Go Back   Steve's Digicams Forums > Digicam Help > General Discussion

Reply
 
Thread Tools Search this Thread
Old Feb 23, 2012, 4:07 PM   #1
Senior Member
 
MartinSykes's Avatar
 
Join Date: Feb 2010
Location: Cheshire, UK
Posts: 854
Default Bayer filter

Just wondering why the common configuration is to have RGGB. If they had RGB and had no filter on the fourth cell, wouldn't that give *much* better low-light performance?
__________________
AutoHDR - Free HDR software for Windows at http://www.autohdr.co.uk on Facebook at http://www.facebook.com/home.php?sk=...36045126467361
My Flickr stream: http://www.flickr.com/photos/martinsykes
Panasonic DMC-FZ38, Sony ­α580
MartinSykes is offline   Reply With Quote
Sponsored Links
Old Feb 23, 2012, 6:29 PM   #2
Senior Member
 
TCav's Avatar
 
Join Date: Sep 2005
Location: Washington, DC, Metro Area, Maryland
Posts: 13,571
Default

Yes, but nobody has started using it yet. See http://en.wikipedia.org/wiki/Bayer_f...matic.22_cells
__________________
  • The lens is the thing.
  • 'Full Frame' is the new 'Medium Format'.
  • "One good test is worth a thousand expert opinions." - Tex Johnston, Boeing 707 test pilot.
TCav is offline   Reply With Quote
Old Feb 23, 2012, 8:09 PM   #3
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

http://www.sony.net/SonyInfo/News/Pr...10E/index.html
JimC is offline   Reply With Quote
Old Feb 24, 2012, 8:22 AM   #4
Senior Member
 
MartinSykes's Avatar
 
Join Date: Feb 2010
Location: Cheshire, UK
Posts: 854
Default

Thanks guys - I guess it's just an idea who's time has come. That HDR movie on the Sony site looks pretty amazing.

"enables the image sensors to achieve a higher resolution at a more compact size"
Not sure what to make of that statement. I guess they're thinking that rather than have an APS-C with better low-light capabilty you could instead aim to have a 4/3 sensor with the s/n ratio of a current APS-C sensor. That could be very interesting if only because it would make long zooms more affordable.
__________________
AutoHDR - Free HDR software for Windows at http://www.autohdr.co.uk on Facebook at http://www.facebook.com/home.php?sk=...36045126467361
My Flickr stream: http://www.flickr.com/photos/martinsykes
Panasonic DMC-FZ38, Sony ­α580
MartinSykes is offline   Reply With Quote
Old Feb 24, 2012, 9:41 AM   #5
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

The design is a tradeoff from comments I've seen from Sensor Engineers, and the idea has been around for a while (from Kodak and others). Basically, you're trading off better color resolution for a little better low light S/N ratio, which can degrade overall image quality in most conditions. Of course, more advanced processing probably comes into the equation, too. So, perhaps Sony has figured out a way to reduce the downsides to the approach.
JimC is offline   Reply With Quote
Old Feb 24, 2012, 12:30 PM   #6
Senior Member
 
TCav's Avatar
 
Join Date: Sep 2005
Location: Washington, DC, Metro Area, Maryland
Posts: 13,571
Default

The green photoreceptors are used to set the luminance value anyway. Removing the filter from one of those photoreceptors means better luminance measurement. Plus, the red and blue photoreceptors are already handicapped; this would give green the same handicap, resulting in better color balance. Plus, the white photoreceptor would provide better dynamic range as well as lower noise.
__________________
  • The lens is the thing.
  • 'Full Frame' is the new 'Medium Format'.
  • "One good test is worth a thousand expert opinions." - Tex Johnston, Boeing 707 test pilot.
TCav is offline   Reply With Quote
Old Feb 24, 2012, 1:00 PM   #7
Senior Member
 
Join Date: Sep 2005
Posts: 1,093
Default

Quote:
Originally Posted by TCav View Post
The green photoreceptors are used to set the luminance value anyway. Removing the filter from one of those photoreceptors means better luminance measurement. Plus, the red and blue photoreceptors are already handicapped; this would give green the same handicap, resulting in better color balance. Plus, the white photoreceptor would provide better dynamic range as well as lower noise.
But, in the Bayer system, the RGGB represents two pixels; presumably in the RGBW system, the four elements represent one pixel. Dynamic range has improved markedly, as has sensor density. It isn't really clear to me which system would provide the greater win.
tclune is offline   Reply With Quote
Old Feb 24, 2012, 1:15 PM   #8
Senior Member
 
MartinSykes's Avatar
 
Join Date: Feb 2010
Location: Cheshire, UK
Posts: 854
Default

RGGB and RGBW represent 4 pixels in both systems.

The 'R' pixel has some G and B added to it from adjacent pixels and so on.

What I don't get though is that the W pixel will register very low light even when the adjacent RGB pixels are registering zero. What colour should that pixel be? I guess it must be interpolating over a wider range to guess the colour where the three filtered cells have given up. I was expecting it to be grey-scale in the shadows but they must have come up with some solution. [edit] maybe I'm wrong there - maybe rgb don't register zero but they are low enough that they are clipped to zero but still have some value that can be used in-camera. hmm.
__________________
AutoHDR - Free HDR software for Windows at http://www.autohdr.co.uk on Facebook at http://www.facebook.com/home.php?sk=...36045126467361
My Flickr stream: http://www.flickr.com/photos/martinsykes
Panasonic DMC-FZ38, Sony ­α580

Last edited by MartinSykes; Feb 24, 2012 at 1:19 PM.
MartinSykes is offline   Reply With Quote
Old Feb 24, 2012, 4:18 PM   #9
Senior Member
 
TCav's Avatar
 
Join Date: Sep 2005
Location: Washington, DC, Metro Area, Maryland
Posts: 13,571
Default

With a Bayer Filter, the data for each R photoreceptor is used to set the R value for 4 image pixels. The same is true for each B photoreceptor. But each G photoreceptor is used to set the G value for only 2 image pixels. Thus, the image sensor captures twice as much detail for G as for R and B. (A common selling point for the Foveon sensor is that it captures more R and B detail than a sensor with a Bayer Filter. A filter with an RGBW pattern would negate that argument.)

Also, if we presume that each photoreceptor is equally as likely to contain noise as any other, then with a Bayer Filter, noise will be twice as likely to be green as red or blue. And if we presume that the level of noise is likely to be the same regardless of the color of the filter on a photoreceptor, then an unfiltered W photoreceptor will sample a lot more light than any of the R, G, or B photoreceptors, but have the same amount of noise, so it's signal-to-noise ratio will be much higher, resulting in a lower overall noise level for the entire image.
__________________
  • The lens is the thing.
  • 'Full Frame' is the new 'Medium Format'.
  • "One good test is worth a thousand expert opinions." - Tex Johnston, Boeing 707 test pilot.

Last edited by TCav; Feb 24, 2012 at 4:22 PM.
TCav is offline   Reply With Quote
Old Feb 24, 2012, 4:21 PM   #10
Senior Member
 
TCav's Avatar
 
Join Date: Sep 2005
Location: Washington, DC, Metro Area, Maryland
Posts: 13,571
Default

Quote:
Originally Posted by MartinSykes View Post
What I don't get though is that the W pixel will register very low light even when the adjacent RGB pixels are registering zero. What colour should that pixel be? I guess it must be interpolating over a wider range to guess the colour where the three filtered cells have given up. I was expecting it to be grey-scale in the shadows but they must have come up with some solution. [edit] maybe I'm wrong there - maybe rgb don't register zero but they are low enough that they are clipped to zero but still have some value that can be used in-camera. hmm.
If a W photoreceptor senses a small amount of light while the adjacent R, G, and B photoreceptors don't detect any light, then the W photoreceptor will contribute equally to each of the adjacent pixels, resulting in a dark gray.
__________________
  • The lens is the thing.
  • 'Full Frame' is the new 'Medium Format'.
  • "One good test is worth a thousand expert opinions." - Tex Johnston, Boeing 707 test pilot.
TCav is offline   Reply With Quote
 
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -5. The time now is 12:50 AM.