Hi there.
Is here anyone who can explain for me step-by-step what the Elsamuko National Geographic script does???
I want to be able to replicate it manually and fully understand what the script does step-by-step...
Anyone knows and can explain?
(i do not understand the scm file in text editor)

Thanks

Looks like the script applies a number of filters to a portrait.
https://sites.google.com/site/elsamuko/gimp/ng
To find out what the script does, open it in a text editor and read what filters it uses and the sequence they are applied in.
SCM scripts are generally quite readable.

well, mister, thank you very much to wasting the time. If i could read it, i wont write here that i cant. I did open the scm and had try to find out what it does, but there is more than filters only. there are some manipulations with layers, some layer deletings, invertings and many other stuf, and does not seems to me its written human readable way - its machine language i can not figure out.
Thats why i am asking if there is someone who can explain the procedure in human language, in chronological steps...
Thanks

Sorry to upset you, I didn't realise you had read the scrip. This script calls 2 other scripts (shadow recovery and smart sharpen) so there would be many steps in obtaining the finished image.
It looks like there are some user adjustable parameters in it as well.
The script does define what filters and layer manipulations are used. (contrast, 3 different layer copy modes, 3 new layers, change modes and opacity of those new layers, add layermask, add red and blue overlay tints)
Maybe all this gets anchored-down into one or two layers, I've never used it.

Not upset, just trying to explain.
That is exactly what i am looking for. A description step-by-step of all operations done by the script (with default values), to be able reproduce it manually...
Anyone who can explain, or already have seen a text explaining it step-by-step?
I found one, but it was for the older version and it was in german... So completelly useless for me on the most recent version

Or maybe an advise or recommendation to some other forum where i could find help with this???
Thanks

1 thing that could help you or any one wanting to understand scheme (script-fu's) is to read the many tutorials on this site Here [Fenceposts tutorials are excellent]

I do not think you are going to get that.

There are headings in the script which I pulled out for you. But you see even an outline is complicated.



A very similar end result is the Dragan Effect and here there are some tutorials in English to help (one attached).

edit: Should have just posted the link to the web site: http://www.mora-foto.it/en/tutorials-gi ... ffect.html I mistakenly thought it was dead

I would start with that and by the time you fully understand the procedure, you will half way to deciphering National Geographic.

No, it doesn't, the code of these scripts is shamelessly copied to/embedded in the code of the script.

To the OP, the problem with your question is that writing a real answer would take a long afternoon, (possibly wasting time on bits you have already understood) and nobody has really got this much free time. Over on Gimp-Forum.net we can try to address specific points.

Here is the description i found in german:
1. Kopiere das Originalbild 2 mal und wandle die Kopien in SW um und invertiere sie.
2. Nun gibst Du der ersten Kopie eine Ebenenmaske, kopierst die zweite Kopie und fügst sie in die Ebenenmaske der ersten Kopie ein.
3. Die Ebenenmaske leicht weichzeichnen (Gauss, je nach Bildgröße ca. 10-30 px)
4. Die zweite Kopie kann gelöscht werden; die erste Kopie mit der Maske nun auf Modus "Ineinanderkopieren" (probiere auch mal "Weiches Licht") setzen, Deckkraft anpassen, im script sind es 60%
5. Die Ebenen vereinen und zwei neue Kopien machen
6. Die untere Kopie kräftig mit USM schärfen (z.B. 2px, 200%)
7. Auf die obere Kopie einen Kantenfilter anwenden (im skript ist es Sobel, je nach Bild kann auch ein anderer gut sein). Das Kantenbild in SW umwandeln und den Kontrast erhöhen, das Bild muss weiße Kanten auf schwarzem Grund zeigen.
8. Nun dem geschärften Bild wieder eine Ebenenmaske geben, dann das Kantenbild kopieren und in die Ebenenmaske einfügen (wie oben auch); das Kantenbild kann jetzt gelöscht werden.
9. Die geschärfte Ebene mit der Kantenmaske auf Modus "Leuchtkraft" setzen, Deckkraft ca. bei 80%
10. Die Ebenen wieder vereinen
11. Den lokalen Kontrast erhöhen durch USM mit großem Radius und geringem Wert (ca. 60 px und 30-40%)
12. Jetzt nochmal zwei Kopien machen und beide in SW umwandeln
13. Die untere auf Modus "Bildschirm" setzen bei ca. 60% Deckkraft
14. Die obere auf Modus "Ineinanderkopieren" oder "Weiches Licht" bei 100%
15. Im Prinzip fertig, zur Verbesserung (empfohlen, im script als wählbare Option), der Bildschirm-Ebene noch eine Ebenenmaske geben, dort eine Kopie dieser Ebene einfügen, die Maske invertieren und weichzeichnen (wie in den ersten Schritten)
16. Zur Verstärkung des Effektes evtl. die Ineinanderkopieren-Ebene nochmal duplizieren und im gleichen Modus die Deckkraft verringern (ca.25%).

I dont understand german, the google translation is a mess in any language i know

lol

Hey there Graechan! Great to see you out and about. Hope all is well.

1. Copy the original image 2 times and convert the copies to SW and invert them.
2. Now you give the first copy a layer mask, copy the second copy and paste it into the layer mask of the first copy.
3. Slightly flatten the plane mask (Gauss, depending on the picture size approx. 10-30 px)
4. The second copy can be deleted; the first copy with the mask now on mode "copy" (try also times "soft light") set, opacity adjust, in the script it is 60%
5. Merge the layers and make two new copies
6. Sharpen the lower copy vigorously with USM (e.g., 2px, 200%).
7. Apply an edge filter to the upper copy (in the script it is Sobel, depending on the image, another can be good). Convert the edge image to SW and increase the contrast, the image must show white edges on a black background.
8. Now give the sharpened image a layer mask, then copy the edge image and paste it into the layer (as above); the edge image can now be deleted.
9. Set the sharpened plane with the edge mask to "luminosity" mode, cover force approx. 80%
10. Reunite the levels
11. Increase the local contrast by USM with large radius and low value (about 60 px and 30-40%)
12. Now again make two copies and convert both to SW
13. Set the lower to "Screen" mode with approx. 60% opacity
14. The upper mode of "Inverse" or "Soft Light" mode at 100%
15. In principle, to make a copy of this layer, invert and make the mask inverted (as in the first steps), for the improvement (recommended, in the script as a selectable option), add a layer mask to the screen level,
16. To enhance the effect, duplicate the interleave layer again and reduce the opacity in the same mode (approx.25%).

USM = Unsharp Mask i believe.
SW = not quite sure maybe black and white?
step number 10 probably means merge the layers

@Rod

SW is indeed black & white

#3 should read: Apply a slight gaussian blur to the layermask (depending on the imagesize 10-30px)

in #4 the tutorial speaks of the mode "Ineinanderkopieren". To my knowledge there is no such Gimp mode. But there was a mode named like that in older Photoshop versions. According to my research its similar to 'Overlay'.

#9 "cover force" is actually Opacity.

#10 Merge down the layers once again.

#14 Inverse mode = 'Overlay' Mode.

#16 interleave layer = 'Overlay' layer

Thank you

This is already a stone in the mozaic

Thank you so much for the time.
Im going to try out this, how ever, it seems to me that from the first two copies only one need to be desaturated and inverted (when watching slow-mo replay of what is happening with layers during the script)

Yes, only the "Shadow" layer has to be desaturated and inverted.

But then next, the line 5 "Merge the layers and make two new copies" gives no sense, as the second copy was already deleted in the step 4 and the first copy with the layer mask stays there as "Shadow Recovery" layer ... so there is no merging that point - right?

Tutorial (i.e. the found description) is very approximate.
There is no merge.
The 2 new copies are created "from visible".(1)

edit:(1) I mean, create "new layer from visible" (because you need 2, copy the first one created)

Well... i am analyzing the slow-mo screen recording of the script execution and here is what is actually happening:

1. Copy the original image 2 times and convert the second copy to BW and invert the image (color).
2. Now you give the both copys a layer mask, copy the second copy and paste it into the layer mask of the first copy.
2.1 Set the first copy mode to "Addition" and set its opacity to 60%

Until here its clear... but then suddenly the second copy disappears and the first changes into inverted BW with layer mask of inverted BW, it is renamed to Shadow Opacity and changes to Overlay mode... (it is not at all clear what happens there...)