Adaptive ISO?

[lbalbinot]

Hi!

I just hit my head to the bathroom sink and got and idea (no, not the flux capacitor!): what if camera sensors had different exposure times on portions of it's pixel array? Say, the top 1/3 of the image has a -2 EV exposure than the bottom part. Just like some sort of virtual ND filter. I know that control over independent pixels might be more complicated, but control over lines of pixels might be easier (just turn them of earlier or later).

I went to Google and very easily I found this:

http://www.apical-imaging.com/UserFiles/ad-iso.pdf

They even used the name I came up with! Damn it! :-)

Why don't we have that on our cameras yet? Could be something a lot simpler, just like the virtual ND I suggested.

Regards,
Luis

[lbalbinot]

Now there's a rumor that Canon will come up with a sensor that can actually do this...  

Luis

[Eric Myrvaagnes]

Now there's a rumor that Canon will come up with a sensor that can actually do this...   

Luis
[a href=\"index.php?act=findpost&pid=220543\"][{POST_SNAPBACK}][/a]

Maybe you can get them to pay you royalties. Just don't show them the Googled item.  

[francois]

Now there's a rumor that Canon will come up with a sensor that can actually do this...   

Luis
[a href=\"index.php?act=findpost&pid=220543\"][{POST_SNAPBACK}][/a]

Try to include a mirror lockup button in your patent. Canon might finally implement it.
 

[Panopeeper]

I suggest to learn the meaning of "exposure" before applying for a patent.

[Ray]

Panopeeper is applying a tactic sometimes used by Jonathan Wienke to discredit the value of an idea.

The current paradigm of color-filtered Bayer sensor, and sensor recording of linear light values, is extremely inefficient.

The color filters necessarily block out about half the light impinging upon the sensor and the brightest stop in any image contains far more values than the eye can discern.

There's a huge room for improvement.

[BruceHouston]

Hi!

I just hit my head to the bathroom sink and got and idea (no, not the flux capacitor!): what if camera sensors had different exposure times on portions of it's pixel array? Say, the top 1/3 of the image has a -2 EV exposure than the bottom part. Just like some sort of virtual ND filter. I know that control over independent pixels might be more complicated, but control over lines of pixels might be easier (just turn them of earlier or later).

I went to Google and very easily I found this:

http://www.apical-imaging.com/UserFiles/ad-iso.pdf

They even used the name I came up with! Damn it! :-)

Why don't we have that on our cameras yet? Could be something a lot simpler, just like the virtual ND I suggested.

Regards,
Luis
[a href=\"index.php?act=findpost&pid=207746\"][{POST_SNAPBACK}][/a]

The Apical technology, rumored to be included on the 5D Mark II, is pixel-by-pixel variable ISO.  This is easier to implement than pixel-by-pixel exposure time by the (also rumored) DIGIC IV processor(s) which are presumably powerful enough to perform digital gain calculations separately on a pixel-by-pixel basis.

Now, if we could just put itsy-bitsy diaphrams behind those microlenses to achieve pixel-by-pixel aperature settings...      

On a separate note, I wonder if Michael R. is playing with the new 5D toy even as we speak...