Here is another example, from tonight's observing session, of "Backward LRGB" - a term I coined for doing things in the "wrong" direction. But the results speak for themselves. The image here shows 3 panels:

1. Top Left - B&W image = stack of 12 x 150s in 2x2 binning mode

2. Top Right - Color Image = stack of 24 x 150s in 1x1 binning mode

3. Bottom Middle = bLRGB (backward LRGB) image formed from both of the previous images.

What this test shows is that, in fact, using the OSC CCD in 2x2 mode is actually more sensitive than a factor of 2 over the individual RGB pixels. I wouldn't expect it to be 4 times more sensitive because each of the component pixels of the Bayer superpixel has a different quantum efficiency. Using an OSC CCD isn't going to give the same kind of benefit as using a mono CCD through filters, where when running in 2x2 binning mode you would expect 4x improvement in sensitivity.

At the same time, however, you can grab all of the basic color info from a single exposure series, instead of having to run through 3 individual exposure series with color filters. So there is some time savings here with the backward technique. (You use what you have...)

But the OSC CCD does give more than a factor of 2x improvement in sensitivity, which is still quite good.

The coloration details in the bottom image are a result of the B&W image at half the exposure of the color field image. And there are more coloration details in the bottom image than in the top right image.

All images were stretched equally in chrominance. The B&W image was further enhanced to reduce the contribution of the myriad field stars. It's intensity modulation shows through in the bottom image, while the colors are granted by the 1 hour exposure of the color image.

I suspect that to get the color information required for a decent bLRGB image, you don't really need to expose the individual Bayer pixels for twice as long as the 2x2 binning mode, as I did here. You could probably get the color info in the same or shorter exposure. I just did this tonight in this way to get a feel for how much more sensitive the 2x2 binning mode is, over the individual Bayer pixels. And this experiment shows it to be more than twice as sensitive.

Backward LRGB looks pretty good! Spending 1/2 hour in 2x2 mode, and another half hour in 1x1 mode should yield more than 1 hour in 1x1 mode. I like that kind of economy.

---

But wait a minute! Isn't the end result at half your pixel resolution?

Yes it is... but in my case, maybe not yours, I'm using a Canon lens that was designed for a Canon sensor. My Canon EOS 6D camera has 6.5 micron pixels. My ATIK 490 has 3.69 micron pixels. So when I dowsample my color images to match the 2x2 binning B&W image, I'm really running at just about the resolution intended for my optical train. Examination of the results with the smaller pixel size in 1x1 binning mode shows that my sensor has more resolution than my lens is capable of providing.

So downsampling by a factor of 2 doesn't hurt very much. It ends up showing Canon-like details with vastly improved temperature control (readout noise and dark current), and vastly superior red sensitivity - which is the main reason I sought to attach my ATIK camera to the Canon lens.

-----

The final image is nothing more than the bLRGB composite discussed above.