M31 from 2017/2018/2019 data with Ha

[ollypenrice]

On 25/07/2020 at 19:22, tooth_dr said:

I've been reading exposure times, as I'm just getting started on LRGB imaging, and thought RGB should be the same (or 2x longer ) than luminance, so I wondering what was the rationale for shorter RGB subs.  l'd have thought that it you can get 10 mins on luminance then you would get 10+ minutes on RGB, and wouldnt be limited to 10mins on H-alpha.

I don't think there can be a singe rule for the weighting of luminance against RGB. It's very target specific. Firstly, are you going after anything very faint, too faint to have any real hope of finding colour in this faint stuff? If so, you can't have too much luminance when it comes to finding this faint stuff. However, if you don't have matching amounts of RGB you'll need to be careful not to burn out all the colour with the L. This will need subtle application of L over RGB and careful stretching of the L. It's easier to add x hours of L to x hours of RGB but, if you're chasing tidal tails, IFN etc, it's perfectly possible to add 10x L to xRGB if you suss out how to do it. Steve is saying the same thing when he refers to APOD images.

Binning RGB? I can see why people do it but here's why I don't: I want star colour to go deep into the stellar cores. If you want it all the way in you won't get enough faint colour, but colour going well into the cores can be dragged further in by processing. If you bin your colour (assuming reasonable sampling rates since huge over sampling will negate this point) you will probably burn the cores to white which is precisely what you don't want to do. So... not only do I not bin colour: I also shoot it in 10 minute subs rather than the 15 to 30 I use for luminance. This is effectively the absolute opposite of binning colour. And it brings another bonus. In my RGB layer I have, in effect, a set of 'short' subs for regions over-exposed in the L layer. There is no reason not to use my far less exposed RGB layer as less exposed luminance. In Ps you don't even need to convert it to greyscale. You can apply RGB in blend mode lulminance.

When it comes to filters and channels and layers in astrophotography I have a golden rule. What am I going to do with this channel?   So suppose I'm going after ultra faint Ha to sit the Double Cluster in an Ha background. Obviously I'm going to need shedloads of Ha. Will it burn out the stellar cores? Sure it will. Does that matter? Not a bit, because I'm going to throw the stars from the Ha into the bin!

Olly

Edited July 28, 2020 by ollypenrice

[tooth_dr]

24 minutes ago, ollypenrice said:

I don't think there can be a singe rule for the weighting of luminance against RGB. It's very target specific. Firstly, are you going after anything very faint, too faint to have any real hope of finding colour in this faint stuff? If so, you can't have too much luminance when it comes to finding this faint stuff. However, if you don't have matching 

In this discussion I was referring to sub length, never total exposure.  I was just curious if your colour filters are capturing say 1/3 of the signal of luminance why would you want shorter subs?

Edited July 28, 2020 by tooth_dr

[ollypenrice]

2 minutes ago, tooth_dr said:

In this discussion I was referring to sub length, never total exposure.  I was just curious if your colour filters are capturing say 1/3 of the signal of luminance why would you want shorter subs?

Sorry, I false-clicked during my post and hadn't finished my explanation of my own take on this. I think I've answered your question in the full version but, if not, let me know.

Olly

[tooth_dr]

5 minutes ago, ollypenrice said:

Sorry, I false-clicked during my post and hadn't finished my explanation of my own take on this. I think I've answered your question in the full version but, if not, let me know.

Olly

Yes Olly, many thanks.  I am sorry for hijacking your thread Steve, this has been in my mind for some time.  I have seen integration info on some nice images, and the lum subs were half the length of the rgb subs.  Previous to this my RGB subs had been shorter than my lum, and I was struggling with the colour from my perceived lack of data.  This is where I started to read and think i was getting it the wrong way around.  But it was most likely my processing rather than the data.  Thanks also for the tip on using RGB as lum in cases, fantastic again.

 

Adam

[x6gas]

35 minutes ago, ollypenrice said:

Binning RGB? I can see why people do it but here's why I don't: I want star colour to go deep into the stellar cores. If you want it all the way in you won't get enough faint colour, but colour going well into the cores can be dragged further in by processing. If you bin your colour (assuming reasonable sampling rates since huge over sampling will negate this point) you will probably burn the cores to white which is precisely what you don't want to do. So... not only do I not bin colour: I also shoot it in 10 minute subs rather than the 15 to 30 I use for luminance. This is effectively the absolute opposite of binning colour. And it brings another bonus. In my RGB layer I have, in effect, a set of 'short' subs for regions over-exposed in the L layer. There is no reason not to use my far less exposed RGB layer as less exposed luminance. In Ps you don't even need to convert it to greyscale. You can apply RGB in blend mode lulminance.

I'm really interested in this.  Recently, for broadband targets, I've only been shooting RGB and no luminance.  Making a synthetic luminance from all of the RGB data has helped but I am now convinced of what was probably obvious in the first place: better results from shooting good quality luminance.

I was assuming that I could bin the colour, as I have done with some success in the past, shooting really short subs - 180s or 300s at most (depending on the scope) because you can stretch the data and then do heavy noise reduction or blur it under the luminance layer.

However, if I read Olly's post right, he's suggesting there is some fidelity in the RGB that is useful both in the colour data itself (giving graduation from core to halo, that maybe wouldn't be present in a heavily blurred RGB) but also the tip to provide less exposed luminance is something I hadn't considered...

More tricks than Paul Daniels 🙂

[ollypenrice]

11 hours ago, x6gas said:

I'm really interested in this.  Recently, for broadband targets, I've only been shooting RGB and no luminance.  Making a synthetic luminance from all of the RGB data has helped but I am now convinced of what was probably obvious in the first place: better results from shooting good quality luminance.

I was assuming that I could bin the colour, as I have done with some success in the past, shooting really short subs - 180s or 300s at most (depending on the scope) because you can stretch the data and then do heavy noise reduction or blur it under the luminance layer.

However, if I read Olly's post right, he's suggesting there is some fidelity in the RGB that is useful both in the colour data itself (giving graduation from core to halo, that maybe wouldn't be present in a heavily blurred RGB) but also the tip to provide less exposed luminance is something I hadn't considered...

More tricks than Paul Daniels 🙂

The thing about real luminance is that it is very strong, the L filter passing nearly three times the flux of the colour filters per unit time. This means that the bright signal (other than the stars) can be heavily sharpened and noise-reduced and the faint can be be dragged out above the background. When I compare synthetic L with real, it's always very disappointing. 

My feeling about star colour is that it's soon lost to saturation, at which point it simply becomes white.

Olly

[kirkster501]

6 hours ago, ollypenrice said:

The thing about real luminance is that it is very strong, the L filter passing nearly three times the flux of the colour filters per unit time. This means that the bright signal (other than the stars) can be heavily sharpened and noise-reduced and the faint can be be dragged out above the background. When I compare synthetic L with real, it's always very disappointing. 

My feeling about star colour is that it's soon lost to saturation, at which point it simply becomes white.

Olly

^^^This.  Absolutely.  And I struggle with this with the TEC140 and will start a new thread on it.