Do have some core routines, but extensions to this depend on the target. In a summarised form, the core goes something like this;
- calibrate subs (darks,flats,bias)
- remove blooms if applicable
- register/align subs
- hot/dead pixel removal
- blink all subs to check for anomalies/discoveries
- combine subs (mostly sigma reject)
- Digital development each combined images to check for gradients. This also give an indication as to the quality of data and a hint of what the end result will look like. Cancel the DD process so no stretches are made.
- If gradients are present in the individual R,G,B frames, I address this before the combine if possible. Don't want to upset colour balance. Sometimes its easier to handle later in photoshop
- Manually normalise the individual R,G,B frames using pixel math. There are some nice tools available to normalise the background on OSC cameras such as CCD's or DSLR's.
- RGB combine, then save.
- Push luminance be in L or Ha through deconvolution routine. Either maxentropy or LR. I only go as far as tightening up the stars. This depends on if the data is oversampled, but in most cases some benefits are gained.
- Save the lum, then blink the original and deconvoluted lum to check for artifacts introduced by the deconvolution routine. If found, check pixel intensities. Deconvolution works well on brighter data. If needed, the data can be scaled to increase pixel values. The run through deconvolution again.
- I then bring both files into Photoshop and commence levels and curves work. Not a huge user of DDP for stretching data. It can work well if you've got good control over your mid point.
- Sometimes use shadow/highlights tool to perform a non-linear stretch on the colour data while maintaining saturation. You don't need to do this, but if you do use levels and curves, its wise to follow up with a saturation tweak so your colours don't get washed out. If they do, you can bring them back by doing a hard DDP stretch on a separate RGB combine image, then relayer it in photoshop as soft light.
- I usually end up with a few layers doing different things such as noise invert masks, SCM, gradient removal, I'm a sucker for adjustment layers when it comes to the finer tweaks. Other than on masks, I stay away from using the brush tools. Personally, if you need to use them, you're getting pretty close to the border line of ethics. The things you can do in photoshop to mask problems and/or enhance things is rather scary. Good data is always the starting point.
There are always a few other things to throw into the mix, but depends on the data. Some data you simply can do much with i.e too noisy for example. I've been working on doing DD stretches for the Ha and Luminance data at different intensities so simply the rich highlight of the Ha data come through the luminance, but not altering the star apperances when using lighten mode. Rather basic, but effective considering Ha stars are typically tight and small compared to RGB. There are many ways to do the same thing, but most have a different effect or result in introducing noise of some sort. Anyway, that's the core for what its worth. Ask me in a few months time and I'll probably find a different method of doing something.
