Toronto Photography Meetup Group

Stumbled onto this on another forum. Interesting.

http://www.photographybay.com/2010/06/1 ... /#comments

Just because it's patented doesn't mean it will come to fruit but I wouldn't be surprised if it did.

I would have to look at the actual patent to see what is actually new but some high end sensors (read military already have this capability) It's similar to making a choice of what ISO you use on a pixel to pixel base, for example dark areas were read out at ISO800 while bright areas were read out at ISO50 this would give you around an extra 4 to 5 stops of dynamic range. and the equivalent to a single image HDR that would have to be tone mapped to visualize on a screen. What iso per pixel could be made on the fly or by first taking a throw away image or done on the fly.

A more straight forward way would be just to have more dynamic range at the analogue to digital converter. 24 bit converters are well within the state of the art.

Note that originally HDR was just a method of storing, combining and utilizing High dynamic range files and current camera are I believe around 14 bits which is well into the single image HDR range. Tone mapping in it's less gaudy form is an equivalent of dodging and burning with some help from the computer to set up the masks.

http://www.freepatentsonline.com/7697063.pdf

nventors:
Sugihara, Hiroyasu (Kanagawa, JP)
Application Number:
11/739586
Publication Date:
04/13/2010
Filing Date:
04/24/2007
Export Citation:
Click for automatic bibliography generation
Assignee:
Canon Kabushiki Kaisha (Tokyo, JP)
Primary Class:
396/429
Other Classes:
348/375
International Classes:
H04N5/225; G03B17/48; G03B19/00; G03B29/00
Field of Search:
396/358, 396/532, 396/352, 396/402, 348/374, 348/375, 348/363

HDR cameras have been around for years. Not just in the military.

There are HDR security cameras that can monitor both interior and exterior at the same time. Try doing that with a normal camera on a bright day and see how difficult it is (impossible). The HDR output is compressed to view on normal viewing devices but all information in the image is viewable.

There are HDR movie cameras like the Viper. It never really caught on as there are no HDR playback devices in existence. Sure, there are a few prototype MDR playback devices but they only offer about 4 orders of magnitude in luminance range. I guess another problem was the enormous storage requirements of uncompressed 32 bit video.

And for stills, there is the Spheron VR. Its a HDR camera that captures 360

No matter how good a screen is the human eye even though it has huge dynamic range has to adjust to differentiate more then about 100 shades of gray so you are still going to have to rely on a histogram.

Yep. Histograms are essential for present day HDR assessment.

The human eye is only capable of seeing 5 orders of magnitude (luminance) before it has to adjust. Not bad when you consider the best monitors are only hitting near 2 orders of luminance (even though they can display over 100 shades of grey). Technically speaking a LCD display is capable of displaying more but the range extends below the sensitivity of the human eye.

In total, the human eye can see 10 orders of luminance. To be considered true HDR, a display or capture device would need to span the human eyes full range. Similar to how stereos need to cover 20-20kHz for accurate harmonics. We would need all visual information from the darkest of visible night objects to the brightest visible object during the day. I say the brightest visible object as there would be no point in reproducing the brightness of looking directly at the sun. Its a practical limit.