We threw the prawn on the barbie, and now it's distinctly less green. (The raw numbers are slightly orange of yellow. We don't want to go much further, because the area is overwhelmingly rich in H-alpha and we don't want to disguise that.)
You know, apart from the rings around the stars and the red flecks still visible...it is otherwise just about the perfect looking wavelength ordered emission line image really
Didn't mind the green in the first place. Sublime work.
Thanks, Marc.
Quote:
Originally Posted by strongmanmike
You know, apart from the rings around the stars and the red flecks still visible...it is otherwise just about the perfect looking wavelength ordered emission line image really
Mike
Thanks, Chief! The funny rings and coloured smarties will have to wait for software improvements.
Quote:
Originally Posted by gregbradley
That's a great colour scheme. I like that better than the Hubble palette.
What mix is it?
Greg
Hi, Greg,
It really is straight Hubble Palette - H-alpha to green, [OIII] to blue, and [SII] to red.
Compared with the previous version, I've used curves to increase mid-tone contrast in the H-alpha, which maintains the green in the highlights but reduces it over most of the image area, thus decreasing the overall preponderance of green. Similarly I've used extra arcsinh stretch in the [SII] and [OIII], which increases the relative overall amount of red and blue. I stopped when the overall image (before colour management) had roughly equal amounts of R:G:B.
Still don't like the magenta stars but the rest looks great M&T
Fair comment, Rick. As mentioned elsewhere, we're working on a new approach to the magenta haloes. That project is on the back burner.
Current software effort has been concentrated on rewriting Prometheus, our register-and-stack program, for 64 bits in C-sharp, to be able to handle larger images, do better data rejection, and get away from the extinct Borland C++. Yesterday it worked, after a fashion, for the very first time ! Bwahhhaha!
Current software effort has been concentrated on rewriting Prometheus, our register-and-stack program, for 64 bits in C-sharp, to be able to handle larger images, do better data rejection, and get away from the extinct Borland C++. Yesterday it worked, after a fashion, for the very first time ! Bwahhhaha!
Why did you pick C#, Mike? I would have thought that image stacking would still be more efficient in a compiled language rather than an interpreted/JIT one. Maybe I need to finally toss my old fashioned ideas about not wasting CPU cycles and memory. It's hard to do when you cut your teeth on embedded systems with 8-bit CPUs and KB of memory rather than GB
Why did you pick C#, Mike? I would have thought that image stacking would still be more efficient in a compiled language rather than an interpreted/JIT one. Maybe I need to finally toss my old fashioned ideas about not wasting CPU cycles and memory. It's hard to do when you cut your teeth on embedded systems with 8-bit CPUs and KB of memory rather than GB
A very good question. Speed: Straight up, processing images in C# is at least three times slower than in C++. But C# makes using multiple cores, and parallel processing generally, a breeze. Next, C# provides the keywords "fixed" and "unsafe", which let you write more or less pure C++ in-line, though it's a bit of a fight. Finally, for real speed on some heavy maths, I use C++ DLL's. Grand result: I can zoom and pan a 16 megapixel image at 70 frames a second. Ease of use: Writing event handlers for the user interface in 1970 Borland C++ was easy. In Visual Studio C++ it is pure torture. In C# it's a pleasant doddle.
I kicked and screamed for about four months, but now I'd never go back.
A very good question. Speed: Straight up, processing images in C# is at least three times slower than in C++. But C# makes using multiple cores, and parallel processing generally, a breeze. Next, C# provides the keywords "fixed" and "unsafe", which let you write more or less pure C++ in-line, though it's a bit of a fight. Finally, for real speed on some heavy maths, I use C++ DLL's. Grand result: I can zoom and pan a 16 megapixel image at 70 frames a second. Ease of use: Writing event handlers for the user interface in 1970 Borland C++ was easy. In Visual Studio C++ it is pure torture. In C# it's a pleasant doddle.
I kicked and screamed for about four months, but now I'd never go back.
Cool, thanks. I haven't had time to do anything more than a little scripting in PixInsight recently but I'm hoping to do some coding again when I eventually retire.
