Gain and Exposure Time for ASI1600MM & LRGB Imaging

[STEVE333]

Hi All - I have an ASI1600MM cooled monochrome sensor that I've been using for narrowband (SHO) imaging. It is working well. Now I want to try some LRGB imaging with an F/5 telescope, but, I don't know what Gain/Exposure to use. I live under Bortle 5 skies.

Any suggestions?

Thanks in advance,

Steve

[JayTee]

Hi Steve,

Does your graph in this previous post of yours not answer your own question? If not, what piece of the puzzle are you missing?
viewtopic.php?t=8035

Cheers,
JT

[STEVE333]

Hi Steve,

Does your graph in this previous post of yours not answer your own question? If not, what piece of the puzzle are you missing?
viewtopic.php?t=8035

Cheers,
JT

Good call JT. The graph in the previous post was produced early on in my noise investigations. I used a more in depth analysis approach when figuring out my NB exposure times. That same analysis can be used to determine the exposure times for the LRGB filters. Will need to measure the skyglow for the L, R, G and B filters, but, that won't take much time. I guess I was hoping to save a little analysis time, but, you set me straight. Thanks!

I'll post the results when the analysis is complete.

Hopefully there will be some clear sky tomorrow night after the rains clear.

Cheers,
Steve

[STEVE333]

I did some preliminary analysis using estimates of the skyglow. The results indicated that about 200 sec exposures might be a good choice with the gain set at "0". However, I need to measure the skyglow to confirm the results. The graph below shows the results of the preliminary analysis.

The vertical axis shows the "goodness" of the master stacked image. The horizontal axis is the Exposure Time of the individual exposures. The "goodness" of the master stacked image has very little dependence on which ASI1600MM gain is chosen. Therefore I am leaning toward the "0" gain to prevent star saturation. The "goodness" of the master stacked image increases quickly with increasing Exposure Time up to an Exposure Time of about 100 sec (the knee of the curve). Increasing the Exposure Time beyond 200 sec produces only slight improvements in final image quality while it will cause more stars to saturate. The preliminary analysis result is:

• Gain = 0

• Exposure Time = 200 sec

It will be interesting to see if using the measured skyglow will significantly alter these preliminary results.

Cheers,

Steve

[STEVE333]

I measured SkyGlow for the R, G and B filters and redid the analysis. The result for my system is below.

The meaning of the graph is as follows:

• The vertical axis is the Image Quality (measured using Signal-to-Noise). The horizontal axis is the Exposure Time used for each image.

• Consider that you are taking images for three hours. If you used 60 sec exposures you would collect 180 images (neglecting dithering or download times). If you used 180 sec exposures, each individual image would be better than the 60 sec exposures, but, you would only have 60 images.

• When you process the images they will be stacked to produce one master Light image. Is the quality of that master Light image the same for the 60 sec exposures and the 180 sec exposures? The answer is NO! I won't bore you with the details, but, the longer exposures always produce a superior result. However, the improvement with longer exposures doesn't keep improving, but, reaches an asymptotic limit. As seen from the graph above the "knee" of the curves is at an exposure of about 200 sec. Longer exposures will provide only a small improvement, but, they will cause more stars to saturate. My choice of 200 sec is a tradeoff between better image quality and fewer saturated stars.

Hope this makes sense. By the way, I used a 200 sec Exposure time and Gain=0 (based on these analysis results) for my recent RGB image of the Leo Triplet which was posted in the DSO sub-forum. The results look good to me.

For those who are interested, in the analysis the effects of dithering time was included.

Feel free to contact me if you want more of the details.

Cheers,

Steve

[STEVE333]

In the analysis above I forgot to take into account the sensor gain (e-/count) [not the Gain setting which I chose as zero. For the Gain = 0 setting the sensor gain = 5 (e-/count). With the correction the graph looks as shown below:

What the graph shows is that the quality of the stacked image using 200 sec exposures is within 5% of the theoretical best that can be achieved. A 5% difference in noise level will be barely noticeable. However, the 200 sec exposure will leave many of the stars unsaturated yielding good star colors when the RGB channels are combined.

The effects of dithering is also included. The effect of dithering is to reduce the amount of time available to take images. The average time between images when dithering occurs is about 60 sec. However, I only dither once for every three images, so, the average dither time is 60 sec/3 = 20 sec.

I know this is a lot of explanation. However, the key point is that increasing the Exposure Time beyond the "knee" of the curve will NOT produce much improvement in stacked image quality (but will saturate more and more stars).

I should mention that these curves are for the ASI1600MM sensor with RGB filters and with my F/4.9 WO Star71 telescope and Bortle 5 skies. The curves would be different for different F/# telescope and/or different light pollution conditions.

Thanks for looking.

Steve

[STEVE333]

One final thought!

Exposure times longer than the exposure time at the "knee" of the curve (call this Tknee) won't provide much improvement in the final stacked image. However, using an Exposure Time of Tknee may still be saturating too many stars. If that's the case, then, your only choice is to make the tradeoff between image quality or fewer saturated stars. Life is full of tradeoffs, and, this is one of them!

At least the curves point to a good starting point (Exposure Time = Tknee or slightly longer).

Cheers,

Steve

[AstroBee]

I should mention that these curves are for the ASI1600MM sensor with RGB filters and with my F/4.9 WO Star71 telescope and Bortle 5 skies. The curves would be different for different F/# telescope and/or different light pollution conditions.

Steve, I'm following this with great interest because as you know, I too have the 1600mm.
So my question at this point, understanding (sort of) everything you've mentioned, is HOW do I do this same analysis for my system and sky conditions?

[STEVE333]

I should mention that these curves are for the ASI1600MM sensor with RGB filters and with my F/4.9 WO Star71 telescope and Bortle 5 skies. The curves would be different for different F/# telescope and/or different light pollution conditions.

Steve, I'm following this with great interest because as you know, I too have the 1600mm.
So my question at this point, understanding (sort of) everything you've mentioned, is HOW do I do this same analysis for my system and sky conditions?

I will be glad to do the analysis for you (it would be tedious for you to do by hand!). I have a program named Mathcad that does math calculations and graphing which makes it easy for me.

For a given telescope what I need is the average counts from a raw frame for each filter along with the camera Gain and Offset settings and the exposure time. This allows me to determine the Skyglow for each filter. Then I can produce the curves for your filters when used with that telescope and the same Gain.

Actually, if you change the camera to another telescope I can probably determine the curves for the new telescope as long as I know the F-number for each telescope.

If this isn't clear let me know and we can figure it out.

Steve

[AstroBee]

I should mention that these curves are for the ASI1600MM sensor with RGB filters and with my F/4.9 WO Star71 telescope and Bortle 5 skies. The curves would be different for different F/# telescope and/or different light pollution conditions.

Steve, I'm following this with great interest because as you know, I too have the 1600mm.
So my question at this point, understanding (sort of) everything you've mentioned, is HOW do I do this same analysis for my system and sky conditions?

I will be glad to do the analysis for you (it would be tedious for you to do by hand!). I have a program named Mathcad that does math calculations and graphing which makes it easy for me.

For a given telescope what I need is the average counts from a raw frame for each filter along with the camera Gain and Offset settings and the exposure time. This allows me to determine the Skyglow for each filter. Then I can produce the curves for your filters when used with that telescope and the same Gain.

Actually, if you change the camera to another telescope I can probably determine the curves for the new telescope as long as I know the F-number for each telescope.

If this isn't clear let me know and we can figure it out.

Steve

Wow, thank you so much for the offer.
If I understand you correctly you would just need a set of exposures through each filter from both of my regular locations, the first being my bortle 8-9 backyard and another from my primary dark sky location which is Bortle 2?

[STEVE333]

That would be perfect. Wow, Bortle 2! Lucky you!!!

Do you use PixInsight or APT? I ask about those programs because I know each of them is capable of measuring the average (or Mean) counts from a raw image. I'm sure other programs can also measure the mean counts, but, I'm not familiar with other programs.

Just make sure you know the Offset, Gain and Exposure used for each image.

Steve

[AstroBee]

I don't use PI or APT... I use NINA or SharpCap for acquisition and Astro Pixel Processor for processing.

[AstroBee]

And I'm lazy so I only shoot with a gain of 139, Offset 10. That's the only master dark library I have built.

[STEVE333]

I would think that SharpCap can measure the mean counts in a raw image. Maybe one of your SharpCap buddies can help with this?

You can just use raw images you have already taken as long as you know the Exposure time used (since you already know the Gain and Offset).

Steve

[AstroBee]

Maybe we could do a Google Meets meeting one afternoon or evening and I could just drop some files on Google Drive?
Do you need a series of exposures through each filter, say 30sec, 60sec, 90 sec, 120sec?

[STEVE333]

Maybe we could do a Google Meets meeting one afternoon or evening and I could just drop some files on Google Drive?
Do you need a series of exposures through each filter, say 30sec, 60sec, 90 sec, 120sec?

A Google Meets meeting would be fine with me. If you put the files on Google Drive I could download them and then open them up with PI and measure the counts.

I only need one exposure for each filter. The analysis takes care of calculating the performance for different exposure times.

Steve

[STEVE333]

Maybe we could do a Google Meets meeting one afternoon or evening and I could just drop some files on Google Drive?
Do you need a series of exposures through each filter, say 30sec, 60sec, 90 sec, 120sec?

I need one more piece of information for the analysis.
1) Do you use dithering?
2) If yes, on average, how long is the time from when one picture ends until the next picture starts?

Example: Consider the case where it takes 60 seconds after an image is acquired for the system to start taking the next image when dithering is performed (the case for my setup):

a) If you dither on every image your average dither time is 60 sec.

b) If you dither on every other image your average dither time is 60/2 = 30 sec.

c) If you dither on every third image your average dither time is 60/3 = 20 sec.

You get the point.

It turns out that the "dither time" has a significant effect on the results.

Cheers,

Steve

[AstroBee]

I dither every second exposure and it usually takes about 30 seconds.

[STEVE333]

Perfect, then your average dither time will be 15 sec.

I wish my time to recover from dithering was as short as yours!

Steve

[AstroBee]

Perfect, then your average dither time will be 15 sec.

I wish my time to recover from dithering was as short as yours!

Steve

I will confirm again when I record the calibration frames for you.