Processing Clusters

[MattJenko]

I have a little question about cluster processing, which is in effect a question about processing stars. I actually not sure what question to ask though, it's more of a feeling that my cluster images (I am a rather big fan of cluster images) are not quite what they could or should be. I have tried a few different approaches, but have decided that I should ask for some help. I think I am sharpening slightly too much, am unsure what to exactly do with the backgrounds (darken to highlight the star colour, or try and leave natural) and how to prevent washout. I am moving from a DSLR to a newly purchased 414ex as of this exact moment, so hopefully this washed out aspect will be remedied with a hardware upgrade.

Here is a recent example. I managed to not overexpose the core, and tried a few HDRWavelet routines to try and get some detail in there. I still think it could be loads better.

Thanks

Matt

[ollypenrice]

I don't shoot luminance for clusters, just RGB. L really just helps blow out the stellar core colour.

I use at least two different stretches, a more aggressive one for the outlying fainter stars and a softer one for the core. The core should not only not be over exposed but should also be made with a very high contrast curve so as to preserve the separation of the central stars to stop them merging into one blob. Black clipping the soft stretch is not a crime because you won't be using the clipped part. This can help the core star separation.

A purist would then paste the soft exposure over the deep one, separate them with a layer mask, paste the soft version onto the layer mask and then blur the mask and manipulate it Curves to get a sweet blend of the two. (The standard M42 technique.)

A pragmatist (or barbarian!) might paste the hard stretch onto the soft and just use a soft edged eraser at low opacity to ease off the over exposed top core, making the brush a little smaller between interations.

At the risk of general chastisement my own forays into clusters have been of the pragmatist variety, viz...

http://ollypenrice.smugmug.com/Other/Best-of-Les-Granges/i-34m3hqS/0/L/M13_TEC_ODK%20V3-L.jpg

Olly

[MattJenko]

Thanks for the detailed response Olly. I can safely say I don't think I would have stumbled over that approach myself any time soon. I assume the theory still holds when using PixInsight, although I am happy to mix and match my processing software usage.

I am thinking that Open and Globular clusters are a good fit for my small chipped Atik I have now got, and also are a good solid set of practice targets for my processing skills as well as being very worthwhile and beautiful in their own right. They also don't require as much overall exposure time than fainter nebulas as far as I am aware. Good images of clusters, such as the one you link to, seem to be rarer than the more popular nebula images, and more often than not don't match the visual appeal of these objects. Thanks as always.

[MattJenko]

I followed a PixInsighty version of your advice using Harry's tutorial on layering in a MaskedStretch over a straight Histogram stretch and the results are considerably better in my eyes - thank you: