Local Lab build

Si c’est juste un manque de compréhension du système de ma part, ne te casse pas la tête !

Ok, I think that using “Threshold deltaE-scope” and “deltaE weakening” sliders seem to improve the result of shape detection.

Another thing: in blur, the difference between 99 → 100 in scope is very big, much bigger than 98 → 99

Je pense quand même qu’une chose qui n’est pas intuitive, indépendemment de la détection de forme : quand scope n’est pas élevé, on obtient en fait comme une superpostion de 2 images sous le spot : l’image d’origine et l’image floutée sont superposées, donc on voit encore l’image nette en-dessous.
C’est très bien pour créer un effet de brouillard, mais c’est différent de l’application normale du floutage je pense.

Je vais regarder cela demain.

Est ce que tu peux m’envoyer le raw et le pp3 plutôt que de parler dans le vide.
Merci

Le raw est celui de @gadolf : [Play Raw] Street scene challenge

Voici le pp3:
_MG_4335.jpg.out.pp3 (53.5 KB)

Avec flou:
_MG_4335

Sans flou:
_MG_4335-unblur

Pour ne pas avoir l’effet brouillard sur le début du parapet, j’ai dû mettre scope sur 100.

@sguyader
J’ai regardé ton “travail”…beaucoup de spots

J’ai changé la progressivité de “scope”, au lieu d’avoir une valeur indéterminée (pour l’utilisateur) de reducdE et passer brutalement à “1” pour 100% (ce coefficient reducdE est la reduction liée au deltaE), j’ai changé l’algo de telle manière que à partir de scope=80, il y a croissance de la valeur correspondant à scope=80 (par exemple reducdE = 0.67) pour atteindre 1.0 pour scope = 100, progressivement

jacques

I made a Topic about Graduated Filter in RT almost 1 week ago , but did not get any response.

Since I consider GF as a part of local adjustment I ask here instead.
Are You developing it any more , or have any plans of doing that ?

I read your earlier topic, but I didn’t really understand what point you tried to make or what you were asking.
But it likely fits better here indeed.
Have you tried the local build to see if it allows you to do what you’re after?

Further, graduated filter seems feature complete. It mimics the graduated neutral density filter on a camera. What more is needed?

There is a lot more that can be done with GF.

Did you see those two pictures - I made a White Balance GF.
How can it be done in 5 sec in RT ? … Can it be done at all ?

What if you take a picture with well defined clouds in the sky.
I wanna add a lot of contrast to the picture , but if I do that the clouds is getting this unrealistic HDR look which i don’t want. Instead i rather add a contrast GF to the lower part of the picture , and leave the clouds untouched.

And this is just the beginning , like I told in the other thread , the pros use GF a lot ( in Lightroom ) and they often make multiple adjustment for each picture.

If you want to formally request a feature or enhancement, you need to do so on github.

Ok , but many of the developer in Github is often visiting this thread.

And as I earlier mentioned , the GF in RT already works great , all they have to do is to add more of the basic adjustment…

Sure they do, but it isn’t their job to interpret what you’re trying to say, then open a feature request. They also should have to trawl the forum for feature requests. They should spend their time writing code.

1 Like

The concept behind LocalLab (@jdc, Jacques, please correct my if I write nonsense here) is a bit different.

It’s about defining a region the modifications should be applied to (an ellipse or a rectangle atm) and a spot inside this region which defines the colours and luminance inside the region where the modifications should be applied to. Currently the definition of the region imho is not flexible enough. That could be enhanced by allowing finer granularity or (a simple way) an angle.

Currently you basically can define two kinds of regions:


and

As you can see in the screenshots the centre spot always has the same x coordinates as the upper and lower spot of the region and the same y coordinates as the left and right spot of the region.

Adding an angle could improve this (though of course more fine grained shapes would be even better)

1 Like

I’m asking you again: have you tried the local lab build? Even if it wasn’t intended to be used for gradients, it can actually be done.
You could try adding a rectangle spot to the top of your image, and use the color toning filter to change tone the colors and sort of emulate a WB change. And you can play with the transition slider to control the gradient.
The same with contrast.

LocalLab regions currently miss the angle which is available in gradients filter

There’s a lot of work to be done in LocalLab branch, but I’m hopeful that it will be done. At least when looking at the current coding pace and feedback :+1:

1 Like

True, it’s not as flexible, but in some cases such as the example @DxO-user mentioned it’s enough:

Here is the RAW file.

I have tried in RT , Local Lab - but can’t get the same result as in Lightroom

Here is a screenshot of my Local Lab
2019-04-11_022038
I guess I must be doing something wrong.