a reply to:
Hyperboles
It is basically what i said it to be - Using my Cell Phone, sticking my head out of the window, I took pictures on the 3rd floor of the building and
another outside on the ground, in a short time span as I could...
The photos where taken AT the sun, with the cell phone set to fixed white balance to prevent any artificial colour shift, iso 100 (lowest it goes on
mine) and a ev value of -2 to give me the lowest exposure the camera will do. Each photo has the sun in frame a top left, Time of day was 12:15 local.
The Phone used as a somewhat old Samsung Galaxy S4 mini, has a 8M pixel CCD.
Drawing a circle around the centre Sun at 500 pixels and a colour reading made with Apple Digital Colour Meter in order to get the RGB measurements.
The measurement is taken along a line directly toward the ground a hillside feature visible in each photograph in order to sample the same spot of
sky. The tool is nice as it allows an average to be taken over a range of pixels. The error bar i placed comes from moving the sample point around to
give a sense of how much the bit values change over the range.
a huge waste of 10 minutes i might add, in order to prove something that anyone with half a brain cell should be able to tell... and with more of an
explanation than you did in that paper...
Photo 1 - Approximate height difference 38ft above ground level
RGB value at the Sun - Saturated
R = 255
G = 255
B = 255
Photometric measurement is thus useless
500 pixels away
R = 56 +-2
G = 94 +-2
B = 120 +-2
RGB values not saturated
Photo 2 - At ground level
RGB value at the sun - Saturated
R = 255
G = 255
B = 255
Photometric measurement is thus useless
500 pixels away, approximately same spot
R = 50 +- 2
G = 85 +- 2
B = 108 +- 2
So the colour changed... OR did it? Well it turns out what actually changed was the brightness of the shot, not the actual colour. If we take each
value for R G and B and find the ratio between each measurement we get
R2/R1 = 0.893
G2/G1 = 0.904
B2/B1 = 0.900
The average of which is 0.899 well within the pixel noise showing there is no shift in colour but a shift in brightness.
When applying the average as a correction we get - We should also round to the nearest pixel since the bit depth of the image is 0-255
Rc = 55.6 -> 56
Gc = 94.5 -> 95
Bc =120.1 -> 120
OK so conclusion after way way way way way more effort than the author of the paper did, with a much more scientific explanation for the measurement,
how it was performed and how the data was gathered...
While the sky is not a discrete source using it in this manner still makes sense since any shift due to altitude above surface to change the local
gravitational field should shift the source of all light. As The sky's blue colour during the day originates from scattering and a long pathlength of
sunlight through it, we would expect the same shift in spectrum as predicted by the paper... according to the paper this shift should be huge, so we
should see it in the sky spectrum.
Following the train of thought in the paper if it is red shifted we expect more red and blue, more blue... what I have shown, despite my statement
that a CCD is not at all a spectrometry instrument, that even on its own terms, the paper is a fine example of poor understanding of not only CCD bit
depth, but also instrumentation, experimentation and analysis. I genuinely fear for the lives of those the author sailed with as a engineer of the
merchant navy, or instructed during flight school...
Note
I can do the same measurement with greater bit depth using a Nikon D610 and the sun, with a lens stopped down and a fast shutter so I am not burning
the CMOS chip, all settings on manual as not to artificially reweighs the colours... I predict exactly the same outcome of... no change at all... If i
feel like wasting my time to prove someone who doesn't understand the concept of a spectrometer... i might re-do it.
edit on 15-4-2019 by
ErosA433 because: (no reason given)