ref image (which I doubled in size for this result): http://www.fontplay.com/freephotos/seve ... 806-03.jpg

First you dup the target image (again doubled in size already) and run Hard sketch:



Then you copy your target image to top layer and then dup that layer. Now run Bilateral smoothing on top layer:



Now you set top layer to Grain extract and merge down; this is your convolution layer.

Now set the convolution layer to grain merge and dup 6 or 7 times:

Now merge visible (did I not say to uncheck visible the reference? lol)

Now, on the multi-convoluted result, run anisotrophic (fake smudge) setting:



Now, run Cartoon preset on this result:



Now copy original on top layer and dup it.

Set Top layer to color and the lower layer to lighten and set opacity to taste:

The result (no extra nuances; I leave that to your liking):

Whoops; got lazy and just substituted name for the anistrophic preview. Anisotrophic smoothing is located under Repair. Sorry for my laziness. lol

Thank you so much Lyle, I'll do your steps, for sure. You aren't lazy, Lyle.

Lyle, you have done it again! Thanks for the tut.

My first step to make the tutorial in PDF file.

Thanks for the PDF Issabella!

Lyle great effect! Another winner.

Thanks for bundling the tut, Issabella. Appreciate the words, Rod.

This is cool!
I automated it with python and added emboss option.

Thanks a heap animicule. I just kept the layers and played with opacities and such for this result. Going to save me a heck of a lot of time now. Haven't tried your emboss step yet, but may toy with it later. Again, thanks.

ref: http://www.fontplay.com/freephotos/seve ... 606-09.jpg

You are welcome Lyle, thank you for showing the steps!

Wow! I love the painterly effect of this! Thanks for sharing Lyle!

Lyle, I love this photo technique and I am really glad that Animicule did that filter! You two have a winner

I'm grateful to Animicule too, Pat. Your results fantatic I might add.

Nice Panthers Pat!
Lyle, it tried to match your flowers at the top of this page and the page before - for hours.. close but no luck.
However, here is a new version with some different options for your entertainment.
There is a choice of 2 different layer mode sets (could add opacities in minutes if you like).
If you find a group of selections that get you close often, let me know what they are and it only
takes a copy minutes to add them in.

Here are a few using first version.

Very interesting results with several photos
which were hazy and with lack of detail in dark areas.

Emboss option was turned off, clean up layers was off
and double size was off also.
Photo was already 4435x2957

Top layer (0) - copy of original subject Normal Mode 51%
2nd layer (1) - generated - Hue - 100%
3rd layer (2) - generated - Color - 100%
4th layer (3) - generated - Value - 50% -- good control layer
5th layer (4) - generated - Hard Light 100% -- control on bias look
6th layer (5) - subject copy - Normal - 100%

pexels-photo-109030.jpg


Also noticed that the as the subject size gets larger, the size of the
'texture' does not expand at the same rate.
I did several at 500, 1000, 2000, 3000 width then sized them back to original size
and it is easy to see.

OH, Boy! You are on a roll again! Very nice images.

BTW, don't feel bad if you can't get Lyle's images copied exactly. I don't think anyone can, or not many anyway. He is sort of "unique".

Might have duped the convolution layer and set it to value and lowered the opacity to taste. Just don't actually recall all the manipulations. lol