Page 1 of 1

rosette nebula using both AutoDev and Dev

Posted: Tue Mar 03, 2020 1:20 pm
by alacant
Hi everyone
In an attempt to come to terms with the differences between AutoDev and Develop I processed the same data using each. Note how the stars and detail are controlled in each. For me, AutoDev is the winner.
https://linuxcostablanca.blogspot.com/2 ... -rose.html

thanks for looking and do post your thoughts and images using both methods.
Cheers,
Steve

Re: rosette nebula using both AutoDev and Dev

Posted: Tue Mar 03, 2020 2:05 pm
by happy-kat
Did you use ROI and play with the ignore fine detail slider?
I am finding autodev needs hand manipulation to bring out the stuff I want in my poor data but in general does do a better job than dev with care.

Re: rosette nebula using both AutoDev and Dev

Posted: Tue Mar 03, 2020 4:47 pm
by alacant
happy-kat wrote:Did you use ROI and play with the ignore fine detail slider?
I am finding autodev needs hand manipulation to bring out the stuff I want...
Hi Yes.
I think all stretching needs some sort of manual intervention.
I just think that AutoDev is streets ahead of the mask-stretch-hope-for-the-best approach of other software.

Cheers and thanks for looking,
Steve

Re: rosette nebula using both AutoDev and Dev

Posted: Wed Mar 04, 2020 2:27 am
by admin
Hi Steve,

That's an excellent example of applying a manual Develop vs using Autodev! :thumbsup:

As you note, you "prefer the stars of AutoDev, and the detail of Develop".
AutoDev, when used for the final stretch, was built to give you the most optimal starting point for further detail enhancement, attempting to allocating dynamic range equally to all areas.

Your example perfectly demonstrates the difference between;
  • the old way of using a global stretch to bring out desired detail in the image straight away
vs
  • the new (preferred) way of using the most neutral stretch to show all detail equally, allowing for targeted, advanced local detail enhancement
The first method is akin to making sculpture by adding on bits of clay to the areas that don't have enough detail. You're mostly repairing detail and you'll likely by applying filter-upon-filter (tacking on bits, removing bits) to get somewhere close.
The second method is akin to making a sculpture by shaping a roughly formed block of marble with fine tools. You're mostly shaping and refining detail. You will be quickly converging on a refined shape.

The latter method is much more targeted and "zen" with a much more defined end-point you will reach much quicker. Case in point, an image I just saw posted on Reddit; "Total Processing Time: ~13 Hours" :shock:

Re: rosette nebula using both AutoDev and Dev

Posted: Wed Mar 04, 2020 10:02 am
by alacant
Thanks Ivo for taking a look and for your comments.

Yeah, it's the bit after the second AutoDev where I'm failing. Here's a better example of where the detail present after Develop doesn't appear after the second AutoDev.
Is this a case of using the next items on the st workflow -contrast, HDR, sharp and Decon- to greater effect?

TIA

Re: rosette nebula using both AutoDev and Dev

Posted: Wed Mar 04, 2020 11:16 am
by admin
alacant wrote:it's the bit after the second AutoDev where I'm failing. Here's a better example of where the detail present after Develop doesn't appear after the second AutoDev.
That Gallery post links to a HDR composite of M42, is that what you meant to link to?
Is this a case of using the next items on the st workflow -contrast, HDR, sharp and Decon- to greater effect?
Indeed, that is what I was trying to get at. Though if you are having trouble, please feel free to share any dataset!
EDIT: The new Entropy should also be considered. To riterate, the idea is that you can have the best of both worlds; good shadows (with more nebulosity than just a regular Develop), good highlights (stellar profiles) and - through local dynamic range manipulation - more detail than just a regular Develop.

EDIT2: Side note, the interior of the Rosette is a cavity; a "hole"/"bubble" where bright, young O and B class stars have blown away the dust in which they were born. Their strong radiation is what powers the emissions. You will find that there is some dust and faint blue reflection in the foreground and background. in that "hole"/"bubble" as it is really a 3 dimensional structure. E.g. it's not as "black" as the background and the AutoDev rendition is closer to reality.

Re: rosette nebula using both AutoDev and Dev

Posted: Wed Mar 04, 2020 1:21 pm
by alacant
Hi
Sorry for not being clear.
Yes, I also wanted to link to that blog post as a further -more extreme- example of the differences between what I'm getting with AutoDev and Dev. The last two images therein show exactly the same problems of detail which I had with the rosette.

If anyone has the time, it would be interesting to see a st guru get the detail out of this after AutoDev AutoDev...

To save confusion and retain context, I've put links to the images in the original post

Cheers and clear skies,
Steve