It’s a good thing that product design will be re-focused on. Never understood the point of choosing the form factor that was chosen, as it presented many ergonomic issues from a picture taking perspective.

Product design-wise, it was cool looking, broke the mould, so to speak, and bore some resemblance to Apple-esque design ala the original iSight, albeit rectangular vs tubular.

But limited capacity, small effective image, and lack of any other interesting features besides taking a relatively small image with re-focusable abilities put it in my mind as an answer looking for a question.

Some things I might pursue, were I tasked with the job of taking the technology someplace where it answered real world questions:

Incorporate more processing power into the unit so that it can calculate and render multiple focal point shots.

Say… render and track 2 or more concurrent focal planes at 120fps apiece so that you can use that data to generate a fast and accurate 3D tracking mechanism for object isolation. A sweep of a given lytro frame will yield 3D mapping data sufficient to perform edge/contrast detection focus to isolate objects and track them. Given known knowledge of the objective lens, one could extrapolate distance and focus points in real-time for a given shape/color/etc. Take this information and use it to drive a focus mechanism for a traditional lens setup.

You would effectively be implementing 3D predictive focus abilities with each single frame vs having to zoom the lens back and forth to make a determination like how phase-detect and contrast detect work currently.

The limiting factor for accuracy and usefulness is speed of frame processing and rendering.

Another use case would be for crime scene documentation. A 360 device using multiple sensor arrays capturing the scene and then merging the data to create a 3D walkable room would be an awesome way to document a scene, present information, virtual tours, etc.

Video. If you can achieve the higher frame rate processing, you can use the camera for depth perception video where you render a standard view, but if someone wants to replay a scene focusing on something else in the frame, then they can.

There are so many things that can be done with the underlying technology. The main weakness is the need to know the exacting focal length and lens characteristics to do the right calculations, otherwise, interchangeable lenses would be easy to use. Though to be honest, with a system like Lytro’s, it would be also very easy to create lens vender lock-in or prevent 3rd party lens production.

Anyways, I hope the concept doesn’t go away. It is an interesting concept, and with Sony producing 36MP APS-C sensors, which would yield a 9MP final image, given a 4:1 photosite to micro-lens factor, the technology becomes more compelling, and not less.

Parting thought: intelligent binoculars and glasses. Combine fast processors, a micro lytro array, high speed OLED display, iris focus tracking, and you can create a very powerful, but with few moving parts intelligent dynamic focus tracking head set. The military use for this would be fairly strong for integration with tech/intelligence in the field. Need to determine range and re-check what was in the field of view? No need to reshoot, you can just refocus the recon data and extrapolate range. Need to paint a target and give accurate coordinates? Yep. Lytro’s technology can be there to achieve that as well.

The same system can also be placed in drones. No moving lens elements combined with massive depth of field = better shock resistance, cheaper manufacture, and ability to focus track and target multiple targets at the same time and provide tactical data from one single lens unit’s video/photo stream output.

The technology has a myriad of uses, but the concept of a square tube for taking snapshots… that is not the killer app.