Petapixel publishes an article on Sony US20140321766 patent application "Image processing apparatus, image processing method, and program" by Jo Kensei proposing different exposure for different pixels in a regular pattern, and a special processing to compensate the difference in image blur:
Could we have your take on that? People start referring to Fuji & Panasonic's similar technologies, but point out those are line-by-line instead being per-pixel.
ReplyDeleteTrue, on the pixel array level, quite a few companies proposed similar ideas. Sony application emphasizes the processing part that reduces the motion artifacts. While it can be done in post processing on PC or smartphone, it would be convenient to have it in camera, may be even in real time.
DeleteCanon usees an similar technique in their Cinema EOS Cameras, but not for High Dynamic purpose. They are summing the two Green Sub-Pixels at different time and out of Phase by 180° Degrees direct on Imager with additional spatial Filtering. This is done to increase resolution, MTF and reducing Aliasing.
ReplyDeleteA Whitepaper (dated 2011) describing this process can be found here:
http://cpn.canon-europe.com/files/product/cinema_eos/menu/EOS-C300-RGB-Resolution-Considerations-White-paper.pdf
I seem to recall a paper on this subject by Shree Nayar and somebody from Sony that was presented at CVPR 2000. Ah, here it is: http://www1.cs.columbia.edu/CAVE/publications/pdfs/Nayar_CVPR00_2.pdf
ReplyDelete