Lytro "infinity" Selective DOF in Post

[Wayne Morellini]

I had this idea the other day...

What if you could make 3D video where your eye movement is tracked so that whatever you're looking at is in focus?
That would be quite immersive I think.

Good thinking Luis, it is a good starting point of a direction to look at, people would have to hold the gaze without blinking (I designed a method around this).

 

[Wayne Morellini]

I can't see a huge demand for this on the pro level any way. If you can get your focus and framing right the first time, you don't need it. Although potentially it could allow you to "put" the camera somewhere you can't physically place it.

Unlike a normal setup, this makes it easy to pick focus with big resolutions and to use fat sensor formats.

 

[Wayne Morellini]

I have a thread on this that covers my cheap pro video camera alternative to this that goes more depth to the different possibilities of the technology and post:

http://www.reduser.net/forum/showthread.php?60421-Revolutionary-feature-in-upcoming-camera-what-about-for-Red

Here is some of the things relevant to: what is being discussed here.

Somethings I forgot to mention: I identified this procedure as close enough for a perfect nausea/problem free 3d overlap for processing, focusing and convergence and the rest, which I what I was aiming for. You can get increase resolution by designing the lens for pixel overlapping to get cinema resolution. Purest might not like that but it can offer very acceptable results when done right. My own scheme offers greater pure resolution separately. With super resolution sensors it does not matter so much. You can use larger sensor formats easily, even medium or large format which allow very high resolutions and makes lens design easier. This maybe the separating issue for the digital cinema camera industry, 6 cm and above.

Noticed the haziness, also the focus on the Oscar in their picture gallery almost looks stepped. Forgot to mention the tilt lens possibilities.

I think I will continue that there for a bit as it is about using it for red, different issues and my own scheme.

IMO this could theoretically be a better way to do 3D motion picture imaging.

Now we need a micro-holographic optical computer with optical storage to manipulate the data and record it.

Been working on the idea over the years for other projects, have a number of ideas for cheap solutions, some revolutionary. Anybody got the odd ten million in seed funding for something that eclipses the camera market here?

I think one of the big problems that Red would have, is how do you compress and store all that data at the moment. A pixel overlap resolution scaling idea would bring the size down many times. The scheme I was considering is visibly smaller again.