|
I think it mostly depends on the target as some are better suited to different palettes. Experimenting is a must. I still enjoy the Hubble palette but regardless of the palette you choose there is an underlying strategy that applies to all narrowband images. The goal is to emphasise differences between the different parts of the nebula. As you point out this is a challenge as Hydrogen Alpha typically dominates the signal. So to counteract this, you need to adjust the weight of each channel to maximise the use of the colour spectrum.
Russell Croman came up with a method which I've used with great success in the past. This approach is different to using photoshop clipping masks which I find gives finer control but in any case both work well.
On the FITS master of each emission line measure the background (in same location across all masters). If the background is 420 for Ha, subtract 320 using pixel math so that all emission line masters have a normalised background of 100. Now that the masters have been normalised. Perform a colour combine with weightings of 4:1:2.5 for SII:HA:OIII respectively. Check the balance of the combined image to see if these weightings aren't biased. If they are, you'll need to scale weaker emission line masters. I use MaximDL for this but CCDStack can do the same (I just fine MaximDL easier for this operation). To scale, use pixel math by reducing the constant and increasing the scale factor percentage. For example SII would have a constant of -400 with a scale factor percentage of 500 which still keeps your 100 count background ADU. You may need to scale higher if the emission is too weak. Once the weaker emission masters have been scaled they should closely match in intensity across the three emission line masters. You can then perform another colour combine with weightings of 1:1:1 as the scaling has counteracted the weighting. Make sure you turn off background auto equalize as you don't want to equalize a background that has already been normalised! Once done, save as a TIF and use this as the background image to a ha luminance for detail. Results should reveal 'punchy' details.
When it comes to colour mapping such as in Photoshop using clipping masks, its relatively simple to get the colours that Martin and others use. A colour map of R:240,G:150,B:30 will get you the muddy burnt orange hue that work well for SII data. Equally, OIII can be tweaked between aqua through to teal. Comes down to a matter of preference. I've always found the Hubble palette as the most aesthetic providing the Ha data is managed correctly.
|