Very interesting, it would definitely explain some of the odd results I've gotten from the script periodically. Replacing the CIEh() with H() in the script does indeed produce results more in line with expectations from the color wheel, as well as the hue angles specified by the buttons. Thanks for pointing this out!
Interesting. I went ahead and did some tests. The h (hue) angle in CIE Lab color space is defined as the angle between the +a axis (red) and the Crominance point. Coordinate conversions between color spaces can be calculated on the EasyRGB website. I created a pure RGB red image in Pixinsight (255,0,0) and measured it with the probe. RGB and HSV values match with EasyRGB and Photoshop pure red but CIE values (both Lab and LCh) in Pixinsight differ from EasyRGB and PS. I am wondering how the CIE color space is implemented in PI?
The Colormask script seems to be untouched since 2017 if you open the JS file. I'd say it could use some TLC. If you look at the source you can see the script indeed uses pixelmath with the CIEh() function, it should perhaps be easier to just change that to use the H() or Hd() function and forego all the the mad lookups to get the right value. The inputs in the dialog seem to match the output of H() better at any rate. After seeing this I'm gonna try and make a patched version for my own use anyway.
I understand- I created this years ago as well..it is old- but no corrections have been made. At the time, I didn't know to look at the script code (or really how to interpret it). So, this was a video that was a journey of discovery. As I said, I kept it behind my paywall for years since I didn't think I was an expert and no one seemed to care it didn't make sense. There are other videos of this nature on my site as well.. perhaps I keep those in the vault also. I am constantly accused as self aggrandizement for this stuff.
@@AdamBlock Sorry to interrupt - not my post - but it is sad if anyone says correcting software errors is self-aggrandizing because (a) these script errors need to be identified so that hobbyists do not waste precious time fighting the software incorrectly believeing we are doing things incorrectly; (b) you have 8,000 YT subscribers so, by definition of the word, there is nothing for you to aggrandize (you're already well known); and (c) disclosure might inspirre program management to correct/improve the script. And hey Benny -- I was thinking of doing the same thing but I am even less skilled at coding than I am at post processing. I hope you will make your mods public (and risk being accused of self aggrandizement, lol).
Hi Adam, I think the script can be fixed by replacing in the PixelMath expression the call to CIEh() by a call to H(). At least for me it produced the expected result.
![Seungmin "그렇게, 천천히, 우리(As we are)" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/kAzmhLHePqU/mqdefault.jpg)
Thank you for doing this Adam -- excellent analysis.
Thank you!
This video was very much needed! Thank you Adam!
Thanks for watching
Very interesting, it would definitely explain some of the odd results I've gotten from the script periodically. Replacing the CIEh() with H() in the script does indeed produce results more in line with expectations from the color wheel, as well as the hue angles specified by the buttons. Thanks for pointing this out!
Amazing find! Thanks for this video!!
Thank you for watching!
Interesting to watch how you make this painting, the technique you use is fascinating. 🙂👍
Thank you Adam! These videos are gold! I wonder if he still wants to update the script...?
Excellent find Adam!!! Looking into the math now 🙂
It is probably a 5 minute fix really. :)
Interesting. I went ahead and did some tests. The h (hue) angle in CIE Lab color space is defined as the angle between the +a axis (red) and the Crominance point. Coordinate conversions between color spaces can be calculated on the EasyRGB website. I created a pure RGB red image in Pixinsight (255,0,0) and measured it with the probe. RGB and HSV values match with EasyRGB and Photoshop pure red but CIE values (both Lab and LCh) in Pixinsight differ from EasyRGB and PS. I am wondering how the CIE color space is implemented in PI?
Thank you Adam. Now I know why I struggled to get this script to work properly.
Thanks for watching.
The Colormask script seems to be untouched since 2017 if you open the JS file. I'd say it could use some TLC.
If you look at the source you can see the script indeed uses pixelmath with the CIEh() function, it should perhaps be easier to just change that to use the H() or Hd() function and forego all the the mad lookups to get the right value. The inputs in the dialog seem to match the output of H() better at any rate. After seeing this I'm gonna try and make a patched version for my own use anyway.
I understand- I created this years ago as well..it is old- but no corrections have been made. At the time, I didn't know to look at the script code (or really how to interpret it). So, this was a video that was a journey of discovery. As I said, I kept it behind my paywall for years since I didn't think I was an expert and no one seemed to care it didn't make sense. There are other videos of this nature on my site as well.. perhaps I keep those in the vault also. I am constantly accused as self aggrandizement for this stuff.
@@AdamBlock Sorry to interrupt - not my post - but it is sad if anyone says correcting software errors is self-aggrandizing because (a) these script errors need to be identified so that hobbyists do not waste precious time fighting the software incorrectly believeing we are doing things incorrectly; (b) you have 8,000 YT subscribers so, by definition of the word, there is nothing for you to aggrandize (you're already well known); and (c) disclosure might inspirre program management to correct/improve the script. And hey Benny -- I was thinking of doing the same thing but I am even less skilled at coding than I am at post processing. I hope you will make your mods public (and risk being accused of self aggrandizement, lol).
Hi Adam, I think the script can be fixed by replacing in the PixelMath expression the call to CIEh() by a call to H(). At least for me it produced the expected result.
Yeah..I think that is right. It is easy to do...now that you know WHY you have to do it. Someone had to figure that out in the first place. :)
I am making a video now... lol
I can not look at zebras again.
haha