Tri- and quad-colour imaging is the preferred means to deep-sky astrophotography. However, one implicit complication is the colour balance which is to be applied when the constituent channels are combined for the creation of the final (L)RGB composite image. The recommended procedure is to use a G2V star for its imaging through each of the filters of interest followed by the analysis of the resultant ADU's in order to derive the associated weighting scheme which will lead to images which are perfectly colour balanced, thus bypassing the differential sensitivities of the constituent channels of the particular CCD camera and filters being used as well as the transmission properties of the telescope (and any Barlows or telecompressors forming the optical train). If more than one telescope is available for imaging, it is adviseable to perform the analysis described below for each telescope (with and without Barlows and telecompressors) so that differential transmisions can also be calibrated.

The above scenario is something which has been addressed by Don Goldman and whose RGB calculator is widely used for the automatic extraction of these weights. However, not only is the actual exercise of tremendous intellectual interest but it also enables the interested amateur to derive these weights for (new) cameras which have yet to be included in Goldman's Calculator (and not to mention that there is also the cost savings in generating these empirical weights on their own).

The example below is based on my SBIG ST-2000XM CCD camera (1600x1200 pixel array, 7.4 micron pixels), SBIG CFW8-A colour filter wheel with LRGB Custom Scientific 1.25" dichroic filters, an AP160 air-spaced triplet refractor and the AIP4Win (V1.4.25) software. At the risk of stating the obvious, the sequence of steps described below can be extended to any monochromatic CCD camera and telescope combination for the derivation and assessment of these critical (L)RGB colour combine weights.

Although a variety of G2V stars exist for such an exercise, perhaps the most popular amongst amateurs for RGB calibration is 16-Cygni (SAO 31898), a 5.93 magnitude star (RA: 19h 41m 48.9534s, Dec: +50° 31' 30.219") near the right wing of the "Swan" which lies approximately 70.5 light-yrs away and is close to the binary star theta-Cygni. This particular star is very similar to our sun in both colour (white) and luminosity (1.61x that of our Sun).

Step 1: Maximum ADU

The first step involves the identification of the exposure which will allow the maximum ADU through any of the filters in use. My ST-2000XM has a maximum possible ADU of 65,536 and an exposure of 3.000 sec yields ADU's of approximately 35,000 (Red), 60,000 (Green) and 45,000 (Blue). To this end, all light and dark frames will be based on exposures of 3.000 seconds.

Step 2: RGB Exposures

For each RGB filter, we will take 16 light and 16 associated dark frames (3.000 seconds for each of these 32 frames). Since photometry is involved (!), we will also require associated flat and flat-dark frames so as to account for potential differential sensitivities amongst the pixels of the CCD array. For the former, we must experiment to arrive at an exposure which will yield a maximum flat ADU lying between 1/3rd and 2/3rds of the maximum possible ADU of 65,536 and which invariably requires an exposure of approximately 20 seconds at dusk. To this end, the associated flat-dark exposures will also be of 20-seconds duration.

Step 3: Calibration and Stacking

Within AIP4Win, we now proceed to calibrate and average-combine (or median-combine if necessary) our various images (light, dark, flat and flat-dark) of our G2V reference star for EACH filter so as to have master frames available for the following steps.

Step 4: Calibration and Stacking

Within AIP4Win, we now proceed to the "Color" tab, select "Color Calculator" followed by "RGB" for the filter set and, finally, proceed with "G2V Star Photometry". A pop-up window will appear requesting the maximum and minimum photometer radii and which require judicious values so as to adequately cover our G2V star. Once the G2V star has been identified for/from each of the filters (click on the G2V star in the master Red frame, click the Red button in the "G2V Star Photometry" pop-up window and repeat in a similar fashion for Green and Blue), we return to the "Color Calculator" pop-up window where we specify the altitude of the G2V star when the light frames were taken (for atmospheric extinction purposes) and proceed with "Calculate Weights". For my particular image train (AP160 with ST-2000XM and CFW-8A), my newly deduced weights for RGB imaging are 0.653:1.000:0.782 when using image processing software which implements these weights as divisors (ex. AIP4Win) and 1.531:1.000:1.279 for image processing software which implements these weights as multipliers (ex. Maxim/DL and CCDStack). In either instance, these weights should yield perfectly balanced tri-color and quad-color composite images.

Step 5: Save and (Re)Use

Once the weights have been calculated, they should be saved ("Save Weights" in the "Color Calculator" pop-up window) so as to permit for their use in all future sessions involving the same optical train. More specifically, we proceed to the "Color" tab on the primary window and select "Process RGB"; each of the master frames for RGB (and L if available) are specified and the colour weights are specified manually or preloaded ("Reload Saved Weights"). It is important to record the altitude when light frames are taken since it will have to be specified as well in order for atmospheric extinction properties to be computed and which ultimately impact the final (L)RGB composite's construction.

Update (Oct 08/2009): The above analysis was performed for my AP160/ST-10XME with SBIG (Custom Scientific) standard LRGB filters and the corresponding fluxes for the master Red, Green and Blue files using five-second exposures (16-Cygni) were 775826, 900086 and 575129 ADU's respectively, thus yielding RGB colour balance ratios of 0.861:1.000:0.640 for AIP4Win and 1.160:1.000:1.563 for Maxim/DL and CCDStack. The enhanced red sensitivity for the ST-10XME is immediately noticeable in these ratios. Examination of the derived RGB ratios for the ST-2000XM above (step 4) similarly reflects the enhanced blue sensitivity which characterizes that particular camera.

Update (Mar 21/2010): With the acquisition of the Baader CCD LRGB imaging filters, a G2V calibration was carried out once again using my AP160/ST-10XME. Although these filters are intended to be used with interline chips (ex. ST-2000XM and STL-11000XM), they are also welcome for the "E" series such as the ST-10XME owing to the fact they are very good at suppressing halos and reflections from bright stars. Using once again 16-Cygni as the reference G2V star for calibration purposes, the corresponding fluxes in the master Red, Green and Blue files using nine-second exposures were 1830578, 1280900 and 914890 ADU's respectively, thus yielding RGB colour balance ratios of 1.000:0.700:0.500 for AIP4Win and 1.000:1.429:2.000 for Maxim/DL and CCDStack.

Note: The above text is being submitted for publication.