15th Feb 2024: Astro Pixel Processor 2.0.0-beta29 released - macOS native File Chooser, macOS CMD-Q fixed, read-only Fits on network fixed and other bug fixes
7th December 2023: added payment option Alipay to purchase Astro Pixel Processor from China, Hong Kong, Macau, Taiwan, Korea, Japan and other countries where Alipay is used.
APP - start to finish comparison: M38 and NGC 1907 to NGC 1931 and IC 417
Here are two version of 11.25hrs of OSC data on M38 and NGC 1907 to NGC 1931 and IC 417.
First version is APP with a lot of PS editing. However, every time I use PS or PI heavily, I always tend go back to a simpler natural form reversing many editing steps/iterations.
This time, I used the wonderful star reducer (with larger 1.2 and 1.3x factor for star radius and star halo) to reduce to 70% size and 90% intensity. The rest is color calibration, mild selective color enhancement, reduction of saturation threshold to 5% and reduction of saturation to 12%. I used the 20-5-0 autostretch with no contrast or sharpening. The result is the second image.
I worked out really well for an entire APP-only process. In the future, I might test some customized stretch, base and background values. What you do think?
@mabula-admin and friends, i think a slider-variable median filter might be great to add just a little noise/smoothening as a simple start to noise reduction, but I am happy with the APP-only process.
Both are nice, but like they are presented now, I choose the first image. Image #2 appears to 'loose' information on the low end of the dynamic range. There's quite some flat area in the histogram (when I load it into affinity-photo). I understand the comparison, in that #2 is meant to be solo APP, but like you already suggest; that is where I would try some customized stretch.
Thanks. I am in the middle. I prefer the stretching of the background weaker nebulosity and the dust, with the more natural stars (in situ processed) of the latter. These refractor stars, large ones with diffused color are not so easy to mask with such a varied background but starless and star merging in image 1 was reasonably OK. I might try to reprocess like image 1 while keeping the stars in place. However, the test got me a longer way in processing with only app than I ever did before so maybe the extra stretching might be more straight forward. Data from the epsilon 130d is a lot easier to stretch with the small spot size in this regards
Here are two version of 11.25hrs of OSC data on M38 and NGC 1907 to NGC 1931 and IC 417.
First version is APP with a lot of PS editing. However, every time I use PS or PI heavily, I always tend go back to a simpler natural form reversing many editing steps/iterations.
This time, I used the wonderful star reducer (with larger 1.2 and 1.3x factor for star radius and star halo) to reduce to 70% size and 90% intensity. The rest is color calibration, mild selective color enhancement, reduction of saturation threshold to 5% and reduction of saturation to 12%. I used the 20-5-0 autostretch with no contrast or sharpening. The result is the second image.
I worked out really well for an entire APP-only process. In the future, I might test some customized stretch, base and background values. What you do think?
@mabula-admin and friends, i think a slider-variable median filter might be great to add just a little noise/smoothening as a simple start to noise reduction, but I am happy with the APP-only process.
I would definitely choose the APP version (i am so biased 😊 though). The 1st version to me looks very artificial in the nebula parts which I just don't like (for me personally, this is overprocessing yes...) and the blue stars are more cyan than blue... so the colors in the APP version are definitely better as well.
The APP version lacks contrast when compared to the 1st though, which you could fix most likely with the CON slider in the preview filter, did you try that Colm @cwm2col ?
Noise reduction is soon to be there in APP for linear data.
I also think the APP version is better for this data set, does not show any overprocessing. The first shows what is there in the data, but extracting it shows it needs a lot more time or darker skies. My fast newtonian is a different story, much easier to process faint stuff in APP only. I have played with this version, stretching a little more, but it is touching on the edge of overprocessed again (attached image). And indeed, the color balanced suffered, which is easier to maintain inside APP only.
Although i look at the orange (?) symbol tool tip and examples of sliders with M42 you added to APP help files, can you tell me what the contrast slider works on in the data specifically. I sometime use sharpening and protection, but for star cores that needs to be done very carefully in the absence star masking for example.
I will try to see if the CON can be useful in this case, and I will repost that image too as the 4th and last image of this test.
I investigated the contrast option, as well as further tweaks in APP and settled on this version as an APP only full process. Managing the green using magenta-green HSL selective color variation with reduced saturation, and minimizing saturation for all colors form 0 - background, while balancing magenta-green in star color calibration helped also, negating any need to go outside the program. More data would help with left-over noise, or indeed the noise reducer tool that is coming. It give a nice natural look to the data.
