I've recently taken a couple of images of emission nebulae, for which I gathered immense amounts of data, through seven different filters (L, R. G, B, Ha, OIII and SII), using the same equipment, from the same location high in the northern New Mexico mountains.
Being curious, I thought I'd process the data in all the ways I know of that an imager might want to present an image of an emission nebula containing significant amounts of both Ha and OIII data (and some SII data).
So I did.
Here are links to two sets of images (NGC6888/Cresent Nebula and NGC7380/Wizard Nebula); both contain NBRGB (with Ha and OIII woven into the broad band data), LRGB (no narrow-band data), HOO (bi-color, mapping Ha to the red channel, and OII to both the green and blue channels), SHO/Hubble and pure Ha (both also have an OIII image, and the NGC7380 has an SII image; it's interesting to see so graphically how different they are). All versions of each image are cropped identically, and set up to allow one-click flipping through the stack.
I worked hard on all images, trying to come up with a fair set of images for comparison, having had no pre-conceived notion of what my order of preference would be.
Also, I don't subscribe to the school of Hubble imagers that eliminates any traces of green in the SHO image, since I feel it would be a bit odd, in an object dominated by Ha emissions, not to have some significant green in the final image (since the Hubble palette maps Ha emissions to the green channel).
My takeaways (I'd be interested to hear yours)
1. I had never done an HOO/bi-color image before, but I like the HOO version of both objects; I'll be doing more of that. But I don't think, in general, that the slightly odd colors of the HOO method results in a prettier picture than the NBRGB version. Well, except I like my bi-color version of the Crescent Nebula better than any of the other versions of that nebula that I did. But, if one doesn't think of it as a "true-color" image, it's very, very pretty, I think (if the object has a lot of OIII emissions, of course).
2. Pure NB data is smoother and more contrasty than BB data, IMO (not really subject to reasonable debate, no matter how much data you gather for a BB filter; the luminance layer for the NGC7380 consisted of 41 fifteen-minute images, and is amazingly noise-free, but still isn't as smooth as the Ha image). So, in these stacks, both "ordinary" LRGB images appear coarse compared to the other versions, in large part because both are located in dense star fields, and BB stars are a good deal larger than NB stars. But I actually like the LRGB version of NGC7380 quite a bit; the LRGB version of NGC6888 not as much.
3. Perhaps related to the previous point, NGC6888 is a very difficult target, with lots of little, wispy bits to the bubble, rather than a solid bubble (like NGC 7635). I don't much like the pure LRGB version, and even the NBRGB version (which I feel is a well-processed image using excellent data) isn't as visually-pleasing as the bi-color, because of the coarseness of the nebula and density of the star field (yes, there are methods of getting rid of stars, but I don't subscribe to that school, either).
4. It's a LOT more work properly processing an NBRGB image than any of the others.
5. Hubble palette/SHO images are interesting, to me, in direct proportion to how much of the sky-blue-lake look they have. I like the SHO version of NGC6888, but not nearly as much as the HOO version, or even as much as the NBRGB version; but the SHO version of NGC7380 is, to me quite lovely, and rivals the NBRGB version.
6. Monochrome NB images can be very beautiful. I've long posted versions of the Ha-only data, but I really liked the pure-OIII version of the Crescent Nebula, since it was so different from the Ha, and also had a lot of data. Also, SII and OII have much larger stars than Ha, which probably accounts for the magenta outline of most stars in an SHO image.
7. Astrophotography is fun.
Edited by WebFoot, 09 November 2019 - 09:32 PM.