I recently purchased a set of Baader 50.4mm LRGB filters to get me through galaxy season. Up until now I mainly did narrowband nebula photography.

Following the presentation from Robin Glover (SharpCap author) I calculated the optimum frame duration for my level of light pollution, scope and filters. The result was a 5 second frame duration for Luminance.

(see https://stargazerslounge.com/topic/332167-talk-by-robin-glover-sharpcap/ for the presentation)

So I shot this target with:

* 100 5-seconds luminance

* 50 10-seconds R

* 50 10-seconds G

* 50 10-seconds B

When looking at the resulting images, it became clear that this short subs may give you a workable SNR, but it resulted in hell to calibrate the images.

Also, I found that doing longer subs revealed more details in the galaxies and increased the SNR. Anyhow, I do not think this is the most efficient way of collecting data.

I may have used Robin's formulas wrong of course.

I had huge problems getting rid of amp glow patterns of the Atik Horizon which were almost impossible to get rid of.

My theory is that due to the shortness of the subs (and thus also the darks), the amplifier of the camera gets quite hot. Since the CCD is kept at constant temperature, but not the other electronics, the amp glow will be dependent on the ambient temperature.

I collected the flats and the darks during the day and was experiencing a slightly different amp glow in my calibration data.

Anyhow, here I was with 250 very short subs and I was not willing to give up on the data.

So what did I do to make the glow get to a bearable level:

* I took dark flats for all flat durations to exactly calibrate my flats (even with the light box set to minimal level the RASA needs just approx.1 second for a luminance flat)

* I stopped using master bias and master dark for calibration. Instead, I used averaged darks (so I did not remove the bias signal). In this way, I could avoid getting black pixels in my calibrated subs

The resulting glow pattern was then minimized by background extraction. Still I needed to apply pretty aggressive curves to get an even background.

All in all: LOTS of gigabytes and time for a mediocre result.

Is anybody else using Robins formulas to calculate optimum sub duration? Did it work out for you?