Local Lab build


(Cyril LAPORTE) #301

OK nickel ! Merci


local lab multipoints
(Gabriele ) #302

Hi All,
I’m using this tool since little time and I think is very useful but I have a question:

Is there possibility in the future We can view alls spots alls togethers and show/hide alls with a simple tool click??


(dngimage) #303

It is nice to be able to use the Local Lab to blur the background.
The only thing is that sometimes it is hard to select what needs to be blurred and what should not be blurred. In this example, the subject’s shoulders are blurred which I didn’t intend to do. A selection tool in this case would be ideal. The user could select the subject better.


Before


After


(Stefan Chirila) #304

do you guys have a debian compatible deb file by any chance ?


(Desmis) #305

Hello
I just push a change

  1. algorithm for blur and noise is now similar to others, therefore, more accurate
  2. I add a new method for blurr and noise : symmetric…it looks like “inverse”, but uses the algorithms of “normal” with scope, etc. Be careful…some results may seem weird :wink:

I add also others improvments; not visible to the user

jacques


#306

RawTherapee_newlocallab_5.3-229-gff26f064_WinVista_64.zip is uploaded at
https://drive.google.com/open?id=0B2q9OrgyDEfPS2FpdDAtMVI1RG8


(Sebastien Guyader) #307

Jacques, did you update rawpedia with this new information?


(Desmis) #308

Sébastien

I’ll do it :slight_smile:


(Desmis) #309

c’est fait !


(dngimage) #310

Thanks @gaaned92 and Jacques!

New Local Lab:

RawTherapee_newlocallab_5.3-229-gff26f064_WinVista_64.zip

From:

https://drive.google.com/drive/folders/0B2q9OrgyDEfPS2FpdDAtMVI1RG8

Is running fine here! :sunglasses:


(Alberto) #311

I can’t find Panasonic FZ1000 lens anymore with this version (RawTherapee_newlocallab_5.3-230-gc77913b7_WinVista_64). It’s not under panasonic nor under leica as it was before, what’s happened?
I hope you can restore it. Thanks

PS: rollbacked to RawTherapee_newlocallab_5.3-166-gb19e1273_WinVista_64 and now I have FZ1000 lens under Leica
PPS: LocaLab and wavelet are so powerful!


#312

I locally makes corrections on lensfun DB waiting for definitive corrections by the lensfun team.
I am going to rebuild newlocallab.
Could you upload a FZ1000 photo on filebin and permit to provide it to lensfun.

edit :RawTherapee_newlocallab_5.3-231-g0f51220d_WinVista_64.zip
Uploaded


(Alberto) #313

I have uploaded a raw file (the same raw file used 28 days ago) here

https://filebin.net/pj9ux103fwta4hdg/P2000678.RW2

You can share it with lensfun and do whatever you need to improve RT.

Thanks for 231 version, i see my lens and many other Fixed lenses that were missing.


#314

@alb1x3 thanks


(Desmis) #315

The U-points and Nik technology have just been bought by DxO © in October 2017 from Google ©.

DxO Announces Local Controls Incorporated at DxO for Spring 2018 with U-Point Technology

Jacques


(dngimage) #316

@gaaned92 and Jacques!

Testing New Local Lab:

RawTherapee_newlocallab_5.3-231-g0f51220d_WinVista_64.zip

From:

https://drive.google.com/drive/folders/0B2q9OrgyDEfPS2FpdDAtMVI1RG8

and it is running fine here! :sunglasses:


#317

Ahoy,
@jdc I’m still testing your locallab branch and really like where it is heading. Oh my, oh my… why didn’t I get into programming languages? I should have! :slight_smile: But let’s get to the point - take a look at this picture:

.
In this case, I would like to exclude the blue area on the left side from the influence of exposure. Adjusting the scope and playing with the algorithm doesn’t do the trick. Let’s add another u-point with negative exposure value:
As you can see, it didn’t solve it. In NIK software we have similar problem:
This can be easily fixed by adding another excluding point:
That kind of operation allows for a greater control. Would it be possible in RT?


(Desmis) #318

Thank you for testing Locallab deeply.
The point you mention is a noticeable improvement, when the “RT-spot” is in an area where the colors are close, such as a pale sky and a snowy landscape.

This improvement is one of the points to be innovated in the future. As for example another way to detect the shapes of the spot object. That’s what I said in “Rawpedia”.
This other form of detection is in my opinion essential if we want to see “excluding points”

"Remarque : il doit être possible de programmer d’autres algorithmes de détection de forme que celui ci-dessus fondé sur les différences de deltaE.
*Par exemple les algorithmes de détection de bords (edge) tels que “Canny” ou “Sobel”;
*Ou encore le principe de décomposition en ondelettes.
Mais dans ces 2 cas tout reste à faire !"

Other improvements can be developed such as saving / restoring a RT-spot, or improvment to remove dust-spot, etc.

Nevertheless these evolutions must be preceded by a fortification of the current “locallab”. This is the subject of an issue that I opened a month ago on GitHub#4094
https://github.com/Beep6581/RawTherapee/issues/4094
Jacques


(dngimage) #319

@gaaned92 and @jdc !

Testing new Local Lab:

RawTherapee_newlocallab_5.3-256-g0a65cb40_WinVista_64.zip

From:

https://drive.google.com/drive/folders/0B2q9OrgyDEfPS2FpdDAtMVI1RG8

It’s running fine here! :sunglasses:


(Desmis) #320

I just pushed a change for the branch “newlocallab”.

I have introduced the concept of “Exclusive RT-spot”, quite close in the principle of “Excluding points” of Nik Software.

The system is not perfect, but should satisfy at least 70% of cases. :slight_smile:

See the French documentation on Rawpedia.
http://rawpedia.rawtherapee.com/Local_Lab_controls/fr#Les_2_types_de_RT-spot

I also fixed a series of bugs

jacques