Cosmos
Posted 09 October 2023 - 01:54 PM
Fantastic. Looks like it's implemented on the lcms2 branch so would need to build that branch to test.
Sigh. I am not a computer scientist or developer…so I sure as heck don't have the JHBuild environment installed! But I do have homebrew installed, which the README in the tarball says is supported. Not sure if that'll load the gui, though…
Cosmos
Posted 09 October 2023 - 02:02 PM
Looks like homebrew only has 1.0.6 on tap.
https://formulae.brew.sh/formula/
On the other hand, the Octave version is way newer than my app version…
Edited by BQ Octantis, 09 October 2023 - 02:05 PM.
Explorer 1
Posted 09 October 2023 - 02:04 PM
With square root preview, I get the benefit of the much closer to reality gray line. But the masses using Siril get the red line.
Unless I misunderstand, your graph shows the effect of the same MTF stretch in display space. But this isn't what happens in practice. Suppose the green line, e.g m=0.25, represents the ideal stretch - the one that looks the best in display space when the display transform is applied correctly. A Siril user will find the MTF stretch that best fits the green line; however, this stretch will most likely not be m=0.25.
Edited by FrankieT, 09 October 2023 - 02:11 PM.
Cosmos
Posted 09 October 2023 - 02:07 PM
Looks like I'd have to open a cask…
https://docs.brew.sh...-Homebrew#casks
Good Lord…looks as complicated as getting a compiler to work!
https://github.com/H...w-cask-versions
Edited by BQ Octantis, 09 October 2023 - 02:09 PM.
Cosmos
Posted 09 October 2023 - 02:18 PM
Unless I misunderstand, your graph shows the effect of the same MTF stretch in display space. But this isn't what happens in practice. Suppose the green line, e.g m=0.25, represents the ideal stretch, the one that looks the best in display space when the display transform is applied correctly. A Siril user will find the MTF stretch that best fits the green line; however, this stretch will most likely not be m=0.25.
You're absolutely right. Eventually, at low values of m (<~0.01), even the display-gamma'd MTF approaches a gamma-corrected MTF with a higher m value. That's probably why no one complains. I'd need a lot finer gradations on my samples to quantify the difference…
BQ
Edited by BQ Octantis, 09 October 2023 - 03:37 PM.
Cosmos
Posted 09 October 2023 - 02:57 PM
Reading up on the MR…
"One use case is shown in !1207 (accurate conversion from linear light to a working space and accurate display on a monitor) in order to assess image chromaticity from a calibrated dSLR. Another use case is support for editing images in wide gamut color spaces." [Source, Emphasis mine]
Be still my beating heart! 
Explorer 1
Posted 09 October 2023 - 07:26 PM
Be still my beating heart!
Siril v1.3.0-alpha-dev preferences
Cosmos
Posted 09 October 2023 - 07:44 PM
I was hoping you'd pull the Windows version!
How's the monitor preview working?
The Intel Mac OS install is cactus at the moment…
Edited by BQ Octantis, 09 October 2023 - 07:58 PM.
Explorer 1
Posted 09 October 2023 - 07:59 PM
I was hoping you'd pull the Windows version!
Cosmos
Posted 09 October 2023 - 08:14 PM
Well...? How are the linear previews in display space?
Explorer 1
Posted 09 October 2023 - 09:58 PM
Well...? How are the linear previews in display space?
For reference, Mark's colour checker image was processed in darktable with only basic processing options applied; i.e., black point, white balance, demosaic, input profile (standard colour matrix), linear Rec2020 working profile and display encoding (sRGB).
Darktable Reference
The raw colour checker file was converted to fits in Siril, bias subtracted, debayered, white balanced and a sRGB CCM applied before saving to a 32-bit fits file. This represents a linear-gamma sRGB image as it would be processed in Siril.
The linear-gamma sRGB image in Linear preview mode. As expected, the image is rendered too dark.
The linear-gamma sRGB image in Linear preview mode with the sRGBlinear ICC profile assigned via the command line. The Linear preview now matches the reference.
Finally, the sRGB TRC was applied to the linear-gamma sRGB image using Pixel Math and saved as 32-bit fits. This represents an sRGB image.
The sRGB image in Linear preview mode with the sRGB ICC profile assigned via the command line. The Linear preview also matches the reference.
Explorer 1
Posted 10 October 2023 - 12:37 PM
To add, this is fantastic progress and amazing work by all the developers to differentiate Siril from other astro software packages.
I suspect there might still might be some bugs to iron out before release. For example, if I load a linear-gamma sRGB image with the preference to automatically assign a linear ICC profile, then I would expect the Linear preview to look the same as if a sRGBlinear ICC profile was assigned manually. However, this is not the case and I get a saturated preview as though the image was assigned an sRGB profile. This could be just be my misunderstanding of the settings so still early days.
Cosmos
Posted 10 October 2023 - 04:23 PM
Very exciting!
I, on the other hand, am still having issues with loading alpha software in Mac OS... 
BQ
Explorer 1
Posted 10 October 2023 - 04:33 PM
I, on the other hand, am still having issues with loading alpha software in Mac OS...
What can I say, ditch that rotten apple
Cosmos
Posted 10 October 2023 - 07:04 PM
Lol! You have no idea.
Cosmos
Posted 12 October 2023 - 05:22 PM
Turns out, there was a library link error in the Mac OS version. So testing with my rotten apple was not for naught.
As I ponder good color management, now I wonder what I will use for increasing chroma if it means asinh is now truly color preserving (vice hue preserving). I know for certain I'm not the only one who uses asinh color boost for these astro OSCs. OKlab chroma? CIE CAM02 chroma? This is unplowed territory for me…
BQ
Edited by BQ Octantis, 12 October 2023 - 06:00 PM.
Cosmos
Posted 12 October 2023 - 08:24 PM
Seems to work just fine in Mac OS Catalina.
Here's the hamburger menu of options…
You can see the severe negative sRGB values required for color matching in the statistics window.
BQ
Edited by BQ Octantis, 12 October 2023 - 08:25 PM.
Cosmos
Posted 12 October 2023 - 08:58 PM
Sure enough, asinh is no longer color boosting in linear TRC.
Auto MTF (in Histogram Transformation) also blows things out--even though Auto Stretch preview mode works fine.
Edited by BQ Octantis, 12 October 2023 - 09:24 PM.
Cosmos
Posted 13 October 2023 - 05:31 PM
Frankie, if you have any feedback, Adrian (the lcms2 developer) requested feedback be posted to the MR page.
https://gitlab.com/f...ge_requests/542
BQ
scout - Today, 04:15 PM
Refractors
dcm_guitar - Today, 04:05 PM
Experienced Deep Sky Imaging
therob - Today, 03:53 PM
Meade Computerized Telescopes
m1thumb - Today, 03:09 PM
Scientific Amateur Astronomy
 Cloudy Nights LLC Cloudy Nights Sponsor: Astronomics |
