Moon Photography: Auto White Balance Mystery

Started Aug 20, 2013 | Discussions
Bob in Baltimore
Regular MemberPosts: 439Gear list
Like?
Moon Photography: Auto White Balance Mystery
Aug 20, 2013

I have a number of photos of the rising Moon with interesting landscape features. I am preparing a talk on this topic for the NASA Goddard Photography Club next month. One of the things I did in preparation for this was to photograph the rising Moon to measure atmospheric extinction ... changes of up to 5 f-stops (!) occur within 15 minutes when you start with the Moon right on the horizon.

I was able to clearly document a phenomena that I had previously observed ... the auto-white balance does some very strange things. Within 2 minutes the color of the Moon (as photographed) changed dramatically. The color of the Moon that I was looking at (when not clicking the shutter) did not change.

I was shooting raw, so the white balance should not affect the raw data, but I am unable to reconcile these two images, even if I force a color temperature in Nikon's raw conversion software.

The images were shot on a rather murky evening, with high extinction. I had traveled to the country to get a clear, low horizon. Unfortunately, I left my tripod adapter at home, and so had to shoot at ISO 1600 to get an adequate exposure when the Moon was as low as 1 degree.

I converted the image in Nikon Capture, but made no adjustments other than those automatically applied. In Nikoneeze that means a "standard" picture control, auto WB, ADL set to "Normal") The ISO, aperture and exposure settings were identical. The Moon's altitude increased from 4.1 to 4.4 degrees in the two minute interval, which should have brightened the Moon by 0.6 f-stops. The measured increase in brightness in the green is 0.4 f-stops, a satisfactory agreement in my opinion.

The color as I recall it was closer to that of the second image, but it is insufficiently red. The first image color is not correct.

I measured the brightness of a spot near the disk center, which is in the lunar highlands. As you can see from the readings in the picture, the the green reading increased a small amount while the red dropped by 23% and the blue more than doubled.

So, what is going on here? I have only one thought, and it is insufficient to explain what is going on. Can anybody offer some thoughts on this?

My one thought: There were a few spots on the first image where the red is saturated, and that seems to have driven the white balance out of balance, so to speak. But I estimate them at only a few percent of the area.

I am going to cross post this to the Nikon FX forum. I know that some of you are in both, but there may be some Nikon gurus who dont read this forum. And many of you here don't shoot Nikon, but I have a hunch that this is not a Nikon specific issue.

(And as long as I have your attention, let me suggest you check out some of my Moon in the Landscape shots at www.ScientiaPhoto.com in the Hawaiian Landscapes collection.)

-- hide signature --

Bob in Baltimore

 Bob in Baltimore's gear list:Bob in Baltimore's gear list
Nikon D610 Nikon AF-S Nikkor 70-300mm f/4.5-5.6G VR Nikon AF-S Nikkor 24-85mm F3.5-4.5G ED VR
NancyP
Senior MemberPosts: 4,186Gear list
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to Bob in Baltimore, Aug 21, 2013

Wouldn't this be due to garden variety differential scattering of red vs blue/green at very low angles (lots more atmosphere to go through relative to overhead)? Why should this be different from sun rise and sun set?

-- hide signature --

NancyP

 NancyP's gear list:NancyP's gear list
Sigma DP3 Merrill
Reply   Reply with quote   Complain
Bob in Baltimore
Regular MemberPosts: 439Gear list
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to NancyP, Aug 21, 2013

NancyP wrote:

Wouldn't this be due to garden variety differential scattering of red vs blue/green at very low angles (lots more atmosphere to go through relative to overhead)? Why should this be different from sun rise and sun set?

-- hide signature --

NancyP

I don't think so. It is too abrupt. The change in altitude was only a few tenths of a degree & 2 minutes.

-- hide signature --

Bob in Baltimore

 Bob in Baltimore's gear list:Bob in Baltimore's gear list
Nikon D610 Nikon AF-S Nikkor 70-300mm f/4.5-5.6G VR Nikon AF-S Nikkor 24-85mm F3.5-4.5G ED VR
Reply   Reply with quote   Complain
Bob in Baltimore
Regular MemberPosts: 439Gear list
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to NancyP, Aug 21, 2013

NancyP wrote:

Wouldn't this be due to garden variety differential scattering of red vs blue/green at very low angles (lots more atmosphere to go through relative to overhead)? Why should this be different from sun rise and sun set?

-- hide signature --

NancyP

Nancy, here is a more complete response.

Yes, it is the same phenomena as we see at sunrise. The scattering and absorption in the red is both stronger and more rapidly changing than it is in the green and the blue. But I have a sequence of shots every minutes, and the change is abrupt. There is little change in the 18 minutes preceding and the 20 minutes following. I didn't hit you with all the data, but it is there.

-- hide signature --

Bob in Baltimore

 Bob in Baltimore's gear list:Bob in Baltimore's gear list
Nikon D610 Nikon AF-S Nikkor 70-300mm f/4.5-5.6G VR Nikon AF-S Nikkor 24-85mm F3.5-4.5G ED VR
Reply   Reply with quote   Complain
wfektar
Contributing MemberPosts: 694
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to Bob in Baltimore, Aug 21, 2013

I'm not familiar with your raw converter so I really don't know, but I agree that this is likely a software issue, not a physical one. But just to argue the other way, just as blue skies turn green when the blue channel saturates, shouldn't the image turn increasingly cyan as the red channel clips? Or does the converter attempt to compensate when it senses channel saturation?

I can come up with scenarios that would turn the moon red but not one that increases the red value for a given exposure (that is, the blue and green are suppressed) -- especially when the red decreases with lunar zenith angle. The eyeball is a terrible photometric device so I shouldn't even ask, but did the luminance of the moon look about the same when you took the two shots?

How's that for a long-winded nonanswer?

Reply   Reply with quote   Complain
Bob in Baltimore
Regular MemberPosts: 439Gear list
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to wfektar, Aug 21, 2013

wfektar wrote:

The eyeball is a terrible photometric device so I shouldn't even ask, but did the luminance of the moon look about the same when you took the two shots?

How's that for a long-winded nonanswer?

I used exposure compensation in Nikon's Capture NX to bring a selected spot on the Moon to the same value for each image, & I combined that with the EXIF values for ISO, aperture and shutter speed, and a correction for atmospheric extinction derived from almost an hours worth of data to calculate the luminance of the Moon for each image if it were taken in the zenith.

The 1st Moon was .3 stops brighter than the second. (It should have been slightly dimmer.) The standard deviation of a single measurement was 0.24 EV. That amounts to saying that the luminance was pretty much unchanged. (Adjustments were made in the green band to avoid problems with near saturation in the red and the higher variability in the blue.)

Hows that for a long-winded answer to a short question? (Yours, BTW, was a helpful response as you tussled with the problem.)

-- hide signature --

Bob in Baltimore

 Bob in Baltimore's gear list:Bob in Baltimore's gear list
Nikon D610 Nikon AF-S Nikkor 70-300mm f/4.5-5.6G VR Nikon AF-S Nikkor 24-85mm F3.5-4.5G ED VR
Reply   Reply with quote   Complain
wfektar
Contributing MemberPosts: 694
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to Bob in Baltimore, Aug 22, 2013

Ah -- I missed the exposure compensation stuff. All bets are off, then. Just out of curiosity, what happens when both are processed identically? No EC, same gamma, fixed WB.

Reply   Reply with quote   Complain
NancyP
Senior MemberPosts: 4,186Gear list
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to Bob in Baltimore, Aug 22, 2013

Right you are. I didn't check the time stamps.

-- hide signature --

NancyP

 NancyP's gear list:NancyP's gear list
Sigma DP3 Merrill
Reply   Reply with quote   Complain
Bob in Baltimore
Regular MemberPosts: 439Gear list
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to wfektar, Aug 22, 2013

wfektar wrote:

Ah -- I missed the exposure compensation stuff. All bets are off, then. Just out of curiosity, what happens when both are processed identically? No EC, same gamma, fixed WB.

In this case, the pix look pretty much the same and the R/G/B values are similar. I have had other cases (I think!) where I could not bring them to look alike when I forced them to the same WB. Of course, I failed to document just when that happened.

-- hide signature --

Bob in Baltimore

 Bob in Baltimore's gear list:Bob in Baltimore's gear list
Nikon D610 Nikon AF-S Nikkor 70-300mm f/4.5-5.6G VR Nikon AF-S Nikkor 24-85mm F3.5-4.5G ED VR
Reply   Reply with quote   Complain
star shooter
Regular MemberPosts: 188
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to Bob in Baltimore, Aug 23, 2013

Bob,

There's no real point in shooting the moon in colour. It has no colour. Best if you switched over to monochrome and use a red filter. This will reduce the haze in the atmosphere and give the image more contrast.

Full Moon - Down Under. Sigma 120-300/2.8 boosted by a 1.4 and 2X converters. ISO 200 exp. 1/250.

Reply   Reply with quote   Complain
wfektar
Contributing MemberPosts: 694
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to Bob in Baltimore, Aug 23, 2013

Well then, it has to be a camera or converter problem. The problem is less the color change (problematic enough) but significant red extinction with marginally shorter atmospheric column. And it's a change that's complete within 2 minutes or <2/3rds of a lunar diameter. There's also blue amplification that's much larger than green, for which you'd need to invoke something like a puff (as the modelers call it) of NO2 or similar that disappeared within those 2 minutes. If it all converted to nitrate (which is stable at night and absorbs in the red) in that time span maybe but that's so far fetched it's in not gonna happen land. Besides, with the later shots when the column has appreciably changed you'd see the difference. Wonder what the raw channel data look like.

Reply   Reply with quote   Complain
just Tony
Contributing MemberPosts: 692
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to Bob in Baltimore, Aug 27, 2013

Visual: The color actually changed more than you perceived it did because of the auto white balancing of your eye/brain system. You can't trust your eyes completely for color measurements.

Photographic: What was going on is that there are hard limits to how much adjustment is available in the auto white balance function. Your earliest exposure was beyond the limit, but then a bit later the color fell inside the range where AWB could follow the color changes.

I bumped into that limit when shooting a video in a yellow-painted room lit by warm fluorescent lights. The AWB came up with a terrible color result. I then attached an 80A blue tungsten color conversion filter to the lens; again using AWB, the color was really quite good then.

Next time don't let the camera make any decisions at all. Shoot raw with a the daylight setting for WB (this doesn't change the images but it does control how the images first appear in your editor), and use identical non-auto settings for the raw conversion of all images.

Reply   Reply with quote   Complain
TJL LTFF
Senior MemberPosts: 1,239Gear list
Like?
Re: Moon Photography: Auto White Balance Mystery
In reply to star shooter, Aug 27, 2013

star shooter wrote:

Bob,

There's no real point in shooting the moon in colour. It has no colour. Best if you switched over to monochrome and use a red filter. This will reduce the haze in the atmosphere and give the image more contrast.

For most of us your point is completely valid; but the moon has various colors, just very subtle. But there is information in those various colors. Going monochrome, in a very high quality photo, could cause a slight loss of very subtle but real detail. I have never taken a photo of the moon where this would matter (at least I don't think I have); but I have seen some examples where it made a pretty cool little difference.

By the way, I like your high contrast photo with the red filter; interesting take; thanks for sharing.

 TJL LTFF's gear list:TJL LTFF's gear list
Samsung NX20 Samsung NX300 Samsung NX 30mm F2 Pancake Samsung NX 18-55mm F3.5-5.6 OIS Samsung NX 20mm F2.8 Pancake +6 more
Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads