Ad
Ad
Ad
Pages: « 1 [2]   Bottom of Page
Print
Author Topic: 120MP Canon  (Read 4606 times)
Wayne Fox
Sr. Member
****
Offline Offline

Posts: 2739



WWW
« Reply #20 on: August 25, 2010, 03:23:34 PM »
ReplyReply

it seems to me there might be some usefulness to a lot of pixels with some cleverness -- improved dynamic range per Fujii, pixel binning at high-iso. 
Perhaps binning could be used for something other than just noise.  Perhaps a sensor that doesn't need de-moisaicing?  What if the idea is to "bin" 4 pixels with appropriate filters as a single pixel and derive the exact color of the final pixel only from the color of those 4 pixels.  So while the sensor has the traditional "bayer" filter arrangement, you have 4 pixels acting as a single pixel.  End result you have a 30mp sensor with no moire, no anti-aliasing filter necessary and no de-moisacing required.

I guess the only reason I say this is it something I've wondered about for some time.  Sort of like adding cores to a single CPU.  Create a sensor where multiple pixels work as single pixels, not only for noise but for this issue as well.

I guess the best way to
Logged

Graeme Nattress
Sr. Member
****
Offline Offline

Posts: 582



WWW
« Reply #21 on: August 25, 2010, 06:03:32 PM »
ReplyReply

Unfortunately you'll still get moire and aliasing doing that, probably more so than using a standard demosaic approach.

Graeme
Logged

www.nattress.com - Plugins for Final Cut Pro and Color
www.red.com - Digital Cinema Cameras
ejmartin
Sr. Member
****
Offline Offline

Posts: 575


« Reply #22 on: August 25, 2010, 06:51:26 PM »
ReplyReply

120MP is basically P&S sized pixels (2.2) in a DSLR sized sensor. Please note that noise is not a fixed quantity, but varies with scale in the image. There are two components to noise in most images -- read noise (noise contributed by the camera electronics), and photon shot noise (quantum fluctuations in the light signal itself). At low to moderate ISO, photon noise tends to be the more visible noise in an image. At a fixed image scale, this noise depends only on how much light is collected, and is therefore independent of pixel size. High ISO might be compromised somewhat; with read noise unchanged from current performance of production DSLR's, read noise at a fixed scale will be about twice that of 7D (since the pixels are about half the linear dimension). Low ISO might not be so compromised for read noise; for instance the 40D has about 10-15% less read noise than the 1D3 at base ISO, so smaller pixels might have slight advantage unless Canon decides that cleaner low ISO shadows are more of a priority (they've been going backwards in this department recently). They could also mitigate the high ISO read noise issue by cutting down on the pattern noise that plagues their cameras at all ISO (though it's worse at low ISO).

And since diffraction is often discussed in the context of pixel size, note that diffraction is a property of the optics, not the sensor; having smaller pixels does not increase diffraction -- on the contrary, all it does is decrease the range of f-stops over which the sensor resolution is the limiting factor in system resolution, rather than the rest of the optics. A side benefit is that the AA filter's blur radius, being tied to the pixel size, will be smaller; and demosaic artifacts will be pushed off to finer image scales where they will be less noticeable.

What Canon really needs to do if they are serious about heading in this direction is to concentrate some development resources on compression technology -- clearly one doesn't want to have to deal with the huge files that result from high MP count cameras, since most of that information is redundant. sRAW is the lamest possible image compression method one can imagine. Much better would be the sort of compression RED uses -- preserves nearly full resolution while reducing file sizes substantially, and allowing continuous shooting without maxing out the image buffer. I did a very klugy exercise some time ago, taking a raw file, separating out the RGGB planes and doing JPEG2000 (wavelet based) compression by about a factor of 10, then uncompressing, reassembling the raw, and demosaic. The result was not bad for such heavy compression; I'm given to understand that RED does something along these lines, but of course they've spent more than an afternoon on it and it works much better.   Wink

One could also wring a little more savings out of recognizing that 14-bit is wasteful at current output DR levels; 12 is sufficient for all Canon DSLR's ever made. Level-thinning methods such as Nikon uses with its lossy compression are another way to shave off a bit or two more per pixel. But to me, the issue is that one doesn't want to keep all the image info at high res (eg, not in skies and other smooth regions), so the question becomes what data to keep and what to discard. Low res cameras make the decision for you, by omitting all fine scale image information; but there are certainly more intelligent methods one can imagine (and some that exist already), where compression technology keeps high res information where it is desired, and discards it where it is not.

BTW, as I understand it Canon is not "obsessed" with APS-H; rather, it is the largest sensor size that they can make without stitching on their fab line.
Logged

emil
michele
Full Member
***
Offline Offline

Posts: 230


« Reply #23 on: August 26, 2010, 06:04:04 AM »
ReplyReply

If you read till the end you'll find this: In 2007, the company successfully developed an APS-H-size sensor with approximately 50 million pixels.
We are in the middle of 2010; where is my 50 megapixels Canon?
Logged

RFPhotography
Guest
« Reply #24 on: August 26, 2010, 07:23:36 AM »
ReplyReply

It's nice to see that the LL forum reputation as the place for the measurebators is safe and sound.   Grin

It's surprising (well maybe not  Roll Eyes ) to see so much serious discussion about a device that will never make it to market or, if it does, not likely in any of our productive lifetimes. 

As someone else noted, this is a 'because we can' exercise.  It's a headline grabber. 

No doubt Canon will do testing with this sensor.  That's what this kind of exercise is.  It's a test bed.  It's like Formula 1 as a test bed for technology that eventually migrates down to our street cars (not as much anymore with the restrictions on costs in F1, sadly). 
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2718



« Reply #25 on: August 26, 2010, 09:14:34 AM »
ReplyReply

120MP is basically P&S sized pixels (2.2) in a DSLR sized sensor. Please note that noise is not a fixed quantity, but varies with scale in the image. There are two components to noise in most images -- read noise (noise contributed by the camera electronics), and photon shot noise (quantum fluctuations in the light signal itself). At low to moderate ISO, photon noise tends to be the more visible noise in an image. At a fixed image scale, this noise depends only on how much light is collected, and is therefore independent of pixel size. High ISO might be compromised somewhat; with read noise unchanged from current performance of production DSLR's, read noise at a fixed scale will be about twice that of 7D (since the pixels are about half the linear dimension).
Emil's analysis assumes that photon collection is dependent on total sensor area and is not affected by pixel size. However, CMOS sensors have transistors in each pixel and a fill factor of well less than 100%. If one makes pixel size smaller but does not shrink the transistor area proportionally by using a smaller process for the electronics, fill factor will decrease. This can be compensated for by the use of microlenses, but full well will likely suffer because of increased charge density. As Catrysee and Wandell explain (CMOS Roadmap), small pixels require a larger f# since the stack height of the sensor does not scale with pixel size, greatly complicating micro lens design.

Furthermore, read noise occurs with each pixel readout, and doubling the lateral size of the sensor while holding pixel size constant would quadruple the pixel count and the total read noise would also quadruple;  a large pixel can be read with about the same level of read noise as a small pixel (see Photometrics), so holding the pixel count constant while increasing sensor size decreases total read noise. Downsizing in software does not reduce read noise by the same amount as binning in hardware.



Logged
EinstStein
Sr. Member
****
Offline Offline

Posts: 263


« Reply #26 on: August 29, 2010, 10:31:37 PM »
ReplyReply

Pixel noise might be handled by deep trench. remember the number of electronic is counting in volumn (3D) not area (2D).
Theoretically, the diffraction problem could be handled by correlation across the pixels. Who said a good image must be clean per pixel?   
Logged
erickb
Full Member
***
Offline Offline

Posts: 235


WWW
« Reply #27 on: August 30, 2010, 02:38:03 AM »
ReplyReply

it is soon time for me to sell all my Canon gear
Logged
Pages: « 1 [2]   Top of Page
Print
Jump to:  

Ad
Ad
Ad