Sub lengths for RGB vs L and Ha

[alcol620]

Hi folks 

 

What's the perceived wisdom on the length (and binning) of RGB subs compared to Luminence and Ha subs, if the L or Ha subs are later to be combined with RGB subs?

 

Thanks for feedback

 

Alec

[johnrt]

It's really going to depend on your sky conditions.

Ha blocks most light pollution out, so the only restriction is the accuracy of your tracking & guiding, aim for 15 minute subs as a minimum if you can, the longer the better.

LRGB is much more sensitive to sky glow, so will need to be more dependant on your skies. I am in a semi rural location and with an IDAS light pollution filter I can manage 10 minutes Lum, I bin the RGB 2x2 so go for 5 minute subs. It's really best practise to not bin Lum or Ha as they are where all the contrast and detail is to be found.

[sharkmelley]

Your sub lengths should be long enough for the background sky fog noise to swamp the read noise.  So L,R,G,B subs can be much shorter than Ha subs (since they will have a brighter and noisier background sky fog).

Mark

[ollypenrice]

Whether or not to bin depends on your pixel scale. If it is already coarse (such as in our dual Tak rig at 3.5 arcsecs per pixel) then binning would be working at 7"PP and that would give brick-like stars. I wouldn't consider it. If using a long FL and/or small pixel camera giving something like 0.6"PP I doubt that there would be any perceptible price to pay. If binning is going to take you below about 3.5"PP I wouldn't do it. myself.

If you are going to use an HaRGB blend then don't bin. The star size mismatch will be problematic and the binned RGB stars won't look good. If it's going to be HaLRGB then that might work better. Personally I wouldn't normally be keen opn any binning with Ha waiting to be added, though. Maybe I'm too cautious.

In LRGB images, particularly if they contain dark starfields (as in galaxy images, for instance) then I don't want to apply L to the field stars so I want nice tight little RGB-only stars, so no binning.

The L-RGB sub length relationship depends on the target as well, I think. If chasing some very faint thing in which colour is a no-hoper (the tidal tail in the Leo Triplet, for instance) then I'll take more and longer L subs than RGB. The shorter RGB will be fine for the stars and galaxies and nothing is going to find any colour in the tail. If you shoot equal amounts of L, R, G and B combining them is easier. Where you have lots more L than colour it can still be done but you need more processing tricks up your sleeve.

Olly

[alcol620]

Thanks Olly, your post provides lots to think about. have just started to use the Atik 383 on my Sky-Watcher Evostar ED80, how does that stack up with your initial comments? My first light with the secondhand Atik was last night taking Ha's (the only narrow band filter I currently have). How would you tackle that target with only LRGB and Ha filters?

Regards

Alec

[ollypenrice]

Thanks Olly, your post provides lots to think about. have just started to use the Atik 383 on my Sky-Watcher Evostar ED80, how does that stack up with your initial comments? My first light with the secondhand Atik was last night taking Ha's (the only narrow band filter I currently have). How would you tackle that target with only LRGB and Ha filters?

Regards

Alec

I check my pixel scales here http://www.12dstring.me.uk/fov.htm but FLO have a calculator as well now. 

With flattener-reducer (which you will want/need for this scope) you'd be at 1.86"PP which is a nice resolution at which to work. It's efficient and yet will hold delicate detail. Binning 2x2 would take you to 3.71 which is getting marginal in my view. It would depend on the target. I wouldn't use it in an HaRGB shoot, for instance, though HaLRGB might be OK. Since I'm not a fan of binning I wouldn't use it myself. I work at this scale with our TEC140/Atik 11000 and never bin. However, I get lots of clear nights...

Olly

[alcol620]

Thanks Olly

Much appreciated

Alec

[Droogie 2001]

I check my pixel scales here http://www.12dstring.me.uk/fov.htm but FLO have a calculator as well now. 

With flattener-reducer (which you will want/need for this scope) you'd be at 1.86"PP which is a nice resolution at which to work. It's efficient and yet will hold delicate detail. Binning 2x2 would take you to 3.71 which is getting marginal in my view. It would depend on the target. I wouldn't use it in an HaRGB shoot, for instance, though HaLRGB might be OK. Since I'm not a fan of binning I wouldn't use it myself. I work at this scale with our TEC140/Atik 11000 and never bin. However, I get lots of clear nights...

Olly

Hi Olly,

Do you comments apply to simply LRGB as well? I am looking at a setup which offers 3.18 Arcsecs Per pixel, this which would then equate to 6.3 Arcsecs per pixel at 2x2 binning if I am correct? (still learning)

I thought 2x2 binning was a recognised method of reducing RGB image capture time with an acceptable loss of resolution. I do not have clear / dark skies but I want to avoid odd shaped stars for sure...

Thanks

[johnrt]

Hi Olly,

Do you comments apply to simply LRGB as well? I am looking at a setup which offers 3.18 Arcsecs Per pixel, this which would then equate to 6.3 Arcsecs per pixel at 2x2 binning if I am correct? (still learning)

I thought 2x2 binning was a recognised method of reducing RGB image capture time with an acceptable loss of resolution. I do not have clear / dark skies but I want to avoid odd shaped stars for sure...

Thanks

If you want nice tight round stars in LRGB is is not best to apply the Lum to the stars, so I would not bin the RGB at that resolution if I were you.

[ollypenrice]

If you want nice tight round stars in LRGB is is not best to apply the Lum to the stars, so I would not bin the RGB at that resolution if I were you.

My thinking as well.

Hi Olly,

Do you comments apply to simply LRGB as well? I am looking at a setup which offers 3.18 Arcsecs Per pixel, this which would then equate to 6.3 Arcsecs per pixel at 2x2 binning if I am correct? (still learning)

I thought 2x2 binning was a recognised method of reducing RGB image capture time with an acceptable loss of resolution. I do not have clear / dark skies but I want to avoid odd shaped stars for sure...

Thanks

Binning colour is indeed an accepted practice and can save time, but the cost is never zero. If you are working at a pixel scale which is reasonably fine once binned (say 2 arcsecs per binned pixel) then I think the cost might be negligible on targets full of nebulosity. On an image with lots of starfield, though, you wouldn't be able to make RGB-only stars which John and I are recommending as good practice.

I would not bin if it took me down to 6.3" per pixel under any circumstances. That is too coarse for me. Besides, at 3.18"PP there is less of a need need to do so.

Olly

[Droogie 2001]

Thanks Olly.

Just needed clarification on you last comment "...at 3.18"PP there is less of a need need to do so". Why is that and does that mean I will have to take equal lengths of RGB and Luminance due to PP size regardless?

[ollypenrice]

Thanks Olly.

Just needed clarification on you last comment "...at 3.18"PP there is less of a need need to do so". Why is that and does that mean I will have to take equal lengths of RGB and Luminance due to PP size regardless?

You're not short of light for each pixel. That's the key point so far as I see it.

It's easier to process an image which has equal quantities of L,R,G and B but if you have a lot more L than colour it can still be done by adding the L in several iterations, boosting the saturation and giving a small blur for noise reduction. You do this several times and then apply the L at full opacity without blur for the last iteration. A beginner to processig might just be best advised to go for equal exposures per sub.

Olly

[Droogie 2001]

Read up a bit more on the Pixel size for the corresponding camera, chip, FL etc. Just when I think I have a solid plan on how to proceed a new snag arrives. That's said at least I am able to factor this in at this early stage.

Thanks as always for your time and invaluable advice.

[alcol620]

The above responses have convinced me to adopt a policy of using only 1x1 binning with the kit I have got. It certainly sounds as if the processing is more straightforward with a 1x1 binning policy. (if straightforward is the right word!)