combining L,R, G,B monochrome images

[jmfmem]

Hello, Forgive me if this has been explained elsewhere  but I have searched and cannot find.  I am getting into astrophotography and trying to use Affinity Photo for some of the processing.  I am trying to find a clear, reliable way to combine the monochrome images taken with different filters as many people do typically with photoshop.

I have the 4 different monochrome images converted to 16 bit tiffs ( L, R, G, B ) and looking for best way to combine them into a color image. 

 

Thanks for any tips and pointers.

 

John Freeman

Memphis, TN

[Leigh]

Welcome to the forums jmfmem :)

 

Have you tried using Image Stacks from File > New Stack? It contains several Stack Modes that may give you the required effect. Feel free to attach your LRGB images to this thread and I will see what's possible :)

[jmfmem]

Leigh, Thank you so much for replying.  I have not tried that but will.  I have attached L, R, G, B tiffs of a piece of china on the wall in my house (I am testing my equipment before going after the stars).  The levels are not adjusted and especially the blue is very dark as I did not adjust the exposure time.  These are done with a monochrome chip with the different L,R,G, B filters.  I have successfully combined the different pictures in an astronomy program called Nebulosity and the colors look appropriate.  My understanding of the normal workflow for combining these images is to adjust levels and then combine them by assigning them to different channels.  If using the masks how do you add the appropriate color to the mask before combining them?  Also, I am a little confused as to the luminance channel.  Is this interchangeable with the alpha channel (I am revealing my total ignorance here).  Does it need to be colored before combination or is the black and white ok?  Thanks so much.  Hopefully I will attach the files properly.  They have little diagonal lines in them because I had to convert them from "FITS" files (the camera makes fits files) into Tiffs and the program I used is a trial program and adds those lines until you pay for it.

Thanks

John Freeman

Memphis,TN

 

Ugh, I just tried to upload files and it said I am not permitted to upload this kind of file.  They were not too large so I am not sure how to accomplish the upload.

[MEB]

Hi jmfmem,

Welcome to Affinity Forums :)

You can zip them to work around the format limitations in the forum. The max file size allowed is 20 MB.

Alternatively you can get a free Dropbox account (2GB) and share a link to the files.

[jmfmem]

Ok I have zipped up the l, r, g, b files of a piece of china on the wall.  The levels of each may need to be adjusted before combining but they should work for people to experiment with a workflow. Thanks everyone for trying to help.

John Freeman

Memphis,TN

l,r,g,b.zip

[TommyL]

I don't know if this thread is still alive, if so did anyone get a result?  I tried every way i can think of but couldn't get it to work in AP, which probably means I'm missing something obvious.

[crabtrem]

I'm lost too.  I find the concept interesting as it is confusing.  But have no answers or good results for you.  I do wish I had an original color composite of your scene, so I knew what I was working towards.  But I doubt I was even close enough to need it.  Good luck.

[crabtrem]

I doubt I'm even going on the right track.  But I was happy to achieve what appears to be some color information at all.

 

I have no idea what it is supposed to look like.  

[jmfmem]

Thank you for the tries and replies.  I have uploaded a color image of the scene combined in a program called Nebulosity(it still needs some tweaking).   In Photoshop and an astrophotography program that I have used (Nebulosity), you bring in each monochrome image made with the respective filter (L,R,G,B) into its proper channel (the channels are in a color management module).  Once they are combined the image looks fairly normal (like an image made with a normal color camera).  You can adjust the levels to get a more "correct" image.  From all the various functions in Affinity Photo, I know the power to do this is in the program, I just can figure out the proper sequence.  Your image with its overall red cast over the background may be on the right track,  but I think  you need to  back down the heavy red and see if combination colors start coming out.  The photo is taking through a telescope down a hall into a dim room so I am trying to use the low light ability of the camera.  I am just hoping I can find a good way to do this in Affinity Photo.  Thanks all for their efforts and once someone feels they are getting close (as crabtree may be) please post the work flow.

 

Thanks,

John Freeman

combinedneb.tif.zip

[anon1]

-

 

[jmfmem]

Thats it, You have it essentially correct.  AWESOME JOB !! ...  Of course there is probably some tweaking to do but it looks like you have combined the color channels successfully. Please let us know how you did it?  Those jagged diagonal lines are there because I converted the images from FITS (an astronomy specific file format) to TIFFS in a demo program that places those line until you buy the product.  This is great!

Thanks,

John Freeman

[anon1]

Ok great, channels and color space conversion are still the two weakest points in my knowledge so please bare with me.

When I opened your images they were opened in a greyscale color space and for whatever reason neither assign nor convert to color space let me choose something like sRGB. So I had to create a new document using sRGB and your files pixel dimensions. I then pasted the 4 files in (of course you could use Adobe or ProPhoto or whatever as well)

If anyone can tell me why I can´t assign/ convert to a RGB profile, I´d be very happy :)

 

Set the blend mode of the 3 color channels to lighten cause thats how additive color models like RGB are calculated. Lightness can easily be added by changing the luma layer to luminosity (I might have misused these two words here, but I´m sure you know what I mean).

 

You now just have to delete the R + G parts of the B layer and so on for the other two.

Therefore you only turn on the pencil and eye symbol of the respective layer in the channels panel (View > Studio > Channel) (in this case composite B ) and then your have to clear "B green" [green composite of the layer that reflects the blue information, your "blue" file] and clear "B red"

(right click and select clear, it´s the lower list, not the upper one where you toggle the visibility)

 

Then I set a marker from the Info Panel (Studio Info > drag the cross sign onto the picture) into a white area and adjusted the curves to make it 255 255 255 and I did the same with shadows (0 0 0) and repeated and tweaked a bit. You can use a threshold adjustment to get a more precise idea on where the shadows/ highlights really are.

 

You could also use nested HSL adjustments on your R,G,B layers to correct e.g for the underexposure of the blue layer but I didn´t get great results in first place so I decided to use the curves approach.

 

Hope that helps, I always try to keep it brief 

[kirkt]

Unfortunately, if you want to do this the "correct" way, you need to do it in 32bit mode, using linear data.  You add the linear images with color filters applied to each of the three grayscale color images and adjust the exposure of each of the three grayscale color images to white balance the image color.  Once you get the color correct, you can apply that to the luma channel and get your composite.  Affinity Photo does not operate in 32bit as far as I can tell - it will open a 32bit image in 16bit, which clips all of the information that is not within the 0 to 1 range.  If you try to do this in 16bit mode, you may have some success, but you will quickly clip your image.

 

Attached is the composite made in Photoshop, 32bit mode for the math, converted to 16bit via Exposure and Gamma tone mapping and exported as an 8bit JPEG for posting here.

 

 

kirk

[kirkt]

Ok great, channels and color space conversion are still the two weakest points in my knowledge so please bare with me.

When I opened your images they were opened in a greyscale color space and for whatever reason neither assign nor convert to color space let me choose something like sRGB. So I had to create a new document using sRGB and your files pixel dimensions. I then pasted the 4 files in (of course you could use Adobe or ProPhoto or whatever as well)

If anyone can tell me why I can´t assign/ convert to a RGB profile, I´d be very happy :)

 

You probably have to change the document mode to Color 16bit and then you can assign a color profile.

 

kirk

[kirkt]

EDIT - I take my previous statement back.  You can kludge Affinity Photo to do the combination.

EDIT 2 - I reinvented the wheel so succinctly described in post #12 by Mbd.  I'm totally spacing out here.  Anyway, Mbd is correct, but the images need to be linear for this to work accurately, and in 32bit.

 

 

I will post in a second, but you basically create a new document in 16bit RGB color and bring each of your R, G and B grayscale images into the document in their own layers.  In the channels palette you clear the channels of the two color channels that are not the current grayscale image's channel.  For example, in the grayscale image that represents your RED channel, you clear the GREEN and BLUE channel from the layer.  Etc.  

 

Set the Blend Mode of each layer to Add (or experiment with Screen or Lighten).  To White Balance, you can add an exposure adjustment to each layer and adjust the exposure of each of the three layers to get the color correct.  Because the grayscale images likely are not linear, the exposure adjustment will cause color shifts in shadows and highlights, so you will have to tweak the color in shadows and highlights to get neutrals as well (in using exposure you are scaling gamma-adjusted images, where the exposure relationship no longer is linear).

 

kirk

[anon1]

lighten and add seems to give the same effect for me here

 

so basically as I described it but 16bit is better than 8bit and in V1.5 you´d even use 32bit.

 

I don´t think changing it to 16bit would make a difference because it would just not let me get out of greyscale profiles, just allows to convert/ assign other greyscale profiles. But maybe that´s by design because there is just no color information, but on the other hand you could arguably express greyscale values in a colored color space as well....

[kirkt]

Here is the AP version, with some work to make it pleasing ... :)

 

[kirkt]

lighten and add seems to give the same effect for me here

 

so basically as I described it but 16bit is better than 8bit and in V1.5 you´d even use 32bit.

 

I don´t think changing it to 16bit would make a difference because it would just not let me get out of greyscale profiles, just allows to convert/ assign other greyscale profiles. But maybe that´s by design because there is just no color information, but on the other hand you could arguably express greyscale values in a colored color space as well....

 

 

But, once you open your grayscale image, you have to make it a COLOR image (Document > Color Format RGB (16bit)) THEN you can assign a color profile.  I was not sure you were in color mode.

 

kirk

[TommyL]

Interesting contributions, had fun and learnt some stuff doing this. My AP version:

 

 

    

[John Schnupp]

Thanks for the info.  Was trying to figure out how to do this.  Took me a bit to figure it out.  Here is what I was able to come up with for picture based on my data from the Helix Nebula.  I think I clipped the black point a bit and it does need some noise reduction.  Also alignment between the layers is not perfect...but with only a 10-day trial it's not giving me a lot of time.  Still, it looks like there is potential for Affinity to be used in astrophotography processing.

 

[tww]

On 3/21/2016 at 7:14 PM, anon1 said:

Ok great, channels and color space conversion are still the two weakest points in my knowledge so please bare with me.

When I opened your images they were opened in a greyscale color space and for whatever reason neither assign nor convert to color space let me choose something like sRGB. So I had to create a new document using sRGB and your files pixel dimensions. I then pasted the 4 files in (of course you could use Adobe or ProPhoto or whatever as well)

If anyone can tell me why I can´t assign/ convert to a RGB profile, I´d be very happy :)

 

Set the blend mode of the 3 color channels to lighten cause thats how additive color models like RGB are calculated. Lightness can easily be added by changing the luma layer to luminosity (I might have misused these two words here, but I´m sure you know what I mean).

 

You now just have to delete the R + G parts of the B layer and so on for the other two.

Therefore you only turn on the pencil and eye symbol of the respective layer in the channels panel (View > Studio > Channel) (in this case composite B ) and then your have to clear "B green" [green composite of the layer that reflects the blue information, your "blue" file] and clear "B red"

(right click and select clear, it´s the lower list, not the upper one where you toggle the visibility)

 

Then I set a marker from the Info Panel (Studio Info > drag the cross sign onto the picture) into a white area and adjusted the curves to make it 255 255 255 and I did the same with shadows (0 0 0) and repeated and tweaked a bit. You can use a threshold adjustment to get a more precise idea on where the shadows/ highlights really are.

 

You could also use nested HSL adjustments on your R,G,B layers to correct e.g for the underexposure of the blue layer but I didn´t get great results in first place so I decided to use the curves approach.

 

Hope that helps, I always try to keep it brief 

Is there any tutorial for this? for some reason I'm not able to manipulate (swap) individual layers neither within a stack nor in a separately loaded jpeg... 

[tww]

On 1/29/2018 at 7:24 AM, John Schnupp said:

Thanks for the info.  Was trying to figure out how to do this.  Took me a bit to figure it out.  Here is what I was able to come up with for picture based on my data from the Helix Nebula.  I think I clipped the black point a bit and it does need some noise reduction.  Also alignment between the layers is not perfect...but with only a 10-day trial it's not giving me a lot of time.  Still, it looks like there is potential for Affinity to be used in astrophotography processing.

 

Could you share how did you made this image? I'm having difficulties with assembling the photos from individual channels...