DID JIM JUST ANNOUNCE A LIGHT FIELD HIGH RES RED CAMERA?

[Ivan Kovax]

here's what I posted on another 'lystro/lightfield thread:


I can state many reasons why this technology sucks and should not be invented and probably will NOT be used by any professionals in any field. But here are just a few quick ones

1. takes the artistry out of a focus puller's job. They're not just automatons, they're artists and believe me, Lystro or not, you're not going to do a better job than a guy with 30 years of experience pulling focus.

2. takes the artistry out of the dynamics of hand crafted professional lenses. You see cinema lenses have specific lens characteristics for bokeh and depth of field, if you use this crappy Lystro technology, what will the bokeh look like, looks digital plain/banal to me and I doubt will ever approximate the artistry of Cooke S4's or Arri Master Primes, etc, etc. Just a bad homogenization and will take the organic CHARACTER out of the film capture.
I foresee real professionals continuing to use artistically crafted master GLASS like Cookes, Angenieux, Arri's etc and not this crappy homogenized technology

I agree and disagree:

1. Yes, focus pulling IS artistry, and I am certain there will still be a need for them on set as focus pulling becomes metadata... but if this is to take off it WILL BECOME editing/ color correction artistry too. Imagine flexibility in being able to do as many focus takes/alterations as needed after the fact... it opens up a world of useful creativity in post.

2. Professional lenses are a thing of beauty indeed. There is no reason to think that this same craftsmanship will not be passed on to light field tech with time. Lytro is first gen tech. I suggest that Jim and Red is in a better position to make good on this than Lytro given that Oakley, at its core, was and is a high end optics company. As a side musing, I wonder whether with this tech at some point the bokkeh itself could be metadata and alterable at any point to the bokkeh and lens type of your choice. Wouldn't it be great to be able to emulate the bokkeh of a Canon T1.2 85mm and then decide that a Cooke 85 works better later on with a simple click on your mouse... and when these are indistinguishable from their analog counterparts, we know we've found el dorado. EDIT having read more about LightField Tech, my suggestion seems unlikely, however there is no reason why red or other innovative lens makers wont continue to make wonderful lenses. It is my guess that MONSTRO will be a Lightfield sensor and will be able to use your existing canon and/or PL lenses

Many will find this premise abhorrent, but if Red is serious about this, and it seems they are, then I guarantee within 5 years that Canon, Sony and the rest of them will follow as well... At that point it it will seem as logical as going from film to digital image capturing was.

My 2 yen.



Ivan

EDIT 2. I can't stop thinking about the AMAZING implications this will have on S3D image capturing. With some clever innovation it could do away with the need for having a 2 camera rig...

 

[Ryan Richardson]

For 2d applications I'm thinking that it may be possible to now have two, or more, planes in focus. If the information captured is what I think it is, the possibilities of new tools, if the there is the software for it, this would be revolutionary.

 

[Wayne Morellini]

You know Jim, I was having doubts about the whole 3k 2d thing after this time for me, but 4k with Lytro 3d like capabilities, that is really future looking, good and good for work, it is a winner against other 4k products. Thanks for that.

 

[Wayne Morellini]

I think that you are approaching it in the wrong way. Some productions cannot afford focus pullers or nice lenses or nice cameras. They have to make do with video cameras, DSLRs etc. If light field technology allows the cameraman to use his camera freely without worrying about focussing, it is a good technology - as long as the image looks pleasant.

+1

I agree, it is a feature, like hdr, color balance etc that suits low end productions, where they can put in some artistic refinement in post.

 

[Mohamed Younis]

+1

I agree, it is a feature, like hdr, color balance etc that suits low end productions, where they can put in some artistic refinement in post.

Digital was once only used for low end productions. If it practical and if it means less people on set and more freedom in post without compromising quality, everyone will use it.

 

[Stephen Pruitt]

Personally, I can't wait for it. Focus on fast-moving subjects shot with wide-open lenses is always incredibly difficult, especially with less experienced talent and crews.

Stephen

 

[Peter Stuart]

imagine the action sequences you could shoot if you didn't have to worry about focus.

 

[Jason Cochard]

I made it through most of the Lytro CEO's dissertation and it seems like the effect is basically created by a microlens array in front of an extremely dense image sensor. Each microlens in the array becomes a "pixel" that contains a record not only of that point in the frame, but of the entire frame. And that record is projected onto the corresponding portion on the image sensor. Effectively, the resolution of the lens array becomes your usable resolution--the rest is extra for focus adjustments. I'd suspect (talking out of my ass from here on in) that the "depth of light-field" is a function of the difference between the resolutions of the microlens array and that of the imager. If we expect a 5K final image, that means you'd need a 5K microlens array. The actual image sensor would have to be ridiculously dense, while staying at the Super 35 size (because the best lenses are built to cover a Super 35 image circle, or thereabouts). Diffraction would limit density at a certain point, but whatever. Anyway, with a 5K sensor like Epic I'd expect either very small frame size or very small light-field depth.

If I'm correct about that depth/resolution trade-off, and I don't know if I am, (and even if I'm wrong) I'd propose a shallow depth of light-field for pro use, because flimmakers will never want to get away from on-set focus pulling, even if it does become food for metadata trolls--it's too intimate of a storytelling tool. (And data-rate will limit it at these early stages in my scenario). That said, a light-field with a light-field of a few feet would give a nice pad for those challenging situations, allowing for correction in post. Of course I'm probably being romantic about focus pulling, and in practice it will become a lost art as people will just use light-field as a reason for sloppy focus pulling.

Sidenote. The data rate will be stupid huge. Imagine if you needed dual link HDSDI for each pixel (ok maybe not that much bandwidth). Overcranking? You'd need a datacenter on set, right?

 

[Tim Whitcomb]

The focus correction is cool and all... but it will not be what you all think... imho it will introduce a new look. the look du jour even. thats cool
but what excites me is the ability to spatially key. and buh bye green screen. THAT application of it's high resolution is beyond $%&ing cool...
sick ass effects are near... and I remember when the original TRON stunned. sigh