Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

First LRGB image - 2 questions

  • Please log in to reply
9 replies to this topic

#1 NGC3031

NGC3031

    Lift Off

  • -----
  • topic starter
  • Posts: 5
  • Joined: 29 Oct 2016

Posted 13 January 2018 - 04:04 AM

First time posting here, although I've been a lurker for a long time. This is my current setup:

  • zwo 1600 mm-c with zwo filters
  • zwo 290 mc as guide camerea
  • TS Photoline 130mm f/7 Apo

I've tested it once already, everything appears to work as intended and I've managed to get 10x30sec Luminance of m81 and m82 before clouds rolled in.

I'm going to have a proper clear night soon and there's two things were I'm not quite sure yet.
Ideally I'd like to go as deep as possible on m81 and m82 and catch some of the IFN. Am I thinking correctly that the 0 gain preset in the ASCOM driver is ideal for this? My line of thought here is that with a lower gain it will take longer until stars are overexposed, therefore allowing me to expose longer and deeper.
My intention is to use a 3:1:1:1 ratio (i.e. 3 hours Luminance, 1 hour RGB each). Should I use the same exposure times for the subs of RGB as for L, or should I use different ones?



#2 pyrasanth

pyrasanth

    Viking 1

  • *****
  • Posts: 606
  • Joined: 08 Jan 2016

Posted 13 January 2018 - 05:43 AM

I don't know your camera so I can't help much with the gain but you need to be careful that your sensitivity (ISO equivalent) is not so low as to compromise data being captured. On my camera (QHY168C) DSO settings are gain 4 & offset 30 and this captures lots of fine detail but under my light polluted skies I need many hours of 10 minute subs typically 80-100 on M81 to even capture the galaxy outer regions and as for the IFN I can probably forget it from my location with the OSC.

 

However with the mono cameras you will have better luck.

 

I would expose the L at the same settings as the RGB but get twice as much L as the RGB. I would bin the RGB as colour data contains little spatial information and use a high res L to pop in the details. The advantage of binning the RGB is 4 times as much data concentrated on the same CCD area and thus a far greater signal acquisition in a shorter time with no significant loss in detail as the detail is in the L channel captured at hi-res.

 

I look forward to seeing the result.



#3 jgraham

jgraham

    Hubble

  • *****
  • Posts: 18258
  • Joined: 02 Dec 2004
  • Loc: Miami Valley Astronomical Society

Posted 13 January 2018 - 08:09 AM

Everybody has their own recipe. I use the old guidance of setting my exposure to give a peak in my luminance that is about 25% from the left edge (the black point), and I generally take 1/4 the number of lums as RGBs (64 L, 16 each RGB). You can try the unity gain settings and see how they wirk for you. In my case, this gave exposure times that were way too long, so I settled on a gain of 200 and an offset of 50, though I may sneak the gain down to 150. Given the geometry of my bsckyard and my local light pollution I generally have about 2 hours to work with on any given target, so I have to work  within that constraint. Still, I have been getting results that I am very happy with.

 

Welcome to Cloudy Nights!



#4 terry59

terry59

    Cosmos

  • *****
  • Posts: 7555
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 13 January 2018 - 08:29 AM

A perspective:

 

The number of subs will drive certain pre-processing setting choices such as the best image integration rejection algorithm. I've found that 15 x 5 minute rgb subs work for me with my CCD camera. Perhaps you can try 20 x 3 minute rgb subs to see if you get a good enough SNR given your imaging conditions.



#5 the Elf

the Elf

    Vostok 1

  • -----
  • Posts: 158
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 14 January 2018 - 08:38 AM

Agree, the linear fit clipping algorithm seems to be the best and  minimum of 15-20 subs is recommended. From that point of view I'd try to take 20 R, G, B. There are discussions whether or not you get better results spending all time on unbinned R,G,B (without L) or by L unbinned and R,G,B binned 2. Chris Woodhouse is writing about it in his book The Astrophotography Manual but without any final conclusions.

The human eye and brain recieved by far the most information from the luminance signal thus I would recommend to spend half your imaging time on L as you suggested. A look into Ruben Kiers 100 best AP targets reveils he is often using differnt sub counts on the colors. From his book, page 86, M81 and M82 in one image: 12L unbinned, 6R, 6G, 6B 8Ha 2x2 of 5 min. M81 only: 8L unbinned, 10R, 10G 10B unbinned of 10min. M82 only: 8Ha 2x2, 8R, 8G, 7B unbinned 5min. Hope that helps. I recommend that book!



#6 NGC3031

NGC3031

    Lift Off

  • -----
  • topic starter
  • Posts: 5
  • Joined: 29 Oct 2016

Posted 15 January 2018 - 11:16 AM

Thanks guys, I ended up using 30 second exposures across all filters and a 3:1:1:1 ratio at unity gain.



#7 bobzeq25

bobzeq25

    Voyager 1

  • *****
  • Posts: 10460
  • Joined: 27 Oct 2014

Posted 15 January 2018 - 11:58 AM

Thanks guys, I ended up using 30 second exposures across all filters and a 3:1:1:1 ratio at unity gain.

That can work.  Generally you want to use longer exposures for RGB than L.  Look at the histograms, the peak should be roughly in the same place.  Roughly is good enough, precisely does not add value.



#8 the Elf

the Elf

    Vostok 1

  • -----
  • Posts: 158
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 20 January 2018 - 08:31 AM

Hmmmm, in my humble opinion it depends on the binning. It is quite a difference between unbinned, 2x2 or even 3x3 data. Furthermore it depends on whether or not you want to blend red or Halpha data with L. Assuming you use no binning for all filters the signal of the colors is lower as roughly 2/3 of the spectrum is cut off. So I would expect a shorter exposure time for L. Please correct me if I'm wrong!

 

Edit: ah, it is a CMOS. Sorry I didn't check what sensor it is. There is no binning vor CMOS of course. Sorry for the confusion!


Edited by the Elf, 20 January 2018 - 01:59 PM.


#9 Jon Rista

Jon Rista

    Hubble

  • *****
  • Posts: 19165
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 20 January 2018 - 12:30 PM

First time posting here, although I've been a lurker for a long time. This is my current setup:

  • zwo 1600 mm-c with zwo filters
  • zwo 290 mc as guide camerea
  • TS Photoline 130mm f/7 Apo

I've tested it once already, everything appears to work as intended and I've managed to get 10x30sec Luminance of m81 and m82 before clouds rolled in.

I'm going to have a proper clear night soon and there's two things were I'm not quite sure yet.
Ideally I'd like to go as deep as possible on m81 and m82 and catch some of the IFN. Am I thinking correctly that the 0 gain preset in the ASCOM driver is ideal for this? My line of thought here is that with a lower gain it will take longer until stars are overexposed, therefore allowing me to expose longer and deeper.
My intention is to use a 3:1:1:1 ratio (i.e. 3 hours Luminance, 1 hour RGB each). Should I use the same exposure times for the subs of RGB as for L, or should I use different ones?

Keep in mind that usage patterns for CMOS aren't the same as for CCD. Most of the recommendations so far in this thread are fine for CCD, but not optimal for most CMOS cameras. Binning simply doesn't apply with CMOS, there is no good reason to use it (at least, not driver or hardware binning.) Stacking smaller numbers, such as 15-20 subs, is not really optimal in any case, CCD or CMOS...it is just the bare bones minimum you can usually get away with and have the outlier rejection algorithms function at minimal capability. It is always recommended for ideal statistical rejection that you stack more subs...the more subs the better in general.

 

For this camera at f/7, you could probably be using longer than 30 second subs, unless you explicitly want the benefits of short exposures. Especially at gain 0. If you want to use short subs for the resolution benefit, then I would recommend using a higher gain...either 76 (half unity), or 139 (unity). For L, you might want to use 76, and for RGB unity, and use the same exposures for each. You could do 30 seconds at a higher gain and stack plenty of them (with this camera, you definitely want more than 20 subs at gains lower than unity to average out the quantization error...and if you are using short exposures, then getting lots of subs should be very easy...I recommend stacking 80 or more to get the cleanest results. And for an hour of RGB each, at 30 seconds, you need 120 subs, which would be great IMO! For L, you would end up with 360 subs.)

 

If you are averse to stacking lots of short exposures, then you might want to consider using Gain 0 for L and Gain 76 for RGB. You could 2-3 minute exposures for all channels. You would then be able to stack only 20-30 subs per RGB channel, and 60-90 subs for L. As long as your L subs were very well exposed, thoroughly burying the read noise in background sky shot noise, then that should be fine. I would try to avoid stacking fewer L subs than that, though, as it is not ideal with Gain 0 on this camera. At Gain 76, 20-30 subs is a bit low for RGB, but you can more easily apply extensive NR to the RGB channels without having a large impact on the final results when combined with L in the end.



#10 baron555

baron555

    Vanguard

  • *****
  • Posts: 2115
  • Joined: 27 Sep 2010
  • Loc: Lockport, IL

Posted 20 January 2018 - 01:27 PM

For the 1600, use gain = 76 and offset = 15




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.







Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics