Narrowband imaging can we mix different exposure lengths?

[HAlfie]

Hello there,

I'm getting used to the ASI183MM in narrowband!
I would like to know : is there anything, in the image preprocessing step, that prevents from using 2 series of different exposures lengths with the same filter on a same target?
I have of course, darks for both exposures lengths.

For instance the last 3 days, I took 60 x 300s exposures of NGC6960 with Halpha & OIII filters. Then I wanted to be bold 😅  and took the liberty to push up a little bit til 450s exposure's length. I ended up this morning with 17 x 450s exposures both in Halpha & OIII. The SNR  improved over 300s exposures and I think from now on, I may use 450s exposure length for all my next targets.

 I just wanted to be sure, from your experience, what are the consequences of doing that?
Could I stack exposures of 300S and 450s in the same preprocessing step (with respective darks of course) or would I need to preprocess them separatly and merge them in Photoshop?

Though I would like to stack all the exposures in the same processing to have the best improved SNR.

Thanks in advance for your answer and clear skies!

 

 

Edited August 27, 2019 by HAlfie
grammar :-)

[PhotoGav]

You certainly can mix subs of different exposure lengths. Which software are you using for stacking? As you say, just make sure you can assign the correct darks to the different subs and all should be fine.

As for narrowband sub length - as you are discovering, the longer the sub, the better the signal detail captured (in general). The question becomes how long is long enough? There is a trade off between signal / sub length / tracking accuracy / sub lost due to something going wrong / number of subs captured in available time (as the more subs the better for stacking) / max exposure time with your set up, and probably more variables! Personally I tend to use 1800s subs for narrowband, and still feel that I would like to go longer! Good luck and I look forward to seeing the results.

[HAlfie]

Hello Gav and thank you for your detailed answer 🤗!
Ok phew, my "bold" move should not be a problem in the pre processing process, that's a relief.

To answer your question, I plan to use Pixinsight in order to akeep using as little as possible different softwares.

Yes I agree totally with you : we have to find the "sweet spot" between a lot of variables! My system can track with almost 100% success sub of 10min 😀! I worked a lot to improve my AZ EQ6 and have tried all combinations possible and the DEC worm with  the DEC tooth Gear gave me the best resultats in PE, along with the aeroshell 33 grease used in aviation.

I have been going through the long topic on the ASI183MM on the Cloudynight forum, with a special attention to Jon Rista's posts. The conclusion was that the latest CMOS sony sensors (used in ASI183MM or 1600MM) don't behave like CCD sensor. They tend to need a bigger amount of image to stack while CCD are more into "the longest your exposure is the better" despite less number of subs stacked. It's a change of paradigm.

In case of the CMOS sensor used in the ASI183MM, the sweet spot seemed to be Gain 54 for LRGB and gain 111 (unity) for narrowband. Below 54 you may face banding issues, and above 160 gain, you may face big ampglow which will not be totally corrected with darks. I remember that in order to squeeze the better from this CMOS sensor, you need to stack at least 60-80 subs per filter and make a lot of dithering like every 3-5 subs.

It questions me though :  what if I stack less subs with longer exposures? When doing HDR processing, is the noise would appears the same way?

I think I may find out  when I will do the pre-processing and processing exactly the same way between : 
17 x 450s (7,5min ) in Halpha = total of 127,5min
25 x 300s (5min) in Halpha = total of 125min on the same target.

Have a nice day and best regards

[PhotoGav]

Yes, variables and more variables! I am a CCD user and don’t have any experience with the newer CMOS technology. There is plenty of advice available out there though. As with all of this, the only real way to find the answers is through real world experimentation. Good luck with it!

What I failed to mention in my earlier post are the the two approaches to multiple sub lengths on the same target. The first is as I believe you have done, ie deep subs with slightly different exposure times - mix in one integration. The second is two sets of subs at different lengths to expose for different areas of a target - process separately and combine in an HDR fashion in Photoshop or similar. 

[HAlfie]

Hello Gav,

and thanks for your answer :)! And yeah you're totazlly right, sometimes we use different exposure times in order to catch slight details such as the heart of M42 for instance :).
In my case I'm more tying to find the sweet spot between exposure time & number of exposures to stack to limit noise.

Here are 3 exposures without pre processing from last several nights : 

300S exposure in Halpha with stretched levels, 800ADU  (I live in a city) : 

 

450s exposure in Halpha with stretched levels, 900 ADU : 

450s exposure in OIII with stretched levels, 900 ADU (after meridian flip) : 

 

Have a nice day and best regards

Edited August 28, 2019 by HAlfie

[carastro]

I do it fairly often if I decide part way through the imaging run that I need longer exposure than what I have been doing.  For instance my latest Flying Dragons image.

I am not sure if some clever stacking software will reject different length exposures, but depending on what you use, you might need to stack them separately and then stack the stacks as tiffs afterwards.  Hopefully your stacking software will play ball.

Carole 

[HAlfie]

Hello Carole,

and thanks for your answer! Yes indeed in the "worst" case I'll do like you said, stack the stacks afterwards 😀.

After stacking the same total duration in 450s exposures and 300s exposure at same unity gain:
-->Noise seems to be the same (I thought that stacking more exposures would reduce it)
-->SNR seemed improved 

Conclusion: I'll stick with 450s exposure and maybe try 600s . The limit is the ampglow which is more important when we increase gain and/or exposure time. I failed to totally remove the ampglow with my darks in both case. Maybe my "old dark" from 1 year ago has changed since.

17x450s (127min) stacked in Halpha gain 111 :

 

25x300s (127min) stacked in Halpha gain 111:

 

Have a nice day and clears skies

Edited August 29, 2019 by HAlfie

[kirkster501]

Yes you can.  You always make sure that you make a "set" of the darks, flats, lights and bias for that session  You calibrate your lights as a set.  As long as other sets of data are all registered with a chosen master, you can mix data of different exposures from ten years ago if you want.

[HAlfie]

Hello and thanks