Camera Start-Up Lytro Promises to "Revolutionize" Photography

For sure.

 

What are they going to do about lenses?

 

Hmmm?? Sounds interesting.

 

Interesting concept, but the example picture, where you can refocus by clicking, doesn't tell anything - you could very well do that with a bunch of ordinary pictures, differently focused, and combined into such a little flash app.

 

you won't need lenses? you won't even need to focus. You just capture all the light going in at that specific point and focus it however you want post production. Who knows how big the raw files will be to do this.

 

Bingo. I just saw a spot on CNN about this. I couldn't get my head around how this could possibly work - but I think you hit the nail on the head. We are so used to thinking about the image after it passes through the lens - not before.

 

I don't know how well this will work in real world situations, I just don't see the technology there to get the pin-point accuracy needed for some images. Basic focus points, yes, but in some images there will be billions of focus points and I don't see it happening.

 

This will probably be the future of photography. Imagine digitally recording the light on a plate; in effect gathering the light before it enters the lens. You could then digitally process that light data through any virtual lens you wanted to with any f-stop, and even varying shutter speeds up to the length of time you allowed the plate to collect data. Post-processing software could even extend the exposure time via calculations.

 

The technology sounds very promising. Until we see it implemented in an actual camera may be some time.

 

Definitely interesting. I wonder if this will at first be more important for that one all important shot. For instance a security camera where the perp is out of focus. Or some military application.

Or that shot where the bride and groom are out of focus, while the flowers in the foreground are perfect.

I'm not sure but with this technology I guess every shot will be focused on a computer, so the RAW analogy above is perfect.

 

http://gizmodo.com/5890028/lytro-light-field-camera-this-is-what-new-feels-like

Pretty amazing stuff. I think once they can incorporate this into higher level cameras i.e. DSLR it's going to be a grand slam for photography. When shooting at anything under f2.8 its pretty hard to make sure the subject is actually in focus, of course until you get back to the computer.

 

I read about this on Pulse. It's not as revolutionary (IMO) as claimed.

It uses a series of special lenses to channel the light into rays, and then samples the convergence of those rays at various points.

In other words, it creates a series of focal planes. So it doesn't really allow you to focus anywhere... it is more like taking a picture with 7 different lenses at once (or one lens set to 7 different focal lengths at once), then you can select amongst them.

If I took a picture of a city looking down a long street, it's not like I could click on any house along the line of sight and bring it into perfect focus. So it's a trade-off.

I actually find the pictures somewhat counter-intuitive and non-user friendly. It doesn't do anything for me, really. I don't think it's the future of photography, but I've been wrong before (once, in 1982).

 

You own a Beta Max? Or were computing with a CPM machine?

 

http://www.dpreview.com/reviews/lytro/

not quite ready for prime time it seems...

 

That's what I was afraid of, hopefully this will spark some interest though, being able to pinpoint focus after the fact would be a god send for photography.

 

Another way of looking at this is that if the camera has a 10Mp sensor and 10 focus points, it takes 10 pictures of 1Mp each. Not a good tradeoff in my opinion.

 

I'd rather they improve sensitivity performance and DR.

My desire is a sensor on which the ISO gain can be controlled on a per-pixel level. Then, using the metering sensor and a logic chip (with some user parameters like +/- stop envelopes), you could decrease sensitivity in the highlights and increase it in the shadows. Imaging the effective dynamic range if shooting outdoors on a cloudy day where the sky is at ISO50 and the shadows are at ISO400.

 

That would be nice, I had a helluva time last week trying to do a photoshoot of 3 Ferraris in the foreground and a sunset in the bg. Had to expose for the sunset and bring out the cars in post.

 

not the best of reviews on dpreview:
http://www.dpreview.com/news/2012/02/29/Lytro-Hands-On-Review-and-Video

 

meh, wake me up when there's a cheap, small cam capable of 1000++fps in hi def.
HeroPro and Twixtor doesn't count, quite.

 

I lighten and darken areas within pictures in photoshop all the time. It works really well. I also "blur" the non-essential elements to make the pictures work better. Did you mean something else?

As for Lytro, it's a decent enough idea, but the current version is too low-res for anything but sharing stuff on facebook, and the screen sucks. The control placements also need to be improved a little, but it's a good job for a version 1.0. Got to play with a pre-production unit for a while through a friend of mine who was considering investing in the company (ultimately, he didn't). I don't like the tube form-factor either.

 

Yah, I didn't mean take a photo and then post process it. This would be more akin to pre processing.

Camera hardware is capable of determining shadows and highlights. It knows before recording an image which pixels will be blown out (white that no amount of post exposure adjustments will fix) and which will be similarly under(un)exposed. If it could selectively decrease pixel sensitivity on the overly bright regions and increase it on the overly dark regions, those details could be recorded in a single snap (as opposed to taking multiple pics at varied shutter speeds and stitching then together).

 

Is that a Lytro in your pocket or are you happy to see me!!!