Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
ofn-luminosity-masks
#1
My own take on luminosity masks. While it pays its dues to PatDavid's canonical tutorial on the subject, it creates the same masks using a different and (hopefully) more efficient method that allows:
  • to create a luminosity selection from the current layer in one single click (or one keystroke if you defined shortcuts)
  • in a future version, to define your own masks
As usual the script can be found here.  Restrictions in this version:

* Only PatDavid's masks, no custom masks yet
* Only for Gimp 2.10 (seems to hit a bug in 2.8)

Enjoy.
Reply
#2
The link needs to be corrected, it goes to a log-in page.
I did download from the usual place.
(This post can be deleted once link is corrected)
Reply
#3
Link ok.
This is the third Ofnuts plugin I am the first to download! :-)
Reply
#4
The DDD mask is selecting whites. (DDD is the only one I have tested.)

What I did:
Create a DDD mask and bucket fill with red.
Did the same for the Tin Tran version for comparison.


Attached Files Thumbnail(s)
   
Reply
#5
To see the mask, you can just use `Selection>Selection Editor (and enlarge the window...).

Found the problem, slight typo in the curve values. Will publish later today, in case another minor thing surfaces.

   


After the fix: The front of the engine, in black, like the sky and the platform, are not in the selection.
Reply
#6
Comparing the expected result (left) with the one obtained, one can notice a slight difference in the middle tones. (Maybe because I created xcf in version 2.10.12 and tested this script on 2.10.14)

But in DDD the difference is quite big.
Lights
   
Mids
   
Darks
   
Reply
#7
As noted there is a bug in DDD. If you want to fix it, before I republish replace the line:

Code:
('Darks/DDD',   'Standard mask',[0.00,1.00,0.33,0.00,1.00,0.00]),

The buggy DDD is more or less DDD+LL.

One can also ask what is the "expected" value, since no script strictly follows the original tutorial.
Reply
#8
(11-14-2019, 03:40 PM)Ofnuts Wrote: As noted there is a bug in DDD. If you want to fix it, before I republish replace the line:

Code:
   ('Darks/DDD',   'Standard mask',[0.00,1.00,0.33,0.00,1.00,0.00]),

The buggy DDD is more or less DDD+LL.

One can also ask what is the "expected" value, since no script strictly follows the original tutorial.

After the change, the DDD is as expected!  :-) (not speaking strictly)  Shy
I have tried to say that similar to the results obtained with both the Saulgoode script and those presented on Pat David's page.

This script gives me a lot of fun, thanks so much for updating it Ofnuts!
Reply
#9
Uploaded a new version. Added feature; define your own mask(s) in a side file. Enjoy.
Reply
#10
(11-16-2019, 06:14 PM)Ofnuts Wrote: Uploaded a new version. Added feature; define your own mask(s) in a side file. Enjoy.

Ofnuts,
But what if the curve also has points on other RGB channels?

The image shows the configuration of my curve. If I only use the values from line # 9 after Points 34, would it be enough for the mask to represent the curve?
   

I defined my mask without using the points of the Red, Blue and Green curves. There was a closeness to what I expected.
   
Reply


Forum Jump: