Combining color-data acquired with a one shot color CCD based on a Sony chip, and luminance data taken with a mono CCD, based on a Kodak chip.

Like in many cases of human discoveries, it was by chance that I found myself dealing with this method of combining data taken with different CCDs which obviously can serve quite well for the purpose of creating decent deepsky-images.
Chance sometimes puts you in a position where you are forced to break with common methods or what is believed to be reasonable by consent and accepted as a rule.

It was the weather in this particular case that on one hand allowed to take a deep and decent luminance dataset, on the other hand the forecast for the following nights and weeks was most unfortunate. For a deep R,G,B dataset there was not enough time, so I decided to use the one shot color CCD to acquire color data.
The object I was out for was a nice small galaxy in Hercules. NGC 6207 is a tiny little guy, that appears as a 3x1 arc minute small object on the night sky is located very close to M13, and therefore it is highly under-recognized, as M13 draws the entire attention of most observers.

(Click image to see details)

I took the luminance data with a monochromatic CCD to take full advantage of a night with nice and decent seeing and also to have a much more sensitive CCD to go really deep. Starlight Xpress SXVF H16 was used for this purpose attached to my 9” TMB folded apochromatic refractor.
The galaxy was just about to leave the night skies and if I wanted to finish this object in this season, I had to break rules. Otherwise I would have had to wait for the next year… not really an appealing alternative.
So I utilized the color data for this L-OSC image that was taken with the one shot color CCD (also at f/9), in a different night, when I imaged M13 in a wide-field shot with the same refractor.

(Click image to see details)

Using an OSC brings an advantage, as you can create an additional luminance layer, a so-called pseudo-luminance, which is extracted from the entire dataset. This can nicely be obtained by means of the channel mixer in Photoshop. By analysing each channel of the OSC-data one can determine which channel-signal is the very best and as a consequence of this analysis mix the channels that look most promising (mostly red and green, and only a little bit of blue perhaps, maybe after smoothing this channel a little bit). This additional luminance information is quite helpful and you can combine this with the luminance data taken with the mono CCD. The more signal you have in the end the more aggressive the processing can be. (An alternative methode to create an additional luminance-image you can use the entire OSC-dataset and simply debayer this.)

The final result looked most appealing to me and so I decided to use this technique for further shake-down.
There is a nice and lesser known planetary nebula in Taurus. NGC 1514, aka as the crystal-ball nebula, that is not so frequently imaged as more popular nebulae like M27 or M57. Nevertheless 1514 is most beautiful and worth every effort.

Luminance data was taken with my 9” TMB folded apochromatic refractor, using a 2x Power-mate barlow lens to boost focal length up to 4100mm.
OSC would never allow imaging objects at a focal ration of f/18, since the Bayer matrix swallows too much light. A much more sensitive monochromatic CCD with an appropriate S/N ratio is essential for such purposes.
The image scale of my system was 0.4 arc seconds per pixel, which is a quite extreme oversampling. This is very good and most welcome when it comes to process this luminance data aggressively with deconvolution and wavelets-filters.

Again I winded up in the typical weather situation that prevented any further imaging in the upcoming weeks. A friend of mine who lives in Hungary and works there as a mount-engineer, was lucky to acquire 90 minutes of narrow band information in O3.  Andras Dan provided me with this data that I could successfully process as an additional L-channel. The signal was quite weak, as 90 minutes is not perfectly much, so it was insufficient to use it as a color channel. But it was feasible to boost the luminance detail a bit.
The final result is a L-L(O3)-OSC image as I call it.

(Click image to see details)

I tried this again in another PN: NGC 2371/2, located in Gemini and also less popular that it’s neighbour NGC 2392, the Eskimo.
Since this PN is quite faint, I could not image for luminance at f/18, so I operated the system at primary focus, which is 2050mm. H16 for luminance and M25C for color.
Again the final result is quite nice as I think; at least well enough to talk about this method.

(Click image to see details)
 
Next example shows once more that this technique also works considerably well for galaxies: NGC 3718+3729+Hickson 56.
I was actually planning on taking this bunch of galaxies in L-RGB technique, as it was a forgone conclusion these couple are not a perfectly bright target.
In the night I acquired the luminance data conditions of the sky were almost perfect, meaning transparency was so exquisite, I could see stars of 6m5 and beyond with bare eyes. Seeing was between 1.2 and 1.5 arc seconds! For some reason however I encountered some technical issues in the following nights when I was out for gathering the RGB with the monochromatic CCD. So I had to use the OSC CCD to do that job. And I was lucky to have a ‘back-up CCD’ at discretion! Fortunately I winded up with enough data for bringing out the colors sufficiently.

(Click image to see details)

One final examply is M27. In this peculiar case however I combined OSC data as luminance information and general color-data for the stars and H-alpha and O3 for boh, color and luminance on the nebula:

(Click image to see details)

I started to collect data on this well renowned planetary nebula last year in July with M25C, reducing focal length of the 9” TMB Apo by making use of an astrophysics reducer: instead of f/9 I operated the system at f/6.8, to be a bit faster.
This year I imaged this nebula with H16, 2x2 binned though, for several reasons: seeing was not really good and an image scale of 1.4 arc seconds per pixel still gives a reasonable resolution (but this would never be my first choice when I intend to go for crisp detail). Secondly, the H16 is not perfectly sensitive to red, so H-alpha simply had to be binned, if I wanted to finish that object in this life.
So in the case of the dumbell nebula, we are looking at 13.8 hours total exposure time, taken with 2 different kinds of CCDs at different focal ratios.
H-alpha and O3 was blended into the existing OSC data as both, color and luminance. In order to preserve the stars (color and shape) I re-blended the OSC star-mask I created back into the combination-image.

Wrapping this up I would not go so far as to recommend this method as a technique of first choice, but in those cases of unstable weather it might be an option.
For very best accomplishments in DSI LRGB will probably always be the method of first choice, though OSC results can be quite pleasing! Also not everyone has both types of cameras at discretion. I do, as my initial attempts in deepsky-imaging were based on OSC. Certainly this methode could also be interesting for all guys having a DSLR instead of a one shot color CCD.

One advantage of the L-OSC might be the fact, you can come to a nice color-data much faster than RGB. This is based on the fact, that my OSC contains a Sony chip that has an exceptional low noise and therefore an ultimate S/N ratio. RGB is taken with mono-CCDs, which usually (other than SXVF H9) functions with a Kodak chip. And these chips are quite noisy, at least compared to the Sonychip-based CCDs.  So much more frames are necessary to take full advantage of a more sensitive mono-chip and yet not let drown the faint signal of a target in the noise of the chip.
If both cameras used for L-OSC, had the same chip, the time-advantage could be negligible, except for the fact, weather could prevent you to gain the entire L,R,G,B data, once the weather turns out to be frustrating.

Dietmar Hager