Today I and presumably quite a few others who also put a deposit down for the L16 received some mixed news today via an e-mail from Light. The bad news is their camera will be delayed until 2017 sometime. We had been told it would ship this summer originally, and I’ve been very excited to work with this new multi-aperture technology which may be the direction small cameras like mobile phones will take in the future. The good news is Light has been able to add some wider optics to 28mm effective and increase the inboard memory to 256 Gb.
Here’s the update information on their site: L16 Update If you follow the link they do have a video on their manufacturing process for the prototypes which is pretty cool – lots of small automated assembly going on in clean rooms.
For those of you have not heard yet about the light camera or multi-aperture computational cameras you can read about it on their Light L16 technology, but here’s a very short summary: Multi-aperture means more than one lens and sensor are employed simultaneously – in the case of the L16 there are actually 16 tiny lens and sensor packages (modules) stuffed into the device which is about the size of a tablet style cell phone but twice as thick. The different data sets from each aperture are then used to create one larger image set that will be higher resolution than just one camera and also have less noise since the noise can be identified and separated from the true image data. Because the lenses of the modules are physically separated on the face of the camera they have different vantages of the same scene much like a human’s eyes do, and that allows some depth information to be calculated. This should enable the photographer to select the plane of focus and focus point after the shot was taken, use shallow DOF, or even maximize the DOF. It’s quite difficult to put longer focal length assemblies into a flat package since the lens needs to be further away from the sensor. Light got around that problem by leaving the lens facing forward as normal but using a mirror underneath to direct the incoming light sideways to a sensor that can be put further away. Rather than use an optical zoom, Light has chosen to use several modules of three different focal lengths (28mm, 70mm and 150mm effective in 135mm format). From their website it appears that the user would have to select one of those focal lengths currently, but they could use computational algorithms to output images with effective focal lengths in between the three physical focal lengths. Light states that up to 10 of the 16 modules will fire for each shot. I can only guess that having all 16 fire wasn’t worth the overhead in terms of computational time and battery life otherwise why wouldn’t they use all 16 all the time?
I’m not sure what Light is doing to remove shake but I’ve noted that they have placed the mirrors on different angles as opposed to all the same way so perhaps this allows them to computationally identify and remove some camera shake?
The L16 will be an interesting camera for sure, and I’m hoping they get the tech right! At least they did build this on an Android platform with capability to share directly to social media and all the onboard memory is welcome as is the 5 inch touch screen. When I finally do get one in my hands, I hope that I don’t drop it right away! Those modules with the mirrors probably have a very fine mechanical tolerance and may be easy to knock out of alignment.