"de-mosaicing" with the new Quattro sensor

Started 7 months ago | Discussions thread
ForumParentFirstPreviousNextNext unread
Flat view
tpani
Regular MemberPosts: 107
Like?
"de-mosaicing" with the new Quattro sensor
7 months ago

I just read Thom's (the Nikon blogger) take on the new sensor. He incorrectly says (as many do, and as many have corrected) that the top layer is blue. Additionally, and more interestingly, he suggests that some kind of demosaicing is needed.

Of course, what the sensor is doing is more like chroma subsampling (rather than Bayer), as in digital video and photography and imaging in general (e.g. the various JPEG formats.) Every 2x2 pixel block has full luma for every pixel, but chroma only for the phole 2x2 block, not each pixel individually. The usual sample images for this of course show that this is visibly worse than having full chroma information for each pixel individually, although I am not sure whether this translates to anything interesting outside the pixel-peeping level.

My question is related to the fact that the usual translation direction for chroma subsampling is from full information to the condensed (less) chroma (simply as a process of reducing the amount of data that needs to be stored). But, what about the other direction? So is Sigma just going to leave the chroma information as is, or is there any point of doing some kind of very approximate smoothing of the chroma information between neighboring 2x2 blocks? Of course, the information is already lost, but is there any point of doing some sort of smarter "demosaicing"?

ForumParentFirstPreviousNextNext unread
Flat view
Post (hide subjects)Posted by
ForumParentFirstPreviousNextNext unread
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow