Incense

by Bart Arondson

submit your photo


Hall of Fame
View past winners from this year

Please participate in Meta
and help us grow.

Take the 2-minute tour ×
Photography Stack Exchange is a question and answer site for professional, enthusiast and amateur photographers. It's 100% free, no registration required.

This is a followup question to Why is the blue channel the noisiest?. The simple answer to that question is that sensors are less sensitive to blue, and therefore require more amplification, which results in more noise. (This is compounded by the fact that typical scene illumination like sunlight or incandescent lamps are lacking in blue.)

So, why are sensors less sensitive to blue light?

Is this just the way many sensors right now are, or is it more fundamentally a "law" of photography?

share|improve this question
1  
Is it just an incidental happenstance of our current CMOS and CCD technology? Is it because of the color filters used, or because of the actual photodiodes? What about back-illuminated sensors? With the typical Bayer layout, there's twice as many green sensors as red and blue; without this, would blue still be worse than green? Why is red apparently better than blue even though there are the same number of photosites? Do CCD and CMOS sensors have a similar response? Or, what about Foveon, which I know has some issues in the red channel? What about, for that mater, blue in color film? –  mattdm Apr 23 '11 at 20:22
    
And can I spell "matter" correctly when typing a bunch of rapid-fire questions? To that question at least, the answer is clear. :) –  mattdm Apr 23 '11 at 20:35
    
I explained last time why the sensors are less sensitive to blue: Much of the blue light is absorbed in the gate structure before reaching the photosite, and switching to backside-illuminated sensors would largely solve the problem. –  coneslayer Apr 24 '11 at 15:26
    
@coneslayer -- thanks. I had somehow missed the link in the bottom section of your last answer. If you have further references for that, that'd be awesome — most non-scientific literature I've seen on backside-illuminated sensors doesn't mention color sensitivity differences. –  mattdm Apr 24 '11 at 15:38
    
This page may be of interest; it has a table of average absorption depth in silicon, as a function of wavelength. You can see that it's dramatically shorter for blue light than red, so you really want the photodiodes to be close to the entrance surface for blue light. learn.hamamatsu.com/articles/quantumefficiency.html –  coneslayer Apr 24 '11 at 16:23

2 Answers 2

The cheapest sensor types by far are CCD and CMOS. They work by taking advantage of the bandgap phenomena of silicon. A sensor is optimized if it's tuned for a wavelength more powerful than the bandgap, but not too powerful. The badgap for silicon corresponds to a 1.1 um light. That's why cameras can typically see in the IR, and require an IR filter to block the signal out.

When the signal becomes much more powerful than the bandgap, sensitivity decreases. Instead of increasing the built up charge, it will tend to just pass through the system entirely.

Blue of the 3 channels has the shortest wavelength, and thus most energy. It's at the point that sensitivity is already decreasing. Further wavelengths, like the UV, are even less sensitive, and thus aren't even imaged (Hence why digital cameras don't need a UV filter)

share|improve this answer
3  
I don't think this is correct. Photons with higher energy than the band gap (i.e. blue light) are absorbed readily by silicon; they don't pass through. The problem is, that on a frontside-illuminated detector, the blue photons are absorbed too readily, within the gate structure, before reaching the photosites. –  coneslayer Apr 24 '11 at 15:13
    
@coneslayer: Good call... It's been a while since I've taken courses in semiconductors, guess I've forgotten a few subtle details... –  PearsonArtPhoto Apr 24 '11 at 15:16

To build on Pearsonartphoto's answer see this application note by Kodak:
Color Correction for Image Sensor - Kodak

This graph shows the natural spectral response of a CMOS sensor(copyright Kodak):
enter image description here
For reference, here is a table relating wavelength to colour(copyright Wikipedia):
enter image description here

The monchrome signal from the CMOS sensor is converted to an RGB signal by siting a Bayer Colour Filter Array before the pixels. This produces, after interpolation, the colour response shown below(copyright Kodak). Note that the peak for blue at 460 nm is roughly 50% lower than the peak responses for red and green. The greater amplification required by this signal produces more noise.
enter image description here

Compare this with the spectral sensitivity of the human eye, below.(copyright E Schubert)
Human eye sensitivity and photometric quantities
enter image description here

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.