Can I use OSC L Extreme Data with Broadband Data

[Grifflin]

Hi All

I have captured a few nights of data on M31 using an L-Pro filter.  The other night we had intermittent clouds so was doing some experimenting and tinkering with the rig but used M31 as my target and, having discarded a lot of subs, I find I have about an hour of decent dual narrowband data using the L-Extreme filter.

The question I have is, can the L-Extreme data be used in some fashion with the broadband data and if so how might it be used?  For example, is there any value in extracting the Ha data and somehow combining it with the broadband data?  Or something else?

All subs taken with the same camera and telescope.  120s subs for broadband and 180s subs for dual narrow band.

Grateful for your thoughts.

Neil

Mods....apologies.  Should have put this is the Image processing section

Edited January 19, 2023 by Grifflin

[Elp]

Many people do this as the lextreme will capture the bright red hydrogen nebulae within the galaxy as well as any o3, then blend the data together with the narrowband data on top of the RGB.

[ollypenrice]

Certainly, yes. How you do it depends on what processing software you use. In Photoshop I'd place the stretched NB data on top of the braodband as a layer and change the blend mode either to screen or to lighten.

Olly

[Grifflin]

Thank you.

I'm using PixInsight for my processing so I guess the process here is to stack the two sets of data separately and process each until they are in a non-linear state.  Presumably I would then combine them using Pixel Maths and just play with the balance (e.g. 60/40 or 70/30 etc).  Would there be a more clever way of doing the combination given there is only an hour of narrowband data versus 11 hours of broadband?

Neil

[ollypenrice]

2 hours ago, Grifflin said:

Thank you.

I'm using PixInsight for my processing so I guess the process here is to stack the two sets of data separately and process each until they are in a non-linear state.  Presumably I would then combine them using Pixel Maths and just play with the balance (e.g. 60/40 or 70/30 etc).  Would there be a more clever way of doing the combination given there is only an hour of narrowband data versus 11 hours of broadband?

Neil

It would not be a good idea just to weight them in an average of some kind.  What you want is to add the NB signal not found in the broadband to the OSC, but without adding the higher noise from the short NB exposure. It's vital to distinguish between the NB signal and the NB noise because you want all of the new NB signal and none of the noise. Any kind of global averaging will just give you a bit of both and do more harm than good.

This must be possible in PI but I've no idea how to do it. In Photoshop it's a doddle.

Olly

Edited January 21, 2023 by ollypenrice
clarification

[The Lazy Astronomer]

2 hours ago, Grifflin said:

Thank you.

I'm using PixInsight for my processing so I guess the process here is to stack the two sets of data separately and process each until they are in a non-linear state.  Presumably I would then combine them using Pixel Maths and just play with the balance (e.g. 60/40 or 70/30 etc).  Would there be a more clever way of doing the combination given there is only an hour of narrowband data versus 11 hours of broadband?

Neil

To follow up on Olly's post above, pixelmath expressions for those two blend modes are

max(img1,img2)

~(~img1*~img2)

[Grifflin]

Thank you both.

PI is crunching through the stacking at the moment so it will be a while then I'll give those expressions a go.  Presumably I will want to do this on non-linear data?

Neil

[ollypenrice]

1 hour ago, Grifflin said:

Thank you both.

PI is crunching through the stacking at the moment so it will be a while then I'll give those expressions a go.  Presumably I will want to do this on non-linear data?

Neil

In Ps I do not add NB to broadband at the linear stage, I do it with fully processed images.  Since I know what parts of the NB will be going into the final image (that is, the parts where the NB is brighter than the BB) I can stretch the two sets differently. The BB stretch must keep a low noise level in the fainter parts because it will be seen in the final image. The NB image can be stretched in the knowledge that its fainter parts will not be included in the final blend, so I can use a very extreme stretch. As long as the faint stuff in the NB is darker than the same faint stuff in the BB, I don't have to worry because it won't be added. 

In a nutshell, the stretches used in NB can (and in my view should) be very different from those used in the BB.

Olly

[Grifflin]

12 minutes ago, ollypenrice said:

In Ps I do not add NB to broadband at the linear stage, I do it with fully processed images.  Since I know what parts of the NB will be going into the final image (that is, the parts where the NB is brighter than the BB) I can stretch the two sets differently. The BB stretch must keep a low noise level in the fainter parts because it will be seen in the final image. The NB image can be stretched in the knowledge that its fainter parts will not be included in the final blend, so I can use a very extreme stretch. As long as the faint stuff in the NB is darker than the same faint stuff in the BB, I don't have to worry because it won't be added. 

In a nutshell, the stretches used in NB can (and in my view should) be very different from those used in the BB.

Olly

Thanks Olly

Yep, I think I get what you are saying.  I'll give that a go although, to be honest, my processing skills are pretty basic at the moment but this will be good practice.  

Neil