Using the moon as a grey card for colour balance?

[xpatUSA]

The intention was to try and capture the eclipse, and as the eclipse proceeded I needed to increase the exposure level so the camera captured enough light for an image.

A pity that the large change in exposure threw color accuracy out the window, rendering any discussion of color balance quite moot.

No colour modification was performed post capture (aside from some increase in saturation levels for the eclipsed moon).

In my world outside of the pale, an "increase in saturation levels" is a change of color.

[Tulloch]

A pity that the large change in exposure threw color accuracy out the window, rendering any discussion of color balance quite moot.

Not sure I understand here - are you saying that just because the EV levels were increased, the colour balance is no longer accurate?

In my world outside of the pale, an "increase in saturation levels" is a change of color.

Fair enough, however all colour channels were increased equally.

[xpatUSA]

Not sure I understand here - are you saying that just because the EV levels were increased, the colour balance is no longer accurate?

Yes, Andrew, it seems that the increase in saturation caused some bottoming because some areas in the posted image have zero channel levels and generally the red channel became dominant.

Hue map of the sequence:



Speaks for itself, I think.

Fair enough, however all colour channels were increased equally.

Looking at the above hue map, I'd say they were not or, if they were, the increase in saturation messed up the color.

HTH ...

[Manfred M]

Tulloch - in an RGB colour model, every time you change one of the values, you have a different "colour". A nasty side effect of the RGB models is that every time you change one of the values, you also change the saturation of the colour; you may not notice it with small changes, but move all three values equally higher or lower, you are in effect showing more saturated colours or more desaturated colours.

This is one reason I sometimes work in the L*a*b* colour space; where I can work with just the luminance channel and not impact the a* (magenta - green) or b* (blue - yellow) channels, i.e. changing the luminance only without impacting the saturation.

There are a number of different ways of describing colours; RGB is one, CMY(K) is another, HSL / HSV is another. Each has its strengths and weaknesses.

While most people don't realize it, many raw convertors are set up that way. Exposure somewhat relates to L* (luminance) while the Temperature is the *b channel and Tint the *a channel.

Human vision is non-linear, so we don't notice changes in the darks to quarter tones or three-quarter tones to lights as much as in the mid-tones, which is what Ted seems to be referring to in when he says "outside the world of pale". The human visual system is weakest in differentiating the lightest colour values.

[xpatUSA]

While most people don't realize it, many raw convertors are set up that way.

Exposure somewhat relates to CIELAB's lightness L*; Temperature relates to the b* channel; and Tint relates to the b* channel.

Thanks, Manfred, I had never considered that relationship ... (pardon the pedantic rewording).

[Tulloch]

Yes, Andrew, it seems that the increase in saturation caused some bottoming because some areas in the posted image have zero channel levels and generally the red channel became dominant.

Hue map of the sequence:



Speaks for itself, I think.

Really? That's a 300 kB 8-bit jpg rendition from a series of 12 x 55MB 16-bit png files made from multiple stacked images of the moon, reduced in size to fit on the Cloudy Nights website. The original images are far too large to post anywhere on the web, but the final step in the process is reduce the bit depth and convert to jpg format for posting on the web.

[xpatUSA]

Really?

Really.

[Tulloch]

I don't.

And I'm not trying to colour balance the moon either. I'm trying to work out how to manually process a raw image from a DSLR/astro camera without having a grey card.

[DanK]

I've been following this thread from a distance, and I no longer can see what it's about.

This is from the initial post:

While it is possible to point the astro-camera at a Macbeth colour chart and produce a colour calibration matrix, this assumes a particular colour temperature which is probably not correct for night-time imaging (ie even though the light from the sun passing through the atmosphere is dispersed in the same way, the sky is not blue at night and so doesn't contribute to the temperature of the light source).

One possible way to calibrate these cameras is by using the moon as a pseudo-grey card target.

That certainly sounds like color balancing to me. But this is from post 28:

And I'm not trying to colour balance the moon either. I'm trying to work out how to manually process a raw image from a DSLR/astro camera without having a grey card.

But in the hopes that I am still partly following this:

--Resizing the image (in this case, shrinking it) should not substantially affect color.
--Changing ISO from 100 to 800 shouldn't affect overall color.
--Changing the shutter speed should not affect color, assuming that overall exposure is kept constant.
--any tonality adjustments in RGB space will affect saturation, as Manfred noted.

I'm completely mystified by the discontinuities in color in post #15.

So, my questions are:

1. What is this thread about, if it's not color balance?
2. What happened to create the discontinuities in post #15?

[xpatUSA]

2. What happened to create the discontinuities in post #15?

Dan, I did notice that, in the four pink moons, there was a plentitude of non-red channel values of zero or close thereto which can occur with applied saturation (deliberately or accidentally). Even if shot at the commonly-suggested 4100K, I would not expect any zero channel values in a properly-exposed moonshot - eclipsed or not.

[DanK]

Dan, I did notice that, in the four pink moons, there was a plentitude of non-red channel values of zero which can occur with generous amounts of applied saturation (deliberately or accidentally). Even if shot at the theoretical 4100K I would not expect any zero channel values in a moonshot - eclipsed on not.

Ah, that's interesting.

I'd have to guess about the 700D, but in the case of the Canon 5D IV (the only Canon for which I have the manual on my laptop), the daylight setting is 5200K. So, if the theoretical value of 4100 were correct, using 5200 would cause a shift to the warm side in the processed image. If the capture were also badly underexposed, couldn't that cause the blue and possibly the green to bottom out? Then amplify to compensate for exposure...

Just off the top of my head.

[xpatUSA]

Ah, that's interesting.

<> If the capture were also badly underexposed, couldn't that cause the blue and possibly the green to bottom out? <>

Dan, He went from 100 to 800 ISO which by itself would indeed underexpose the sensor by -3EV.

But he also cranked the shutter time from 1/1000 to 0.6sec which is over +9EV. In other words, the pink moons were exposed (at the sensor) 6EV more than the normal-looking gray moons.

HTH.

[DanK]

Dan, He went from 100 to 800 ISO which by itself would indeed underexpose the sensor by -3EV.

But he also cranked the shutter time from 1/1000 to 0.6sec which is over +9EV. In other words, the pink moons were exposed (at the sensor) 6EV more than the normal-looking gray moons.

HTH.

Indeed. I should have caught that. Then I'm completely mystified.