Relationship between read noise in e/rms to sky electron rate in e/pixels/s

[Giles_B]

I'm delving into the dark arts of LRGB imaging with a 1600MM Pro and trying to work out optimal exposure times, guided by this post https://www.dpreview.com/forums/thread/4377640

Having taken some test shots I've determined my SQM=18.5 [Bortle 7]

Using the calculator linked to in the post, I've determined my sky electron rate in e/pixels/s for LRGB

L filter = Sky Electron Rate 8.99 e/pixel/s
RGB = 3 e/pixels/s

 

However, when I look at the technical data for the 1600MM I get either a base read noise of 1.2e or a graph that plots read noise in e-rms against gain (below). I can't work out how to relate the e/pixels/s number to e-rms, and the more I read the more confused I'm getting - could someone walk me through an answer?

 

[The Lazy Astronomer]

2 hours ago, Giles_B said:

I'm delving into the dark arts of LRGB imaging with a 1600MM Pro and trying to work out optimal exposure times, guided by this post https://www.dpreview.com/forums/thread/4377640

Having taken some test shots I've determined my SQM=18.5 [Bortle 7]

Using the calculator linked to in the post, I've determined my sky electron rate in e/pixels/s for LRGB

L filter = Sky Electron Rate 8.99 e/pixel/s
RGB = 3 e/pixels/s

 

However, when I look at the technical data for the 1600MM I get either a base read noise of 1.2e or a graph that plots read noise in e-rms against gain (below). I can't work out how to relate the e/pixels/s number to e-rms, and the more I read the more confused I'm getting - could someone walk me through an answer?

 

As far as my understanding goes, the e-rms figures given can be used in place of e-/px/s. Essentially, each pixel of a CMOS camera has its own circuitry and therefore its own independent read noise figure. The rms figures are a way of averaging these to give a number for the sensor as a whole.

That said, do not get too obsessed with optimal sub exposure time. In high light pollution areas, you'll likely find the theoretical optimal time is <5s which, frankly, is impractical for DSO imaging. Spend 10 or 15 minutes one night doing some test exposures of somewhere between 30s - 5m and look at the image histogram to check it's not too far lifted from the left. Decide what your best compromise is between a good exposure, mount tracking ability, and storage space.

Personally, when I do mono (with a 294mm), I use 2m L and 5m RGB (bortle 5-ish).

[Giles_B]

Many thanks. As you say, the exposure time does seem very small. I'd previously been experimenting with 20s L exposure and 40s RGB, but thought I should try to be a bit more scientific. The experiments are a bit rubbish as tracking was off, but they should still be useful to check the histograms and see if I can find a good compromise between the science and the practice!

[Clarkey]

As @The Lazy Astronomer says, in Bortle 7 the read noise will be swamped in seconds. In reality you just need to choose a sub length that allows you to take enough total integration time without needing a super-computer to process. Personally, in my B6 back garden, I use 2 minute subs for most targets and 3-5 minutes for NB.

Have you watched the Robin Glover video about AP. Well worth a look if you haven't.

 

[900SL]

17 hours ago, Giles_B said:

Many thanks. As you say, the exposure time does seem very small. I'd previously been experimenting with 20s L exposure and 40s RGB, but thought I should try to be a bit more scientific. The experiments are a bit rubbish as tracking was off, but they should still be useful to check the histograms and see if I can find a good compromise between the science and the practice!

I use the mean/median histogram values for a trial sub as shown in the ASI Air app. There might be some useful posts here

 

[Giles_B]

1 hour ago, Clarkey said:

Have you watched the Robin Glover video about AP. Well worth a look if you haven't.

 

Thanks - I hadn't watched it but will do so tonight

[symmetal]

Here's a chart I made for the ASI1600 which takes all the maths out of it. 🙂 If you expose until the read noise from the sky background is around 5x the read noise then the read noise itself becomes insignificant. At that exposure duration you may as well start another exposure as there is nothing to be gained from exposing for longer.

You can easily get your image sky background ADU by hovering the mouse over the background of your image in your capture program or use the image median value, (as long as there is not a lot of nebulocity present).

The average (mean) ADU value of your master bias frame is also needed, so that the actual effect of camera offset can be incorporated into the results. If you're using offset 50 then the chart should be fairly close but I've included the Excel file below so you can download it and fill in your actual bias value if you wish to get a more accurate result.

The exposure calculator is a quick way to get the optimum exposure from a sample image. Just fill in its sky background ADU and exposure time, the required ADU (like 1965 from the chart) and it will calculate the exposure to achieve that.

Different filters will need different exposure times to achieve this optimum background ADU of course. Luminance will be around 1/3 that of RGB, and narrowband will be much longer.

The chart has been approved by vlaiv who is a proponent of 5x read noise swamping for optimum exposures. 😃 The 3.16 swamping factor is a holdover from early threads on the CN forum, but as vlaiv pointed out 5x is a better figure.

ASI1600, Sky Background ADU v2.xlsx

Alan