30 replies to this topic

[smr]

Hi,

 

I'm trying to go for a tri or bi colour image with my first light data from my 2600MC Pro and Optolong L-eNhance Filter but I'm not sure on the work flow.

 

I use Photoshop only. I have downloaded Annie's Actions. In PS I split the R, G and B channels and then stretch them all a bit and save them as individual images. I then reopen the stacked data as is (rgb) and stretch that. Then using Annie's Actions there is a bi colour or Hubble palette action. It asks for Ha, SII,  OIII data, which out ot the split R, G, B do I assign to each filter?

 

Or if its more doable I would like to create a bi colour image, red and blue or similar, or orangey blue etc.

 

Thanks for any help.

Edited by smr, 02 March 2021 - 05:00 AM.

[imtl]

You cannot separate the Ha and SII data. They are both mixed in the R channel.

[james7ca]

I'd try a bi-color just using the red and blue channels, you can mix both of those to produce green. I often use a 50/50 mix of Ha and OIII to produce a green channel which I then combine equally with the red and blue. But, it will also depend upon your subject, since you want an emission nebula that contains both Ha and OIII.

[Huangdi]

I created a Photoshop tutorial for this specific purpose. It's fairly straight forward and didn't go too deep into the way of achieving it, perhaps it can help you!

https://youtu.be/bEinmXFXEyk

[smr]

Thanks guys, 

 

Yes a bi colour is fine, I just want something different to an all red nebulosity, I like the look of the red and blue look. 

 

I'm imaging the Soul Nebula, but I can't manage to get the nice blue colour in the middle of the nebulosity.

 

This is what I have, not really understanding what I'm doing...

 

Is it right to split the channels first and then apply level adjust and stretching to each of R, G and B ? That's what I did, and then saved each. I then open the master stacked file again, sort of like a Lum, and then use Annie's Actions and the 'Ha-OIII-Bi Colour' Process. It first asks for my Ha image (presumably this is my R image?) and then asks for my OIII, so I merged the G and B images together to create an OIII. 

 

 

[smr]

I created a Photoshop tutorial for this specific purpose. It's fairly straight forward and didn't go too deep into the way of achieving it, perhaps it can help you!

https://youtu.be/bEinmXFXEyk

 

I tried following the tutorial but I got lost a couple of times so I couldn't follow it.

 

When you first equalise the images you save them as Ha and OIII, without saving as layers, that's fine, but the images are equalised when you save them, then it jumps to both images being layers(?) in a different image, with no equalisation set. Its confusing me. Also how did you create that image with the Ha and OIII layers, I mean, how come there is no background layer ?

[smr]

Has anyone got a link to a tutorial which explains how to do this to someone who has never ventured into splitting channels etc. before. I still don't understand how to do it despite spending two days trying to get my head around it. 

 

Or can anyone explain in clear steps how to do it.... I have my stacked image open in Photoshop. What do I do from there to create a bi colour image? 

[F.Meiresonne]

I gave it a go in Startools, but that won't help much here...tends to come out red ok for HA  but OIII as blue cyan...i did not particulary like that part ..actually

[JamesTX]

Filters like the L-Enhance already produce a bi-color image.  Your image is very red because the O3 signal is very weak.  Getting more data to build up more o3 signal should help but the soul nebula isn't a great candidate for this.  There really isn't a whole lot of O3 in there.  Attached shows 8 hours of L-Extreme on soul nebula.  This attachment is unprocessed.. just auto stretched but it should give you an idea of how little O3 is in there.

 

Try shooting Thors helmet as a test.  I know its small for your setup but there is a lot of O3.  That will give you an idea of how much O3 will show up with your current config.

 

Many of the soul images you see with deep rich blues are using the Hubble palette.  Ha is getting mapped to green, and many people remove green which leaves the blue.  You would need to collect S2 data (separately) in addition to the Ha and O3 that your L-Enhance is picking up.

Attached Thumbnails

• 

[Huangdi]

I tried following the tutorial but I got lost a couple of times so I couldn't follow it.

 

When you first equalise the images you save them as Ha and OIII, without saving as layers, that's fine, but the images are equalised when you save them, then it jumps to both images being layers(?) in a different image, with no equalisation set. Its confusing me. Also how did you create that image with the Ha and OIII layers, I mean, how come there is no background layer ?

Hi, sorry if you couldn't follow it, its tough to make the video easy to understand for everyone, let me try to explain it to you directly.

 

The equalization is never used as an actual stretch, its just to see what's there. 

 

Basically what you want to do first is split the RGB image you got after stacking into two different images. One (the red channel) is the H-Alpha, the other (Green channel) is the Oiii. You then want to stretch these images separately. The goal is to have two stretched images that have roughly the same background brightness.

 

Then its time to put them back together. You can map these colors whichever way you want, there is no right or wrong here.

 

You have two images, but an RGB image has three channels. So you need to create a third "synthetic" green image. To do that you can simply take both Ha&Oiii and use them at 50% opacity (one is 100%, the other one, in the layer above, set to 50%).

 

Now you have three images. Say you want:

Red=H-Alpha 

Green=Synthetic

Blue=Oiii

 

You start with the synthetic image. ctrl+a to select the entire canvas, ctrl+c to copy everything. Go into the channel selection of the blue image, click on the green channel and paste in the Synthetic image. Repeat this with H-Alpha and you have a full rgb image. This is likely going to look awful...the easiest way to balance the colors (in my experience) is with the channel mixer and the color balance tool. Once you found a decent balance, you can start adjusting the colors to your liking.

 

If you have bad signal to noise ratio, you can also use the h-alpha as a luminance layer. This isn't really the way it should be done, but it works either way. Just put it on top of the RGB image and set the blend mode to luminance.

 

I hope this helped

[smr]

Filters like the L-Enhance already produce a bi-color image.  Your image is very red because the O3 signal is very weak.  Getting more data to build up more o3 signal should help but the soul nebula isn't a great candidate for this.  There really isn't a whole lot of O3 in there.  Attached shows 8 hours of L-Extreme on soul nebula.  This attachment is unprocessed.. just auto stretched but it should give you an idea of how little O3 is in there.

 

Try shooting Thors helmet as a test.  I know its small for your setup but there is a lot of O3.  That will give you an idea of how much O3 will show up with your current config.

 

Many of the soul images you see with deep rich blues are using the Hubble palette.  Ha is getting mapped to green, and many people remove green which leaves the blue.  You would need to collect S2 data (separately) in addition to the Ha and O3 that your L-Enhance is picking up.

Thanks, although there is this image with 10 hours data, 294MC Pro and L-eNhance..  This is the kind of look I am talking about. I've got 7 hours with my 2600 so far and so I just wanted something like this where not everything is red (ie. the result I get when I stretch the stacked RGB)

 

 

[JamesTX]

Thanks, although there is this image with 10 hours data, 294MC Pro and L-eNhance..  This is the kind of look I am talking about. I've got 7 hours with my 2600 so far and so I just wanted something like this where not everything is red (ie. the result I get when I stretch the stacked RGB)

 

 

Do you have the link to this astrobin page?  This image doesn't take me to the authors page.  I'm interested in seeing the data points.. might yield some clues on how he did it.  You could always reach out to the author of that image on astrobin and ask. 

[JamesTX]

Maybe he used a mask.  I tested it out real quick on mine.  I created a range mask, then used curves to pull back on the red.. this allowed the blue to come out.  Then boosted the saturation a bit.

 

 

Attached Thumbnails

• 

Edited by JamesTX, 02 March 2021 - 04:20 PM.

[Ivo Jager]

One (the red channel) is the H-Alpha, the other (Green channel) is the Oiii. You then want to stretch these images separately. The goal is to have two stretched images that have roughly the same background brightness.

Please be careful here; non-linearly stretching separate channels is not OK from a documentary photographical point of view. It creates artifacts by suggesting emission concentrations that are not really there.

There is a reason why we don't stretch channels separately for terrestrial photography either (unless for 90s era artistic purposes of course). It yields different colors in different areas of the image that have no bearing on reality.

 

Processing a bi-color is straightforward if you separate luminance and chrominance;

 

* Create a synthetic luminance dataset by blending O-III and Ha. Process that in mono to tease out your detail (it will be mostly determined by the Ha signal). Stretch this all you want.

* Create a chrominance dataset mapping Ha to red, and O-III to blue and also green (for a red/cyan bi-color). Manipulate this dataset as you see fit (but please don't non-linearly stretch the individual channels differently). Boosting the blue/cyan is trivial by simply color balancing (a linear process which should always be performed in linear space).

* Create a composite that uses the processed synthetic luminance dataset for your detail (for example the L component in CIELAB space) and the processed chrominance dataset for the coloring (for example the A*B components in CIELAB space).

 

Hope this helps!

Edited by Ivo Jager, 02 March 2021 - 07:41 PM.

[smr]

Please be careful here; non-linearly stretching separate channels is not OK from a documentary photographical point of view. It creates artifacts by suggesting emission concentrations that are not really there.

There is a reason why we don't stretch channels separately for terrestrial photography either (unless for 90s era artistic purposes of course). It yields different colors in different areas of the image that have no bearing on reality.

 

Processing a bi-color is straightforward if you separate luminance and chrominance;

 

* Create a synthetic luminance dataset by blending O-III and Ha. Process that in mono to tease out your detail (it will be mostly determined by the Ha signal). Stretch this all you want.

* Create a chrominance dataset mapping Ha to red, and O-III to blue and also green (for a red/cyan bi-color). Manipulate this dataset as you see fit (but please don't non-linearly stretch the individual channels differently). Boosting the blue/cyan is trivial by simply color balancing (a linear process which should always be performed in linear space).

* Create a composite that uses the processed synthetic luminance dataset for your detail (for example the L component in CIELAB space) and the processed chrominance dataset for the coloring (for example the A*B components in CIELAB space).

 

Hope this helps!

I appreciate the reply Ivo, I've got no idea how to put it all into practice within Photoshop though. What I need is a video tutorial of your steps but there aren't any. 

[smr]

Do you have the link to this astrobin page?  This image doesn't take me to the authors page.  I'm interested in seeing the data points.. might yield some clues on how he did it.  You could always reach out to the author of that image on astrobin and ask. 

Yup, link here: https://www.astrobin...fualnr/?nc=user

 

I'll ask him how he processed it. 

 

Edit - just realised I don't need to, he has posted his entire workflow here:

 

https://photo.m-j-s....color-workflow/

 

If anyone can translate this to a followable Photoshop workflow I'd massively appreciate it.

Edited by smr, 03 March 2021 - 05:53 AM.

[cybermayberry]

Hi,

 

I'm trying to go for a tri or bi colour image with my first light data from my 2600MC Pro and Optolong L-eNhance Filter but I'm not sure on the work flow.

 

I use Photoshop only. I have downloaded Annie's Actions. In PS I split the R, G and B channels and then stretch them all a bit and save them as individual images. I then reopen the stacked data as is (rgb) and stretch that. Then using Annie's Actions there is a bi colour or Hubble palette action. It asks for Ha, SII,  OIII data, which out ot the split R, G, B do I assign to each filter?

 

Or if its more doable I would like to create a bi colour image, red and blue or similar, or orangey blue etc.

 

Thanks for any help.

I would try and use the blend if tool to make a mix of the blue and green channel data, then copy that blend into the Green and Blue color channels of a duplicate image, then do further adjustments in camera raw for color temperature and tint, while masking areas you don't want effected. This will give you a fake HOO effect. You will need to do additional hue and color adjustments to tweak it to what you want.

[TareqPhoto]

Why people are using a color camera and one of those dual/trio/multi bands filters when there is a mono camera option with NB filters? Beside price and time of course. 

[cybermayberry]

Why people are using a color camera and one of those dual/trio/multi bands filters when there is a mono camera option with NB filters? Beside price and time of course. 

I think you just answered your own question

[TareqPhoto]

I think you just answered your own question

Price and time?

[JamesTX]

Price and time?

Price.

 

Camera (which costs more)

Filter wheel (~$300)

LRGB + NB filters

 

Prices start to add up, especially with 36mm or larger filters.  Puts it out of reach for many.

[TareqPhoto]

Price.

 

Camera (which costs more)

Filter wheel (~$300)

LRGB + NB filters

 

Prices start to add up, especially with 36mm or larger filters.  Puts it out of reach for many.

Ok, fair enough 

[smr]

Ok, fair enough 

Not only that, but my aim is to buy a fast scope at some point, an OSC and f/2.8 or 3.3. 

 

I also image in Bortle 5 from home with my Holiday Home at a Bortle 3/4 location. 

 

If i was in Bortle 8 I would have bought a Mono Camera.

 

Just so happens to be that the best Horsehead and Flame Nebula image I have seen, for one instance, was taken with the 2600MC Pro. That's compared to every other single image I have seen, including plenty of Mono images.

[TareqPhoto]

Not only that, but my aim is to buy a fast scope at some point, an OSC and f/2.8 or 3.3. 

 

I also image in Bortle 5 from home with my Holiday Home at a Bortle 3/4 location. 

 

If i was in Bortle 8 I would have bought a Mono Camera.

 

Just so happens to be that the best Horsehead and Flame Nebula image I have seen, for one instance, was taken with the 2600MC Pro. That's compared to every other single image I have seen, including plenty of Mono images.

And that is why i am planning to buy OSC regardless i am in Bortle 8/9, it might help at some points for some targets, not all.

[dcm_guitar]

Yup, link here: https://www.astrobin...fualnr/?nc=user

 

I'll ask him how he processed it. 

 

Edit - just realised I don't need to, he has posted his entire workflow here:

 

https://photo.m-j-s....color-workflow/

 

If anyone can translate this to a followable Photoshop workflow I'd massively appreciate it.

I'm NOT a photoshop expert, nor am I a Pixinsight expert; not even close!!!  However, I've played around enough in PS (I keep my subscription) to expect that while the above linked process might be possible in PS, if you want to process your OSC images like that you're going to need Pixinsight.

 

For your image processing question, my recommendation is to simply buy Pixinsight and start learning.