skin retouching

[federicof]

Hi everybody,
i've already wrote on my presentation post that i'm a retoucher,
so, this is one of my last work, i've tried to obtain a "porcelain doll" aspect without losing the skin texture and details,
in this work i've used a loto of modern retouching tecniques, split frequency, dodge&burn etc,
when i'll be able to post a link, i'll be pleased to show you the live retouch video of this work if someone will be interested to it.
Thanks for your attention.
PS i'm not the author of the original shoot

 

[chrisdesign]

The quality of your retouching is of the best quality I've seen up to now in this forum. Congrats! The achieved porcelain look is great, the skin texture excellent, though not possible without working with split frequency technique. Would love to see your tut.
The only thing I don't like is the "hard edges" of her mouth. But maybe this is just me?

 

[federicof]

Hi Chrisdesign,
thanks for your reply and appreciation!
yes, split frequency is is one of the fundamental steps,
the workflow is about:
1) different RAW elaborations for different areas
2) split frequency to fix blemish and imperfections
3) Dodge&burn to micro-reshape of whole face and neck
4) some various creative passages to adjust color, tone etc
yes, maybe the lips have the edge too hard, but i wasn't looking for a very natural final result, I aimed to a kind of elegance of face features, like the dolls... but alive!
Many time, when i work on different projects, images for fashion editorials for example,
i'm very carefull to don't exagerate in that kind of details, but this time, i was more free!
However thanks for your opinion!
The video that i've realized is not a real tutorial, is without talked audio, only music, it's a kind of timelapse from start to the end of the work, i'll post on the next post, that will be my fifth!

 

[federicof]

here is the video!
http://youtu.be/ODNGNkWPBqg

 

[Crotale]

Pretty decent job, but I noticed a few areas where you went too far.

1. The left portion of the forehead merges with the temple over the brow looks very flat now. Needs more shading to match the original lighting effects.

2. The lip's left corner is too smooth and there is no lipstick there now.

3. The chin looks too flat or lopsided compared to the lighting of the original.

 

[federicof]

Hi Crotale,
thanks for your reply,
about the shadow on the temple you have probably reason, i sought an homogeneous whitening effect, and maybe i've exagerate in that point so the light seems too flat.
For the smooth lips, as i've already write to chrisdesign, it's been done on purpose, the chin is flattest, but i like it, I don't think that it's so lopsided.

 

[Tom Mann]

I always have fun comparing retouching methods, especially when the standard is so high, as it is in this case.

I'm always on the lookout for faster methods, perhaps even partially automated. The last time Chris posted one of his excellent retouch jobs, I had just received the newest version of Portrait Pro (ie, v.12), so I gave it a whirl, but, as I recall, I hardly supplemented it with any other methods -- in that case, what I posted was almost pure Portrait Pro.

This time, I'm more familiar with Portrait Pro and I also decided to supplement it with a few other semi-automated techniques / plugins. These included Topaz Detail (to partially replace the split spatial frequency approach), Nik Dynamic Skin softener, Nik Output sharpener, and "Perfectly Clear" (to quickly get a very bright look). All of these plugins were applied using layer masks and the BlendIF sliders to control their application.

I should also add that I have red hair [err ... had ... it's now white :sad: ] and many people in my family are very fair, pink-skinned, never-go-out-in-the-sun redheads, so I have some strong opinions on hair and skin color (as will be obvious when you compare the colors) :naughty: .

Anyway, enough blathering. Mine is on the left, FredricoF's s is on the right.

With respect to the time required to produce what you see (starting from the original), compared to a fully manual method, I probably didn't save much time by using lots of plugins. It took me around 40 minutes to get this far (lots left to do), plus another few minutes to put the two images side by side and post it.

T

 

[federicof]

Hi Tom,
i've read carefully your post,
i know very well all (or about all) plug-in that you quoted, i work with photoshop since 1998 and i've seen a lot of add-on appear (and disappear) in the filter panel!
I've watched your work and sincerely i found the quality of the skin texture very low, very plastic, maybe on the thumb is not fairly evident, but simply zooming on the image it becames obvious.
Last month i had the pleasure to work on a photo for a big exposition that had been printed 2 meters high, do you really think that your image will looks good on big (or normal) print?
I'd like to precise one thing: on the image i've posted the retouch is much exagerate, it's been a precise choise, a "style exercise" we say in italy (i don't know if translatable) but usually the retouches that i make for work are much natural than this one and the methods that i use (and that are used by a lot of photographers and visual artists) allow me to modify any single details, any different areas as much as i think, usually a very natural retouch should be invisible to the glance of the observer, plug-ins, in every case, make the photo fake.

Your critic is mainly about the time requested... the image took me about 3 hour...
with a retouch more natural i usually close a work in 1h30min-2h, your retouch took you about 40 min and as i can see there is still a lot of work to do (as you also tell), details, the hairs on the neck and other areas with visible imperfections, do you really think that your method allow you to save a lot of time?
Obviously these are only my opinions, is right that anyone has the own, but i'm not persuaded by your method and i'll continue with mine!
i hope i've been polite, i'm not english mother language and i've not a great control on the language!:banghead:
i've a suggestion: if you are a so big fan of plug-ins, you should try "portraiture" by Imagenomic i think it's absolutly one of the best.

 

[Paul]

 

[Tom Mann]

Hi Federico -

Actually, I agree with you on most of the points you made. In fact, even tho I stopped at 40 minutes, if I had worked on all the other problems, it could easily have taken the same amount of time as your method.

However, my idea is to periodically check on the progress being made by the developers of automated retouching software and see how close it's getting in quality to a fully manual approach, and see if it offers any advantage in speed even at reduced quality, say suitable for high volume portrait operations such as processing 1000 8x10 inch photographs of graduating high school students. The requirements for an each 8x10 inch print viewed only by a family are much more relaxed than the requirements for a print 2 meters tall.

So, I suspect there is less disagreement between us than you may think.

Best regards,

Tom

PS - by the way, both the before and after images that u posted were quite low resolution. Did you do your retouching for this example at such low resolution?

 

[Paul]

The thing with shop is we all think we are the bollox at it don't we, we love to show our efforts hoping for good feedback and dread the day we get the little yellow emoticon rolling on the floor at us

 

[Tom Mann]

Paul, I don't think you understand the purpose of this little experiment. It wasn't at all intended to be a "gee, look how good my method is" type of post. There are always tradeoffs between speed and quality. There aren't a lot of people (especially on this forum) with the capability of testing all these plugins, and I wanted to see just how bad the tradeoff actually is these days. I did this both for my own curiosity as well as to guide others.

Before the OP responded to my post, I stated my conclusion, "...compared to a fully manual method, I probably didn't save much time by using lots of plugins...", which I think is in essentially perfect agreement with the OP's conclusions. What I implied, but didn't explicitly say was that if one is willing to take a reduction in quality (as shown in the image I posted), then one might be able to save some time. It's up to each individual to determine whether their particular situation demands the highest quality, or is getting something of modest quality out in 40 minutes more valuable than a higher quality product in 3 hours.

T

 

[Paul]

Paul, I don't think you understand the purpose of this little experiment. It wasn't at all intended to be a "gee, look how good my method is" type of post. There are always tradeoffs between speed and quality. There aren't a lot of people (especially on this forum) with the capability of testing all these plugins, and I wanted to see just how bad the tradeoff actually is these days. I did this both for my own curiosity as well as to guide others.

Before the OP responded to my post, I stated my conclusion, "...compared to a fully manual method, I probably didn't save much time by using lots of plugins...", which I think is in essentially perfect agreement with the OP's conclusions. What I implied, but didn't explicitly say was that if one is willing to take a reduction in quality (as shown in the image I posted), then one might be able to save some time. It's up to each individual to determine whether their particular situation demands the highest quality, or is getting something of modest quality out in 40 minutes more valuable than a higher quality product in 3 hours.

T

Get your tin hat out Tom, I Understand fully it was not a dig at you or anyone else, we all work in different ways because shop and associated plug ins if you use them allow it to be that way.

 

[federicof]

yes Tom, i think that there isn't a great disagreement between us at the end,
every thing i've wrote should be refered to that cases we have the right reasons and requirements to do a work like this, usually i can use 2 hours/photo only for payed works that need that retouch level, a beauty editorial for example.
In other case, like the photos for the students that you write before is obvious that the solutions should be others, otherwise a work of 1000 photos became a year of work!


Personally, however, I retouch much the skin only for portraits (when request) beauty and fashion images (or for personal works and experiments like this) in other kind of photos i concentrate only on other aspects, like atmosphere, sharpening, colors, etc.


the image i work on i think it was a photo from Nikon D800, about 7000x5000px.

 

[Tom Mann]

OP: "...Personally, however, I retouch much the skin only for portraits (when request) beauty and fashion images (or for personal works and experiments like this) in other kind of photos i concentrate only on other aspects, like atmosphere, sharpening, colors, etc...."

Like you, I also do this much skin retouching only if specifically requested. In fact, sometimes, if you retouch the skin of an older woman without permission ... even if only a little bit ... they will be very upset. Some feel that they have earned every wrinkle. Others want a little bit of softening, but not too much. They are looking for a "believable" final image.

BTW, since I was working on a jpg version of the image that was only about 1300 pixels high, whereas you were working on an image that was either 5000 or 7000 pixels high (depending on which way the camera was oriented), this factor of at least 15x in area could easily have been part of the reason that the small scale structure of the skin was better preserved in your version. Also, were you working on a NEF or a 16 bit Tiff, not just a JPG? I'm not trying to "beat a dead horse", but I do like to get to the bottom of things.

For the sake of another experiment, is there any chance you might be willing to let me try the plugin method on the full resolution version of the same image? I wouldn't spend the time trying to deal with all the small retouching details scattered throughout the image, just a small patch of skin. I certainly understand if you would only feel comfortable sending a watermarked version, or a narrow vertical strip cut out of the original, but I think this would be a very interesting additional experiment. I don't want to be pushy, but might you be comfortable doing this?

Tom

 

[federicof]

obviously working on jpg rather than RAW the result will be bad than the other, but i think that is the same because however plugins and filters destroy pixels, with a biggest image... you obtain to destroy much more pixels!
However, i apologize, i can't send you the NEF, i'm the retoucher only, i've the permission to share my work and the before/after reusult, but i'm not allowed to share the original shoot.

 

[Tom Mann]

I understand completely, but just in case the comparison was possible, I wanted to ask.

Different algorithms respond differently to structures of different sizes. Some preserve smaller features better than others. In my experience doing and teaching image processing since the 1980s and PS since the 1990s, I've learned that with complicated algorithms / processes such as we are discussing, the only way to tell how a particular algorithm (for example, Portrait Pro) will respond to increased resolution is to do a proper, well controlled test. My guess is the exact opposite of yours, but without an actual comparison of the same image at two resolutions, we are both just guessing.

A Sincere Thanks Again,

Tom

 

[federicof]

Hi Tom,
yes, we're pratically players in two different team, contrast for the different point of view but with fair play, i like it!
In the near future i'll be very busy for work, but as soon as will be possible, when I will have the time and a new RAW file free to work on, we will share it to compare methods, battle on equal footing, ok?


best regards

 

[Tom Mann]

I love it! That sounds great!

more later,

Tom

 

[ALB68]

I don't profess to be in the same league with you pros, but I still like attempting these. Tried to maintain a more real look.