With your permission I'd like to make a few suggestions with regards to workflow.
One of the biggest attributes I've found with regards to final output was with that of base image definition. Where final IQ is entirely dependent on the definition(or grain as I like to call it) of the base file. With this in mind, I usually start by assessing what options I might take to improve and achieve the highest quality base file at the pre-post-processing stage of things.
One example of this might look like the following:
Post process RAW and carrying out all RAW adjustments while eliminating any and all grain altering adjustments(sharpening, tonal and distortion corrections etc etc). The objective here being to preserve and/or promote the absolute, most neutral image grain possible. However all shadow and highlight adjustments would best be carried out in RAW prior to export(since photoshop does not access RAW data).
One other factor that I consider important with respect to high ISO processing is the use of Adobe's latest chroma NR engine. And though I keep mentioning Denoise in most of my samples, I also think it's well worth mentioning that I couldn't achieve near any of these results with Denoise alone(less ACR6 or LR3). Therefore the two are synonymous with respects to the quality of my own experiences.
And finally if there was one other "big issue" worth talking about(with regards to Denoise), it would be the creation and use of presets(can't stress this enough), since the software is quite demanding(slow) and the last thing us older system users need, is to wait minutes at a time while trying to tweak our way through images.
Beyond this, I think that end results really come down to targeting and addressing the various noise types within a scene. My usual approach is to break it down into the following three sections:
1. Deep or heavy shadow regions.
2. Moderate or midtone areas(skin and average tonal regions)
3. Highlight and details(hair, eyes, lips etc etc)
I usually keep a few variations of these in preset form within Denoise, which I then try out(quick and dirty) to see how they look. In most cases, I can complete an image in less than 5 using presets. Though sometimes(if I need to adjust things or go out of bounds), it can take a little longer. But for the most part, the presets are what I use.
And lets not forget the power of masking too. If you use a graphics tablet, then you may already know how versatile this can be(pressure sensitive), as we mask in NR and details between layers in an image. Which can of course be done with a mouse(plenty of people do it), but... it's always more convenient with a pen if you have one.
And finally... don't forget to wash those files with
GordonBee's TingeRemoval tool I know Denoise does banding and tinge(black level) correction, but both of these don't seem mature enough to replace the legacy methods(GordonB and Dfine debanding) just yet imo.
Hope this helps...
Last edited by JohnBee; 08-08-2010 at 12:33 AM.