Trying to understand LRGB imaging

[mikey2000]

As ever, my mind is wandering ahead to the next upgrade.  I've got a rock solid guided mount now and I'm thinking ahead to an upgrade from my Sony A6000 (mini mirrorless unmodded dslr-sensor camera)

 

Why upgrade?  two reasons.  1) Remote control of the camera is pitiful - I use a hard-wired intervalometer then pull the memory card later on for file transfer.  2) I don't fancy astromodding it (due to reason 1).

 

So, I keep hearing about mono cameras with filter wheels.    I expect I'm looking at a fair bit of cash for a camera, automated wheel +filters.  But I'm not sure of how this sort of thing works.

 

Here's my guess/best understanding so far for how it would work for a 2 hour session:

1) Set up camera and filterwheel for remote operation via some clever bit of software.

2) Shoot 1 hour of L with just a light pollution filter

3) Bin the camera 2x2 and shoot 20 mins each of R, G then B.  (Again with a LP Filter)

4) Calibrate the 4 sets of pics (enlarging the R G and B files to 2x rez, to match the L photos) with their own sets of Flats files (sharing Darks, Biases)

5) Make 4 separate stacks (eg in DSS)

6) Channel-combine all 4 stacks

7) Process from there.

 

Is that more-or-less correct?

 

It seems like a lot of extra processing work.   I tend to get 2 hour imaging sessions.  I wonder if I'd be able to get similar results from a one-shot colour camera?  Maybe, maybe not.

I'd also be interested in adding Ha filters too.  I suppose the mono-camera is the correct way forward here but any advice is always appreciated.

[spillage]

I do not bother binning with the asi1600. The software I use is APT. This is clever, I am not, so it just works for me. You set it up to run X number of exposures using X exposure time using X filter then hit start. You can have different profiles so this only needs to be done every known and then.

colour vs mono is a can of worms...whilst the mono may appear slower it will collect more data per filter than a colour. From what I know the colour is actually 2 green one blue and one red so with my mono and a filter all my pixels will be the colour of the filter with no waste.

You will also be able to use narrowband filters ie ha, the ha filter will allow you to take images in moonlight whilst your colour filters/camera will suffer from light pollution. 

I found one of most important things I did was to erect permanent pier. This allows me to easily go back to target night after night with minimal setup time. I can be up and running in about 10 minutes from going outside so I can just grab that 30mins of clear sky with out much fuss.

File organisation is important but once you have done a few targets you find it fairly easy.

With DSS and mono you just need to make sure you use a reference frame to keep everything aligned.

I am sure others with greater knowledge will be along to give better advise.  

hth

Spill.

[carastro]

It's pretty much correct Mikey, and your choice of going mono with filterwheel will result in some great improvements.  Also the advantage of a mono camera is the ability to be able to do narrowband.

Having said all that, you can bin the RGB or not bin it, but binning is quicker.  I am not sure that only 1 hour luminance is really going to be long enough, but you can find out by experimentation.  I do tend to bin my colour as I can get a coloured image in less time, I also tent to bin Oiii and Sii.  

Quote

6) Channel-combine all 4 stacks

Channel combine RGB and place luminance on top as a luminance layer (use the drop down blending tool).  

I would recommend an electronic filterwheel so that you can automatically collect the different filters without human intervention once you have set it all up.

APT is good, I use Artemis, but it has no extra bells and whistles such as plate solving, dithering (at least on my version), mount control etc, but does me just fine, and some people like to use SGPro.

Good luck.

Carole 

[Star101]

I agree with Spill,

But I do use a OSC camera and have used my Astrodon 3nm Ha filter many times to produce great images with the OSC camera. I just remove the R channel  and work with that. As was said though. Because of the bayer matrix, only 1/4 of the pixels are used for the "R" Channel ( Ha ) so the resolution is reduced. 

If, say, you want to use narrowband, again a OSC camera can be used.....But, IMHO, a Mono camera would probably be a better deal.

[mikey2000]

Ok , I wasn’t too far off the mark (except maybe a bit optimistic with 2 hour sessions)

 

i can set up and go in about 15 mins at the moment - a lot of it is faffing with cables.  That’s not going to improve then!

 

Can I ask two more questions:

LP filters - do people still use them for LRGB?  (Not for Ha as I understand)

How do they attach with a filter wheel?

 

[mikey2000]

And also, how do coma correctors fit into this scheme of things?  I have a baader MPC mk3 with T2 thread

[carastro]

To be honest, I have a LP filter but keep forgetting to use it when at home with my CCD camera.  You don't need it for narrowband.  Mine is a 2" filter which I would screw on the end of the nosepiece if I remembered to use it!!!!

Carole 

[spillage]

I am in the same boat as Carole when it comes to my lp filter.

I too have the same cc. You will end up with camera,filter,cc if you have a separate guide scope. This all screws together as one unit and can be removed from the scope.

You just use spacers to make sure the distance from the chip to the cc is correct. This can be a nightmare and took me several days to get right.

[carastro]

Another idea, if you only get a 2 hour window and want better data, is to take luminance only for 2 hours and do RGB on another evening.  OR, even better.  2 hours luminance one night, then 1 hour luminance and 1 hour RGB on 2nd night.  You main detail will come from the luminance channel.  

These days I have started doing just 150secs binned on RG&B filters.

Carole 

[spillage]

I find that now everything if fixed and I can use platesovle then my approach is more relaxed and less of a panic to cram as much in as possible. I will quite often go out and just collect a few exposures each of lrgb and just keep adding to it.

[Filroden]

2 hours ago, mikey2000 said:

3) Bin the camera 2x2 and shoot 20 mins each of R, G then B.  (Again with a LP Filter)

6) Channel-combine all 4 stacks

The only additional comments I can add are: 

• Only bin colour if your camera does it in hardware as it reduces read noise. If binning is done in software you can equally do it in processing (I think).
• Don't combine L with RGB at the start of processing but once they are both in the non-linear stage because there are processes you will do different to the linear stacks such as deconvolution for L and more aggressive noise reduction for RGB

[MattJenko]

You are pretty much there. If you are confident that your newt setup has no focus difference between filters, then you can always do L,L,R,G,B,L,L,R,G,B... or something along those lines and dither every cycle or something like that. That way you can be sure to have something by the end, as opposed to the soul destroying L, R, G and no Blue for 2 weeks because of the weather

[ollypenrice]

I process the L in Ps and the RGB in Ps separately because the objectives are different. In L you want depth and sharp detail. In RGB you just want low noise and strong colour..

Olly

[tomato]

1 hour ago, MattJenko said:

You are pretty much there. If you are confident that your newt setup has no focus difference between filters, then you can always do L,L,R,G,B,L,L,R,G,B... or something along those lines and dither every cycle or something like that. That way you can be sure to have something by the end, as opposed to the soul destroying L, R, G and no Blue for 2 weeks because of the weather

Because I am impatient to see a final image and just know the UK weather will leave me at least one channel short, I follow this methodology. The downside is I then tend not to persevere with acquiring the lengthy total integration times which IMHO you need if you want produce top notch results. I have recently started Ha imaging which you can do pretty much unhindered around the full moon when LRGB results will be quite badly affected.

I started out with a cooled OSC colour CCD, again because I was keen to see some results quickly, mono with filters is pretty straightforward once you get into the routine, but the initial financial outlay is higher.

Good luck with your chosen path, You  must already know AP is a wonderfully complex and challenging hobby, best not to mention the ?.

[wimvb]

A bit late to the party, but here's my € 0.02.

For ccd, binning rgb makes sense, but less so for mono cmos.

If you have light pollution, with a cmos, leave out L alltogether. There's enough detail in rgb, which is less affected by LP. You can always create a synthetic luminance. With a 5 position filter wheel, you'll then have extra positions for an Ha and an Oiii filter.

If you fear that clouds are going to spoil the fun, get an electronic filter wheel and rotate 5 x R, 5 x G, 5 x B, 5 x R, ... etc. This way you'll at least have a full data set every session.

Also, a cmos will allow you to take short exposures at high gain, making guiding optional.

Processing rgb is simple: calibrate and stack the three channels seperately. Then combine as the first step in postprocessing, and continue as you would with osc/dslr.

Btw, until last summer, I too imaged with a dslr that had no computer connection. Investing in an astro camera turned out one of the best moves so far.

Connecting a cc to a (mono) camera:

CC -- spacers -- filter wheel -- camera

[mikey2000]

Interesting stuff, wimvb!   With my dslr background, I’m sure I’m more used to CMOS and maybe the learning curve will be less.   (Every time I read about ccd, I see stuff like unity gain and full well depth etc etc etc...)

 

I can feel another thread coming on - “which cmos astro camera?” (2018 edition)

[wimvb]

As they say: "Don't sweat the small stuff" LRGB imaging isn't that difficult. As I said, I made the transition last summer and haven't regretted it.

As for which camera, that very much depends on your budget. I chose the ASI174MM-cool. It has a medium size chip, good dynamic range (full well ), and was within my budget. Got the ZWO mini electronic filter wheel and lrgb filters with it. If you have the funds, the somewhat larger ASI1600MM-cool (or QHY equivalent) would be a better choice, simply because of the larger sensor (smaller sensors than the ASI174's are more difficult, because of the limited fov). If you go larger still, you also need larger filters. This can become expensive.

You'll pick up the "language" in no time. Gain is just another word for ISO. (Or is it the other way around?)

As I wrote before, I think that RGB imaging is less sensitive to light pollution. Having a cooled camera also means that you can create a darks library (you need darks with cmos, but you don't need bias), and you can shoot these whenever you have the opportunity. No need to shoot directly after an imaging session to match the temperature.

Processing is also easier, because you don't have the colour mottle of dslr images. And stacking is very much automated. Just tonight I was testing the Batch Preprocessor script in PixInsight. I just loaded up all my subframes (darks, dark flats for 4 filters, flats for 4 filters and lights for 4 filters, all in one go) and let PI do its magic. Out came 4 calibrated and stacked master files for L, R, G, and B.

[TareqPhoto]

Here is a link to my first LRGB image i did of M42, the link including the LRGB stacked files, they are done with 1 sec exposures, it is just for test, i can try to collect the RAW files un-stacked and upload them if you want, someone just aligned those stacked files for me but i don't know where i left them, take a look on them if you want.

https://drive.google.com/drive/folders/1D0yziNirdH_n62VYzlKfK_haYGqMY_6H?usp=sharing

And here is the result of the files after that person who aligned them for me did process it, and i just cleaned it a bit from noise [only applied noise reduction on PS], i am really happy to see how this turned out for just 1 sec exposures and ST80 scope.

 

 

[TareqPhoto]

Really i can't wait to order a new scope [APO] then i can test or image with LRGB all the night clear possible if i can, it is a long time indeed, but seeing the results after proper processing which i hope to learn it soon then you will forget about the long time, you will believe it wasn't wasted after all.

[ollypenrice]

9 hours ago, mikey2000 said:

Interesting stuff, wimvb!   With my dslr background, I’m sure I’m more used to CMOS and maybe the learning curve will be less.   (Every time I read about ccd, I see stuff like unity gain and full well depth etc etc etc...)

 

It's the other way round. You have more parameters to think about with a CMOS because the gain is adjustable, whereas it's factory set with a CCD. Either way I don't think that's a big deal. You really shouldn't give in to the idea that either CCD or LRGB are complicated! (I may well be the least computer-literate person on SGL, as plenty of members will confirm.) You select the filter, start the cooler, focus, frame the target and launch a set of exposures.

If your system is parfocal you can indeed scroll LRGB,LRGB, automatically. Alternatively you can exploit the LRGB system more ingeniously by shooting L and blue when the target is highest. Use the best of the seeing on these filters. (And note that decent filters are parfocal though the optics may not be, in which case an OSC camera can never be optimally focused.)

I advise beginners to shoot equal amounts of LRGB because it makes processing easier. The L will not overpower the RGB and dilute the colour. Once you're into it there are ways, in post processing, of combining deeper L with RGB but this is not compulsory.

Olly

[wimvb]

4 minutes ago, ollypenrice said:

Either way I don't think that's a big deal.

It isn't. Just as most people image at a (more or less) fixed ISO, so in lrgb you'll image at a fixed gain. The only time I change gain is when I don't want to guide (or guiding is acting up during one of those precious clear nights). Cmos gives the option to choose between fewer long exposures, or more shorter exposures.

And because of those few precious nights, I switch between filters quite a lot. I can't afford to wait for best conditions; those may be a year away. Imo, a mono camera with electronic filter wheel is the best way to ensure a full data set after each imaging session, if you can't rely on the weather.

[alexbb]

On 1/27/2018 at 03:21, wimvb said:

If you have light pollution, with a cmos, leave out L alltogether. There's enough detail in rgb, which is less affected by LP. You can always create a synthetic luminance.

I would still recommend shooting L even if you have light pollution. In the end, you still get 4 times the amount of R and B and 2 times the amount of G in the same amount of time. The reason why shooting with R, G and B filters may be less affected is that many filters leave out a gap between G and R where the sodium lamps emit the most. You can buy a LP filter that cuts out some of the most annoying light pollution wavelengths and still passes a relatively equal amount of light for R, G and B. I've one IDAS-LP P2 for this purpose, but now I'm not sure of the benefits of using it compared to plain IR/UV cut. Attached is a mosaic of 2 panels, one at the top, one at the bottom, one taken with the LP filter, the other one is with just the L (IR/UV cut) filter. Both are taken with the same light pollution levels, that would be orange/red in my case. 60x60s each. Which one do you think is with the LP filter and which one with the L filter? Of course, the top panel has some extra short subs used just for the core.

I'm sure that with 20x60s for each RGB filter/panel I would be far from this amount of signal under my skies.

[alexbb]

As for how to use the RGB channels with the L channel, I see 2 ways.

• One is when you want to tease out some details you might lose in the lum because:
  • Your scope is not truly apochromatic so not all the wavelengths focus in the same plane. You can focus for each filter and bring a narrower range of wavelengths closer in the same focus plane therefore you could resolve more details. You should use this method to bring some additional details in the bright areas and for tighter stars. You will need to refocus for each filter for this.
  • You image a bright area which becomes overexposed with all the light recorded, but not overexposed when you filter out with RGB. This doesn't really happen too often as you will most probably overexpose with the RGB filters as well. Or you can go with shorter exposures for RGB than for the L and make sure you don't overexpose.
• The other one is when you can not bring out any additional details or you're not interested to. Then you can even bin the RGB and also you don't need to refocus for each filter, only for L.

[mikey2000]

Lots of hints and tips.  Thanks.  I'm looking forward to getting started.    It's payday tomorrow....