Histogram Moving To The Left Throughout The Night. Help.

[Ross1204]

Hi,
I'm posting this question in the hope I can gain a little clarification on what causes the histogram on my exposures to gradually move further to the left throughout the night. I have saved a few screenshots below which show the clear shift. 

My current thinking is that it could be either light pollution, accidental setting change in the camera, thin cloud (could still see stars though) or bad seeing conditions?

However the last image though is using the same ISO and exposure length as the two before but as you can see the histogram is way over to the right becoming too overexposed. This one is a little more confusing in some ways as not sure why as the above stated settings are usually fine.

Any advice would be much appreciated?

 

1-11-18 - 21.53pm - Sussex - Pointing South/South East

 

Histogram - 2-11-18 - 2.45am- Sussex - Pointing South/South West

 

Histogram - 8-01-19 - 7.08pm - Sussex - Pointing South/South East

Telescope - Skywatcher
Mount - HEQ5 Pro
Camera - Canon 700D (Astro Modified)  - No Filter 

[symmetal]

The histogram peaks correspond to the level of the sky background. At 02:45 the sky will be darkest so the histogram shows the lowest level of background. At 7:08 pm there will still be some skyglow from the Sun even though it may be classed as Astro Dark. Astro Dark isn't a fixed level of darkness but is a threshold value below which it's considered dark enough for Astro. For a couple of hours after the Astro Dark threshold start time the sky will get darker.

Alan

[Stargazer33]

The two lighter images which seem to show light pollution are both pointing SE. The other one is pointing SW. and is after 2am. The lights in our road are turned off after 1am. If this is similar in your area the sky would be darker at 2am than at 7pm. Also I would imagine that you have a more built up area to your SE.

[Stub Mandrel]

It also depends on the transparency, if there is thin cloud and you have sources of LP anywhere (e.g. you are in England) it will increase the level of the background even if stars are barely affected.

[Ross1204]

Thank you all so much for replying and helping me understand what was causing this change in the histogram. Every single point makes perfect sense, relieved man!

With that said what would be best way to shoot going forward?

• Lower ISO?
• Shorter exposure time?
• Purchase filter?

I have generally been really happy with 3min subs at ISO800 ,but now seeing a overexposed histogram with these settings does make me think I may need to change it up, cos last time out I didn't really get much usable data.

Thanks again

Ross

[symmetal]

As the background skyglow (the histogram peaks) is the faintest object detail you can resolve, once this skyglow peak is not clipped to the left of the histogram there is little to gain by exposing for longer. The signal to noise will be improved to a degree, but your dynamic range will decrease and more stars will be clipped and bloated. In CMOS astro cameras it's normal to expose long enough for the camera read noise to be swamped by the skyglow and then take many exposures at that exposure duration. This is dependent on the gain and offset of the camera. For me, with the ASI1600, the swamping skyglow level is around 1400 ADU (in 16 bit) This would correspond to only around 2% on the histogram (linear display) For luminance exposures at unity gain this can be as short as 15 seconds in light polluted skies. I have fairly dark skies so can expose up to about 60 seconds before the skyglow limit is reached. With your one-shot colour  DSLR these exposures would be about three times longer.

I would aim initially for the centre of the histogram peak to be around the 15-20%% mark on the APT histogram (in linear mode), but not clipped on the left. On your darkest 02:45 image it's shown at around 30%. This would mean shorter exposures than your current 3 minutes at ISO800 or dropping the ISO. ISO800 is generally considered optimum for noise on Canon DSLRs so I would leave it at 800 and drop the exposures to 2 minutes or 90 seconds and so enable you to take more exposures to stack in the same imaging session. This will give you less star clipping, a larger dynamic range, and the increase in number of exposures should improve your noise when stacked.

As your background skyglow, shown by the histogram peaks is fairly neutral in colour I don't believe a light pollution filter (if that's the sort of filter you had in mind) will achieve much. As your DSLR is modded you could try narrowband Ha which will allow you much longer exposures before you're limited by the skyglow.

Other people may have other views and disagree with me, but that's my pennyworth. 

Alan

[Stub Mandrel]

6 hours ago, symmetal said:

As your background skyglow, shown by the histogram peaks is fairly neutral in colour

The Plieades should have a preponderance of blue, but G &R seem to be up, which suggests sodium light pollution to me (it sits pretty much on the overlap between R & G curves). A cheap LP filter is only about £15 so I would say is worth a try.

[Ross1204]

Thank you all for taking the time to help, very grateful.

So, I reckon next time the clouds break giving me a chance to get out there I shall stick to ISO800 and just shorten the exposure times. I have previously exposed at ISO200 for 180s and the resulting data was full of noise and lacking detail compared to all the other nights. Hopefully shortening the exposure times does not kill the camera's ability to pick out the faint details in DSO's.

Thanks again.

Ross   

[symmetal]

4 hours ago, Ross1204 said:

I have previously exposed at ISO200 for 180s and the resulting data was full of noise and lacking detail compared to all the other nights. Hopefully shortening the exposure times does not kill the camera's ability to pick out the faint details in DSO's.

Like you when I was taking short exposures with the ASI1600 the single frames did look very noisy and flat compared with longer exposures on my CCD and wasn't hopeful that stacking many more frames would make up for it but it does seem to work.  The detail which seems 'lost' in the noise does re-appear when stacking 50 or more frames. So sticking to ISO 800 and taking 3 times as many 1 minute subs in the same period as your 3 minute subs, I think you will be surprised at the result. Just ensure your histogram isn't clipped on the left (on any colour) with the shorter subs, so that the skyglow signal  is significantly more than the read noise.

Alan