purposes.

http://www.cookeoptics.com/cooke.nsf.../cookepixel207

[CVB]

JD, We are thinking about developing something like that... the biggest problem with accelerometer based systems is drift. Positional data ends up getting extremely inaccurate after a minute or two. That drift is basically due to noise and temperature effects and it gets compounded when the accelerations are integrated to get position values. This isn't a huge problem on large scale movement but for small scale <1" the data is useless. There are a number of companies that are working on making tracking systems that are accurate to <1cm, portable, and inexpensive. Unfortunately the technology is still a couple years away from the mainstream.

[Ace]

JD, We are thinking about developing something like that... the biggest problem with accelerometer based systems is drift. Positional data ends up getting extremely inaccurate after a minute or two. That drift is basically due to noise and temperature effects and it gets compounded when the accelerations are integrated to get position values. This isn't a huge problem on large scale movement but for small scale <1" the data is useless. There are a number of companies that are working on making tracking systems that are accurate to <1cm, portable, and inexpensive. Unfortunately the technology is still a couple years away from the mainstream.

Curt. What if on each take, you set up 6 "nodes" around the set. These would be portable thumbdrive size transmitters equidistantly seperated. From that the unit on the camera would recieve timecode transmitted from these nodes accordingly, and measuring the latency you could pretty much create coordinates of the cameras location on set. Ive just described GPS on a micro scale. A mobile GPS setup.

[CVB]

Ace, the tricky part with the time of flight solution is developing a device that can measure time on the scale of picoseconds (thats trillionths of a second). Its actually a bit harder than regular GPS because of the shorter distance invloved (20 feet vs 20,000 miles). Most of the work I've seen in this area is being done at a university level and some of the technologies have moved into startups... so its still a couple years out. All of this depends on the accuracy you are looking for - for tracking a camera I think you need something that has a resolution of around 1 cm or better to approach the accuracy of match moving. There are tons of systems on the market that can locate object using the same method you are talking about.. but they are alot of $$$ and they have a resolution of about 1-2 meters.

[JD Holloway]

mmm
I gotta think about it more.
Thanks Curt.
The local GPS was an idea too. Works well enough for GEO Analysis, might be able to adapt it to distances of less than 100 foot.

J