Hi,

I have been very deep thinking how to produce skeletal animation capturing. I came up with an affordable idea but I wonder wheather it will (can) work.

Here it is:

4 camera setup, in a square, 1 camera in the middle of left side of the square, second camera opposite- in the middle of right side od square.

The other two cameras are placed just the same way, at front- back side of the square.

I would capture the videos, and by using four monitors, I would run all 4 captured videos synchronously. Then, frame by frame I would place the bones of the animated object to fit the actor pose every frame.

I wonder though wheather this is doable. I have few concerns:

markers? Would I need to place markers on actor for being able to accurartely analyze a curent pose? How many of them? Two per bone (the two ends of bone)?

Will I need an "orthogonal stick" helper on a bone marker, to be able to analyze its rotation around itself of it occurs (for example turning a key in a lock is such a rotation)?

I am confused about markering setup, what markers should be like to help me?

But mainly, is my solution doable and can produce an animation asset if I work hard? On the cost of it not being the on-air actor-to-animated display.

Maybe Im wrong but cant you use an Xbox 360 or Xbox one Kinect to get motion data. Seems easier that way and cheaper then buying 4 or 5 cameras.

thanks guys but I just want criticism towards the solution I promise myself to. Is it all cool, instead of it not being real time previewable, but, AS I HOPE, being well snatchable? (so if I capture some awesome performance I will be able to reporduce it, no matter of skill needed)