File talk:Spectrum of blue sky.png

Hi,

I wonder if you could elaborate on 'This spectrum is not calibrated for intensity'? Does that relate to its giving intensity as a 'count' rather than in terms of energy delivered? A naive look at the spectrum shown seems to suggest that the sky should be green! I'm speculating that the apparent mismatch between 'intensities' recorded here and those expected (stronger violet than blue, stronger blue than green, etc.) could be because the shorter-wavelength radiation has more energy per photon, but am I right?

Clarification could help resolve a dispute here...

Thanks!

--Oolong 08:57, 21 October 2006 (UTC)[reply]

Ha! How weird! I was JUST looking at that flickr pic the other day! What I mean by 'this spectrum is not calibrated for intensity' is that I did not remove the nonlinear respose curve of the CCD to the frequency of light hitting it. In other words, the frequency measurement of the thing is very accurate in this mode but the AMOUNT of light hitting the detector is not accurately reproduced because the detector is much more sensitive to green light than light of any other wavelength. Look at the spectral response curve of the detector here on page 8 [1]. I am working on calibrating the detector for intensity (see here: [2]). I have, in fact done a spectrum of blue sky which is calibrated for intensity and yes, the peak IS in the blue region rather than the green seen here but it is very preliminary and I am still working out how to do it properly. So I will upload a new calibrated version of this in the future but its going to have to wait a bit; its always cloudy here!! :) --Deglr6328 21:57, 21 October 2006 (UTC)[reply]

Hi, I keep reading that the sky doesn't appear purple because of the construction of the human eye. But this chart seems to indicate that the effect is real--that the sky's spectrum peaks in the middle of the optical, then is attenuated at higher frequencies. What causes that high frequency attenuation? Is it the glass the spectrum is taken through? Pulu (talk) 02:54, 15 December 2008 (UTC)[reply]