After acquiring my monochrome ATK16 IC camera I had to learn how to make colour images with this camera, which is amazing enough in its own right: making colour images with a b/w camera!
This technique is called 'LRGB' and is based on using colour filters during imaging and special post processing techniques.
During my learning curve I have made some notes, so that I would not forget what I learned.
No, this is NOT a full fledged LRGB Tutorial, but only describes the main scenario flavoured with some hints and tips.

IMPORTANT: for my DSO Tutorial - with many imaging and processing tips - click .... HERE !!!!


  1. I have a nice set of Astronomik LRGB 2c Filters and an Astronomik 13 nm H-alpha Filter and a Scope Teknix 5-position filter wheel, which I insert between my optics (LX200 or William Optics Zenithstar 80 ED II APO) and my ATIK 16IC camera
    Master deep sky imager Jim Thommes remarked:
    "I can tell you that my Astronomik Ha and IR block filters are parfocal, so I think Astronomik tries to get it right.
    You may have good results with your Astronomik RGBs.
    Also, I would agree that you should not need the IR filter on your Astronomik RGBs".
  2. I capture an AVI with each filter; in order to remember which AVI was captured with what filter I use suffixes _H _L _R _G _B in my AVI file names.
  3. Question: what exposure times to use with what filter?
    Jim Thommes:
    "I recommend going deep on luminance exposure for detail, but go light on R,G,B. If you overexpose RGB's things tend to white and colours can be muted. Another thing to think about is the effect of the luminance channel on the RGB. A luminance value of 255 when luminance blended with an RGB will always yield white. Sharpened stars that go from 255 to black in the edge pixels of the star will be white - star colours will go away. Thus a brighter star likes to be blurred a little to the edges to bring out colour - or at least not overly sharpened."


For fine tuning the exposure times between the filters you need to CALIBRATE the filters.
Jim Thommes said about this issue:
"If you image a G2V star with equal times for R,G,B (probably intentionally defocused) you should get a balanced colour. (i.e. at half max, each colour should have a weight of 128). If there is no balance, you can use Curves in Photoshop (at input level 128) to balance these colours for your test exposure. Record the output setting and now you have a scale factor for calibrated colours for use in Photoshop . Now you can relax an take all your filter exposures at the same time and calibrate them later in Photoshop."
Photoshop Tutorial by Jerry Lodriguss
More about calibration and also a list of G2V stars (sorted in RA).

There is a LOT more to say about this, but Rob Kantelberg has a nice Tutorial about this subject on his website [Tutorials => LRGB imaging. He even supplies an exposure calculator !].


Meanwhile I have accumulated some experience myself in imaging with my ATK16IC camera, so here it is:
Note: because I am already familiar with K3CCDTools I decided to use K3CCDTools as my imaging software

  1. 1. ATK16IC settings in K3CCDTools.
    Peter Katreniak recommends the following:
    1. The color/monochrome/RAW radio buttons only affect the preview of colour cameras
    2. Adaptive deinterlace feature is useful for CCD cameras based on CCD chips with interlacing (like TV broadcasting). In such cameras the CCD is read in two steps and thus the second step exposure is slightly longer - which causes interlacing pattern (especially on shorter exposures). This problem can be observed on DSI monochrome cameras. ATIK 16IC has progressive scanned CCD, so this option should stay unchecked.

  2. 2. Reaching focus
    I found out that my William Optics Zenithstar 80 ED II APO did not have enough back focus, so after some experiments I decided to buy a 2" (5cm) Extender [for imaging I do not recommend to use a diagonal as that might introduce unwelcome reflections etc.]
  3. 3. Focusing and camera orientation.
    Of course select a bright star and do some coarse focusing.
    The readout time of the CCD for full sized images is about 3.5 seconds, which is not nice when focusing.
    So you should select a small size video format: 160x120 pixels: aha, that goes a lot faster!
    Now use the ZOOM function of K3CCDTools to enlarge the small video screen and fine focus.
    I recommend to use a Diffraction Spikes Mask for final focusing.
    Use the the reticles to adjust the East/West orientation of the camera, so that you have North UP.
  4. 4. Live Histogram.
    1. 1. Invoke the Live Histogram and use Auto mode
    2. 2. Aim at your object and expose for say 60 seconds
    3. 3. Look at the lowest value of the Live Histogram: it should be at least 250 - 300; if it is less than increase the exposure time. Oh: local sky glow can surely give Low values of 4500, so be it ... sigh ...
    4. 4. Look at the High values: should not exceed 52000 or so
    5. 5. Now stretch the histogram till you have a nice preview picture; when you are happy: OK ! Else increase the exposure time till you are happy: experiment!

  5. The above screen dump was made while imaging M3; please note the Live Histogram.

    Live Histogram
    At the top you see the L and H values: these are the actually measured values of the just captured image.
    At the bottom are also L and H values: these were manually selected by using the slider, so that I could get some preview.
    When you select "AUTO" the top and bottom values will be (almost) identical, depending on atmospheric conditions.

  6. 5. FINALLY: Time to Start Capturing !
I store the captured frames in an AVI, because Peter 'K3CCDTools' Katreniak has incorporated a special K3CC codec which stores in 16-bit RAW format which is exactly what I need.
I always take great care to include in the filename of the AVI which filter has been used (L, Ha, R, G or B) so that I will not be confused during the postprocessing.


1. Align - Stack - FITS Liberator.
I align and stack the LRGB individual AVIs with K3CCDTools [see "Selecting, Aligning and Stacking" in my DSO Tutorial] and save as individual FIT files: 16 bit signed.
The next step is to bring out the details with the Photoshop Plugin FITS Liberator.
I highly recommend this plugin, which is freely downloadable; I also recommend the FITS Liberator Tutorial by Noel Carboni.
I save each file as a TIFF file with the suffix "fitslib", so that later I know that this is the result of the FITS Liberator processing.

2. Combining the R G B channels.
Don Waid has a tutorial about this Channel topic.
In short: in Photoshop open a new canvas, RGB, 16 bits and of the proper size.
Select: Window - Channels and you see a panel with the combined RGB and also the R G B separately.
Paste your Green image into the Green channel, Red into Red and Blue into Blue.
I usually have to manually align the channels, which is very easy with the arrow keys on your keyboard.
Hint: activate/deactivate the channel you are working on for blink comparison and fine adjustment.
Currently this is rapidly becoming my favourite scenario.
I save the result as: RGB-aligned.psd
When you are happy with the RGB result just add the Luminosity layer in the same fashion.
Hint: use about 50% Opacity of the Luminosity layer for proper positioning, then put Opacity at 100% and set the Blend mode to Luminosity.
I save this result as: LRGB-aligned.psd

3. How to avoid the "peach" or "salmon" hues of emission nebulae.
When you have mastered the above you will want to avoid "peach" or "salmon" hues of emission nebulae for which Robert Gendler has a very nice tutorial to include the h-alpha layer directly into the RGB colour data.
For my own convenience I have made the following extract telling me what to do:

  1. New canvas, mode RGB.
  2. Copy the Ha and the Red layers to this new canvas, with the Red layer on top, blend mode Normal or Lighten.
  3. Play with Opacity and Blend modes till you have normal looking stars and nice nebulosity and contrast.
  4. Flatten this H-alpha-Red blend and save it as the "new red channel".
  5. Combine this "new red channel" with the existing G and B channels to get an RGB composite.
This Ha+Red operation is easiest in Photoshop CS or Photoshop Elements 2.0 with the freeware FITS Liberator where you can stretch the previously created FIT files and the perform the layering as explained by Robert Gendler thus creating a "New Red".
Now open a new canvas, select Layers/Channels and copy the New Red into the Red Channel; after that stretch the Green and Blue FIT in FITS Liberator and copy/paste them into the Green and the Blue Channels respectively.

4. Combining 2 Different Exposures.
Jerry Lodriguss has a fine tutorial about Combining 2 Different Exposures which is especially useful for objects with large differences in brightness, such as M42.


1. Removing halos by using the Inverse Selection method.
My good friend Adrian Jones had imaged the Double Cluster but wound up with blue halos around the brighter stars.
I gave it a try and here is my Photoshop scenario (of course I use Adrian's image with his full consent).
I slightly downsized Adrian's image so that it fits better on my screen ...

The image as captured and post processed by Adrian Jones.
The blue halos are prominently present.

Step 1: Select - Color Range - click on saturated star - slider at 192

Step 2: Select - Modify - 2 pixels (I used 2 pixels because of the size of the halos)
Step 3: Select - Inverse
Step 4: Hue/saturation(Ctrl-U) to de-saturate and darken the blue [one might do the same with red, which I did not do yet, but I invite you try do this yourself]

Here you see the inverted selection and the hue/saturation settings I used to reduce the blue halos.
Please note that only the BACKGROUND is being worked on: the inverse selection takes care of this.
PS: of course this INVERSE selection technique also allows you to STRETCH the background which is nice when you want to bring a nebula or nebulosity more forward.

Step 5: Next I sharpened a bit [Filter|Unsharp Mask], which increased the brilliance.

Step 6: The Result

The blue halos have disappeared.
This blink comparison shows the before and after ...

Adrian Jones commented as follows:
Excellent, Jan!
Your method avoids the matching-back difficulty altogether by working on the halos AND whole sky background together - very clever.
And it even flattens out slight colour unevenness in the sky background at the same time - an extra bonus!
I'm really impressed by the results and will certainly use this method in future for star clusters.

2. Removing halos while leaving a galaxy untouched.
This tutorial is an extension to my "Removing halos by using the Inverse Selection method ".
My good friend Stephane Poirier had imaged NGC7331 but wound up with reddish halos around the brighter stars.
When using my "Removing halos by using the Inverse Selection method " also the Galaxy was affected, which we do not want of course.
The problem was how to add the Galaxy to the already existing selection.
Here is what I did ...

Step 1 and 2: as in : "Removing halos by using the Inverse Selection method ".
Now we have to add the Galaxy to the selection, which I did as follows:
Step 3: Activate the 'Magic Wand' function
Step 4: with the cursor in the image: do a right mouse click: you get a pop-up menu
Step 5: Select 'add to selection' and hold the SHIFT-key down (and KEEP it down !!!)
Step 6: Activate and use the Lasso tool - I used a feather of 20 pixels - (now also press the left mouse key) and encircle the galaxy
Step 7: Now you can release the SHIFT-key: the galaxy has been added to the selection!
Step 8 and 9: similar to Steps 3 - 5 in : "Removing halos by using the Inverse Selection method ", but this time I concentrated on the Red colour.

The Result: Here you see the 'before' and 'after' of this scenario (just move your mouse in and out t see the difference):

3. Enhancing nebulosity by using the Inverse Selection method.
I use a similar technique as in the previous paragraph, but with a twist :o)

  1. Load the image into Photoshop
  2. DO: Select - colour range - click on saturated star - use slider to select all the brighter stars
  3. DO: Select - invert (now you have selected the sky :-)
  4. DO: Select - modify - expand - 1 pixel
  5. Now do your Curves ! The nice thing is that only the sky (and thus the nebulosity) is stretched, because the stars are NOT selected :o)
  6. When you are happy reduce the noise a bit (DO: Filter - Noise - Noise reduce) you can see what you are doing :o)
  7. DONE

4. Hi Pass Masking: bring out dust lanes.
For your (and my own) convenience: here is a quick scenario of Don Waid's Hi Pass Masking Tutorial

  1. Open the image in Photoshop
  2. Duplicate Layer (Name it: Hi-Pass)
  3. Blend mode: Overlay
  4. Filter|Other|Hi-Pass Radius 8.0 pixels (This sharpens the whole image)
  5. Layer|Layer Mask|Hide All (This creates the Layer Mask)
  6. Make sure the your Foreground Colour is White and the Background Colour is Black
  7. Select Brush with proper size; Opacity and Flow both 100%; Mode Normal
  8. Manually sharpen the detail with left mouse click
  9. Flatten Layers
  10. Save image

I followed this Hi Pass Masking method, with the following result (just move your mouse in and out to see the difference):

5. Sylvain Weiller's solution to deal with light pollution.
Sylvain lives in Paris and here is his PDARK Photoshop solution against light pollution.

6. How to remove a Gradient with IRIS.
Here is the IRIS Tutorial of which I followed the "Gradient removal on very wide field images" (mouseover shows it all):

Here is in short the IRIS Gradient Removal Procedure: [works on grayscale images only ]
LOAD image-with-gradient
Wavelet command of Processing menu (all values at 0.0, last value at 1.0, checkmark Auto verif .)

LOAD image-with-gradient
>SUB SKY 100
>SAVE image-with-gradient-removed

7. Get Vibrant colours without increasing contrast.
Here is a link to a tutorial how to Get Vibrant colours without increasing contrast by using Lab colours.
Here you see the 'before' and 'after' of this 'Get Vibrant colours' scenario (just move your mouse in and out te see the difference):

8. From elongated to round stars.
In spite of your best tracking and/or guiding you sometimes wind up with egg-shaped stars, alas ...
Here is my scenario to turn your eggs into nice round stars (but apply with care !!):

  1. Open the image in Photoshop
  2. Window - Layers [Pops Up the Layer Window]
  3. Layer - Duplicate Layer [Background Copy]
  4. On this Duplicate Layer: Set Blend Mode to Darken [use the pulldown menu in the Layer Window]
  5. Filter - Other - Offset and activate "Set to Transparent"
  6. Enlarge the image by "Ctrl +" so that you can see what you are doing
  7. Increase/Decrease the Offset values [best with the arrow keys, max. value 2] to manoeuvre the image in such a way that your stars are as round as possible
  8. Close the Offset window
  9. Now fine tune via: Edit - Fade Offset and play with the slider till your stars are even rounder, then close with OK
  10. Flatten the layers: Layer - Flatten
  11. Save Image
  12. Set Blend Mode back to Normal [use the pulldown menu in the Layer Window]
  13. Because of the above scenario it might necessary to finalise with some Curves/Levels to your image.
  14. Save Image
Here you see the 'before' and 'after' of this 'Elongated Stars Repair' scenario (just move your mouse in and out):

And here is the final image:

9. Bring out the colours of bright stars.
And Don Waid gives live demonstrations and explanations of what he does in Photoshop.
Here is in short steps what Don recommends to do to bring out the colours of bright stars:

  1. Open the image in Photoshop
  2. Duplicate Layer (just for safekeeping and a possible redo)
  3. Select - Colour Range - click on bright star - fuzziness 20
  4. Select - Modify - Expand 4 pixels
  5. Select - Feather 3 pixels (75% of expand value)
  6. Curves: lower intensity of stars: the upper part of the curve only
  7. Saturation: try 50, 60 or more (experiment)
  8. Filter - Other - Minimum 1 pixel
  9. Edit - Fade Minimum - 50% (experiment)
  10. Deselect
  11. Save image


M67 - Open Cluster.
First Light for my ATK16IC and also my first LRGB image.
March 25, 2007 19:31 - 20:32 UT.
I selected M67 because it contains a G2V star: nice to calibrate my filters.
Imaging: William Optics Zenithstar 80 ED II APO, f/6.8 [focal length 545 mm] with 10 cm extender, Astronomik LRGB filters; ATK16IC b/w camera, K3CCDTools.
Exposures: Luminosity 5x210 seconds; RGB 5x100 seconds; all unguided.
M67 - Open Cluster.
March 26, 2007 19:46 - 20:16 UT.
My second LRGB image; once more I selected M67 because it contains a G2V star: nice to calibrate my filters.
Imaging: William Optics Zenithstar 80 ED II APO, f/6.8 [focal length 545 mm] with 10 cm extender,
Astronomik LRGB filters; ATK16IC b/w camera, K3CCDTools.
Exposures: Luminosity 18x30 seconds; RGB 10x30 seconds; autoguided.
M67 - Open Cluster.
March 25, 2007 19:34 - 19:59 and March 26, 2007 19:46 - 20:16 UT.
My first LLRGB image:
I combined the 210 seconds luminosity layer of March 25, 2007 with the 30 seconds LRGB image of March 26, 2007.
I determined the rotation between the images with Astrowave and then applied the exact rotation in Photoshop and layered.
Imaging: William Optics Zenithstar 80 ED II APO, f/6.8 [focal length 545 mm] with 10 cm extender,
Astronomik LRGB filters; ATK16IC b/w camera, K3CCDTools.
Exposures: Luminosity 5x210 seconds plus18x30 seconds; RGB 10x30 seconds.

The following QCUIAG friends helped me getting started in LRGB:
Dave Nash (Dave's LRGB page is here)
Jim Thommes
Peter Lloyd
Rob Kantelberg