Auto Dev ROI/Mask

[Troy Galebach]

I am not the most experienced member posting here, in fact this may be a totally stupid request.
Consider this situation:
I have an image of M31 it is really a good image as long as you don't mind the core of the galaxy being blown out.
It would be useful to be able to mask that core section and NOT stretch it as much as you do the galaxy's disk. Sort of a reverse ROI if you will. Instead of lighter processing of things around your ROI, you want the area around the ROI processed as normal, but the core processed so that you lessen the blown out core area. I chose Auto Dev, but perhaps this feature, if deemed useful, could be invoked in other modules as well.

I have no idea if this is a sane request or not. I was talking with a friend about this situation and decided I'd ask the question of a group of folks who could weigh in on the usefulness or not, of the idea.
Regards,
Troy

[hixx]

Hi Troy,
this is probably a pretty good idea...
many modules in ST ( I think including AutoDev) prevent clipping or have special tools to recover highlights ( like "Highlight Repair" in Color.
However, I think for this local M31 core part we should be OK, because what You asked for is already Implemented, but not in AutoDev: In the Contrast & HDR modules there are parameters dealing with recovering highlights (especially HDR's "Tame" parameter) . So my understanding is, that AutoDev's job is not to make it look great at this point, but to provide a most optimized, unclipped global stretch and hand that over to let the local modules do the local optimization job with their dedicated, taylored tools.
The drawback for a user is, You'll need to be able to anticipate the capability of the further modules and leave things open '( "not optimal") for further touch but having played around with a few dozen dataset, this experience should develop pretty quickly.
I guess, implementing a local tool in AutoDev (which is a global stretching tool) might make the math behind the things more complicated (actually no clue, but I suppose that's the reason why it's split into the various modules)....

Clear Skies,
Jochen

[Mike in Rancho]

Hi Troy,

I'm not sure. Especially about the mask part. I think there's been much discussion about why ST doesn't have, and likely never will have, masked stretching. It pops up every now and then here, though maybe for slightly different reasons than your issues, which seems more in the HDR realm.

For which, in most cases using ST, the controls between AutoDev, Contrast, and now 1.8 HDR are more than sufficient, such that masked layer-blending HDR techniques aren't necessary.

In fact, I find that ST's taming capabilities are so strong that one has to remember to not get carried away, and back off some. I've easily ended up with globular clusters and yes, M31's, that have cores way too deadened to look "right" when I step back and view the overall image.

Of course the first place to start is acquisition. If pixels are truly blown out they can't be recovered, really, and the only option would be a more classic layer-blended HDR of some shorter exposed integration. Though query whether you would be better off just using only the less exposed stack in ST for full processing.

The controls within AutoDev - from ROI to IFD to Outside ROI percentage, not to mention the hammer of gamma and shadow linearity - seem to provide an awfully wide variance of global stretch possibilities. And as Jochen says that is just your setup for following through with Contrast and HDR.

Feel free to provide an M31 stack that you think is being troublesome, maybe we can take a shot at it? Perhaps also a sample or link to an M31 that you would be hoping to emulate as the stretch you would like to achieve.