auto focus

[Craig Schober]

posters have been tossing around the idea of an autofocus system for a while. here's a piece of the technology that might help.

http://www.technologyreview.com/Infotech/19476/

so now we would have angle, speed and direction all transmitting wirelessly back to lens. now all you need to do is calculate distance from behind ear to eyes and distance from device to lens. set your racking speed curves via software or on-the-fly using something like a red grip.

you might not be able to replace a good focus puller but you sure could replace a bad one.

[Erik Widding]

posters have been tossing around the idea of an autofocus system for a while. here's a piece of the technology that might help.

http://www.technologyreview.com/Infotech/19476/

so now we would have angle, speed and direction all transmitting wirelessly back to lens. now all you need to do is calculate distance from behind ear to eyes and distance from device to lens. set your racking speed curves via software or on-the-fly using something like a red grip.

you might not be able to replace a good focus puller but you sure could replace a bad one.

This is really not that useful. The three dimensions of motion that it can track are: relative to the person, not the camera; incremental, not absolute. It would require three additional axes (all angular, using gyros or a magnetometer) to get the pose information that you suggest - and one is still left neeeding to know where the person is relative to the camera, which requires another six axes of pose information.

[CVB]

Using accelerometers for position feedback is very sketchy with todays sensors. The accelerometer gives acceleration which has to be integrated two times to get position... after all is said and done you'd be lucky to be within +/- 10 feet after a couple minutes of use.

[Craig Schober]

i'm no engineer but i was assuming wireless latencies can be measured to determine distance. it works with gps, granted, at least 3 satellites triangulating but distance measuring has also been used successfully in indoor microcell networks and even bluetooth applications. i guess it's pointless unless accuracy gets to +/- 1 cm.

[John V]

The times that would need to be measured with a wireless signal are far to fast to find distance