measuring dynamic range of camera for skyglow experiment

[h2olilykh]

In doing the skyglow experiment---
Have followed the procedure exactly...filled the camera's view with the white paper in indirect sunlight, took pictures with aperature at 2.8, ISO at 200, and series of photos at shutter speeds from 10 to 1/4 by factor of 2. I also took 2 pictures at each shutter speed to use as control to check my findings--as the experiment suggested.
When I ran them through the Image J program, ALL the histograms for each shutter speed came up with exactly the same mean value.
Looking at the example graph in the experiment of average pixel gray level vs. shutter speed this isn't what should have happened.

I did take other pictures of the white paper in a different location--following the same procedure as above. The mean was different; however, still all the same exact value for that location.

The pictures of the clear night sky, at different times of the night, but in the same location, did come up with different mean values--so is not the camera.
Any ideas?
Thanks

[deleted-71882]

Hello h2olilykh,

You certainly have a puzzling result. (Also a clever user name.) I'll try a guess as to what is happening.

Since you get the same mean value, something has to be making the camera automatically adjust itself to get that result. Please check carefully to see if you have some automatic feature of the camera turned on. The camera has to be in the manual exposure mode.

My camera and most inexpensive digital cameras don't have a manual mode: it's always automatic, so I couldn't use my camera for this experiment.

Suggestion: Try taking a picture of anything at different apertures and exposures. Can you find a small aperture and exposure time that results in a black or very dark image? If yes, then it's in manual mode. If no, then it definitely is doing the exposure automatically.

It's possible that pictures of the night sky are also being taken in automatic mode, but the night sky isn't light enough to allow the automatic exposure to correct for the low light level.

Please try the Suggestion and let me know if it helps to diagnose your problem, WW

[h2olilykh]

Sorry, I did not include that my camera does have a manual mode, and it was in it for both parts of the experiment. The white paper, and the shots of the night sky at different times. I left it that way so as not to include a different variable in the experiment.
The only thing I changed in the manual mode was the shutter speed before taking the pictures.
Thanks for answering so quickly!!!
I did try changing the shutter speed and aperature and can get a black picture of a white floor in the sunlight.

(It's amazing how many people don't get the user name!!!)

[deleted-71882]

h2olilykh,

Put the black picture into Image J and look at the mean value. If the black picture still gives a mean value at the middle of the range, then either Image J is wrong or you're interpreting it incorrectly. Obviously, a black picture should have a mean value near the bottom of the allowable range, probably zero.

If the black picture gives a low mean value, then look at all your pictures of the white paper. Are any of them black? Then they have to have low mean values. If they are all well illuminated, then somehow you're not doing all the same things as when you got the black picture. Do you follow this reasoning?

I think there's some little detail here that's being missed. You seem to have a good grasp of the general ideas.

Let me know how it goes, WW

[h2olilykh]

All the pictures of the white paper were the same color-no matter what the shutter speed was. The pictures were taken as the experiment stated to do--indirect sunlight, fill the view finder.....as I said above.

The dark picture I took at your suggestion today had an appropriately low mean in image J. To get this picture I changed the aperature and shutter speed to make sure there was no automatic feature involved, and to get the settings dark enough to make a dark picture.

Because the paper was white, and the aperature was wide open at 2.8, the fastest shutter speed I used was 1/4. Should I have started less than 10 and gone down to 1/1000--faster shutter speed-- to see a change in the picture--and therefore maybe a change in the Image J findings?

[deleted-71882]

h2olilykh,

You wrote, "All the pictures of the white paper were the same color-no matter what the shutter speed was."

Yes, you should have used exposures shorter than 1/4 sec. I think that will solve your problem.

Since all the pictures of the paper were white, that means that you need to reduce the aperture and exposure time to capture the transition from black to white. Before you take the series of pictures to analyze digitally just try different apertures and exposure times and look at the pictures.

• Set the camera to an aperture halfway between the maximum and minimum that your camera has.
  Take pictures at all the exposure times the camera has and look at the result.
  Note what exposure time is at the transition from black to dark gray.
  Note what exposure time is at the transition from light gray to white.
  If you can't get either complete black or complete white, adjust the aperture and try the range of exposure times again.

Now you know what settings will allow you to get the range you want. Record the series of exposures and do the analysis in Image J.

Hope this works, WW

[h2olilykh]

By changing the aperature setting to different from what the experiment stated to do (2. that would add a variable to the experiment, right?
Isn't this part of the experiment the control-- to verify the night-time pictures measurement for skyglow?

(The night time pictures are taken at 2.8 and ISO of 200 just like the pictures of the white paper.)

Thanks

[deleted-71882]

h2olilykh,

Yes, you're right that changing the aperture introduces a new variable. Use 2.8 and take exposures down to the smallest exposure time your camera allows. If the image is still not black then you can reduce the aperture.

Keeping the aperture fixed allows you to perform the experiment with just one variable: exposure time. It could also be done using aperture and exposure, but I think that calculating those factors may involve math beyond where you are.

Stick with 2.8. Let me know if you can't adjust the exposure time to get the full black-to-white range.

WW