Mitsubishi Electric Develops Deblurring Flutter Shutter Camera

[Ken Lee]

“We have UV filters. We have polarizing filters. What about time filters?” said senior research scientist Ramesh Raskar at Mitsubishi Electric.

See http://www.digitalcamerainfo.com/con...of-Toronto.htm

[darr]

Bad time to put any investment into high-speed films!

Thanks for posting Ken!

[Donald Qualls]

Hmm.

So, the final version could be a filter with PC or hot shoe connector and a battery; when the camera shutter fires, the "flutter" filter starts its sequence -- and this could potentially be used on *any* camera.

Actually, darr, this would work *better* with high speed films, because you'd still be able to use a shorter shutter speed or smaller aperture. Picture shooting, say, a NASCAR race and being able to deblur the cars well enough to read the driver's expression?

[darr]

Actually, darr, this would work *better* with high speed films, because you'd still be able to use a shorter shutter speed or smaller aperture. Picture shooting, say, a NASCAR race and being able to deblur the cars well enough to read the driver's expression?

Thank you for enlightening me Donald.

BTW, I like your Spicy Chicken Curry recipe. I have cooked a similar recipe in a crock pot, but will give this recipe a try.

[Bruce Watson]

So, the final version could be a filter with PC or hot shoe connector and a battery; when the camera shutter fires, the "flutter" filter starts its sequence -- and this could potentially be used on *any* camera.

Actually, darr, this would work *better* with high speed films, because you'd still be able to use a shorter shutter speed or smaller aperture. Picture shooting, say, a NASCAR race and being able to deblur the cars well enough to read the driver's expression?

So... I could still take that 1/4 second exposure, but now I don't have to wait for 45 minutes for the wind to die down? Now that would certainly be something I'd be interested in!

[Donald Qualls]

That's the general idea (though you'd probably have to extend to 1/2 second, to allow for the approximate 50% time slicing of the flutter shutter, and maybe a little more to allow for its transmissivity when "open").

However, to use this, you'd also have to be willing to digitally process the resulting image, and from the examples, the result will be grainy and not very "large format" like in appearance. Don't know how much of that is from the small original format and further cropping, however, and how much due to the deblurring algorithm.

[Struan Gray]

So... I could still take that 1/4 second exposure, but now I don't have to wait for 45 minutes for the wind to die down? Now that would certainly be something I'd be interested in!

I see two problems.

The first is that a global deconvolution will only work if the whole scene has the same blur. That will work for camera shake (up to a point) but not for wind blur where different parts of the scene are moving by different amounts. There are ways of cutting out individual parts of the scene, de-blurring them individually and putting them back together again, but for a scan off a piece of LF film you are talking significant amounts of computation.

The second is that the film has to be sensitive enough to record a blobby trail of light for each pixel in the final image. You are taking what would be a spot of intensity on a sharp negative and spreading it out along a line. Fine if you're doing handheld stuff in bright sunlight, but not much use for shadow detail on a typical dusk landscape, where reciprocity failure will kill the spread-out spot so it hardly records.

That said, I doubt LF photographers are the target audience in any case :-)