Local Lab build

(Alberto) #421

RT Locallab 893 : If i duplicate the control spot 1 i get the values copied to 2, if i adjust the spot 2 and do the same thing (duplicate it to 3) it disappear…


I have a feeling that the duplicate button only duplicates the first spot to the second and not the selected spot to the next.

(Desmis) #423

@alb1x3 @McCap
I just comme back from holidays :slight_smile:

After commit and some changes, residual now works

For duplicated, I tested and found no bug… but I admit user manual is a little curious and deserves to be more ergonomic, to review when implementing a best GUI to manage spots .

You must to duplicate:

  1. create a first spot with it’s settings - suppose it is “1”
  2. move “control spot” to current + 1, in this case “2”
  3. click on “duplicate current spot” - all settings of the spot is duplicate, except center which is always at centerx=200, centery=200
  4. move and change settings of this spot
  5. if you want again to duplicated, move “control spot” to +1, in this case “3”
  6. click on “duplicate current spot”
  7. move and change settings,
  8. etc.
    but perhaps there is a bug I don’t found ?


(Morgan Hardwood) #424

If it’s the same issue, it was fixed in dev 16 days ago, and merging dev into locallab should fix it.

(Desmis) #425

I confirm residual is fixed :slight_smile:

(Alberto) #426

You’re right, “duplicate current spot” function works. I think something like “Copy from previous control point” would be more easy to understand. Something to select the original point would be even better.

(dngimage) #427

Open source now. Could it work with RT’s Local Lab?

(Gabriele ) #428

Hi jdc,
how do you make such plot?

(Desmis) #429

Excuse my bad english, but I don’t understand ! which plot ? for what ?


(Gabriele ) #430

The slice of the prophoto for L=88 …is not your?

(Desmis) #431

yes of course, but why a plot ?

(Desmis) #432

Now I understand, what you mean…I think

This slice is made with an old software, “Profile Editor 5.0”, you can see and compare all profiles (ICC, ICM), for Luminance between 0 and 100
For exemple for L=14, Prophoto, Widegamut, Rec2020, sRGB

(Desmis) #433

Also, with L=43, same profiles

(Gabriele ) #434

Thanks ,
closed source :neutral_face: and not for Linux
I use icc examin but I’m no able to make static slices


This is a Tensorflow model, RT would have to grow a Tensorflow branch, so probably not.

(dngimage) #436

I like how semantic image segmentation is used (portrait mode) to separate objects in the foreground from the image background.

Sometimes I use RT Local Lab’s Control Points to try to do that. BTW Thanks again Jacques!

More about SIS with DeepLab here:

(dngimage) #437

Using local lab:


I am getting purple highlights. Even in b&w Film Simulation (Kodak Tri-X 400).

I wonder if it is related to this:

Traveling and @ work now, no time to do much testing.

(dngimage) #438

I went back to:


and the problem is gone.

So the problem is present in the latest:





@jdc can you merge dev as @agriggio commited a modif for “the clipped highlights look psterized” bug?

(Desmis) #440

I just push this new merge :slight_smile: