SRB Photo wrote:
And as a note for the MM crowd, sparked by an incessant conversation elsewhere, it occurred to me that yet another way of halving image contrast prior to employing the native High Pass filter is to use the Custom filter with a matrix of:

0 0 0 0 0
0 0 0 0 0
0 0 1 0 0
0 0 0 0 0
0 0 0 0 0

Scale 2 Offset 0

The advantage is that this is an extremely efficient way of doing things (or at least, should be - I won't go out on a limb and vouch that a filter this old is actually optimized) and is another 100% Smart Object-compatible solution.

So how does this fit in the current workflow?

Julian Marsalis wrote:

So how does this fit in the current workflow?

That depends on what you're trying to do.  Mostly it's for the Smart Object crowd, though I'm sure some enterprising folks will find more uses for it.  If you're doing "splits" so as to heal / clone onto the separated layers, this doesn't have a lot of value.  I'm told that it's also easy to script with, though as I'm not much of a script writer myself, and the person who told me that hasn't actually implemented it, I can't verify that fact.

SRB Photo wrote:

That depends on what you're trying to do.  Mostly it's for the Smart Object crowd, though I'm sure some enterprising folks will find more uses for it.  If you're doing "splits" so as to heal / clone onto the separated layers, this doesn't have a lot of value.  I'm told that it's also easy to script with, though as I'm not much of a script writer myself, and the person who told me that hasn't actually implemented it, I can't verify that fact.

It is relatively easy. Using the image below as the example...



This is how you would script it..

var set=new Array();
var layerRef=app.activeDocument.activeLayer;
set=[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24];
scale=100;
offset=50;

layerRef.applyCustomFilter(set,scale,offset);

Wherever you place a "0" into the matrix, it will come back as a blank box for that position if you reopen the filter (Smart Object Layer). I recommend keeping the 0's in the array to make it easier to keep track, though you could just leave out that array element. It also makes it easier to do various calculations on them if you should want to. It's easier to add 2 to 0 than it is to " ".

Photons 2 Pixels Images wrote:
It is relatively easy. Using the image below as the example...



This is how you would script it..

var set=new Array();
var layerRef=app.activeDocument.activeLayer;
set=[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24];
scale=100;
offset=50;

layerRef.applyCustomFilter(set,scale,offset);

Wow.

That does look easy.

Maybe even easy enough that a caveman like me could do it.  I'll look into it.

Thank you!!

SRB Photo wrote:

Wow.

That does look easy.

Maybe even easy enough that a caveman like me could do it.  I'll look into it.

Thank you!!

I know this is a bit off topic, but for a bit of randomness....

var set=new Array();
layerRef=app.activeDocument.activeLayer;
scale=Math.round(Math.random()*10);
offset=Math.round(Math.random()*100-50);
for(i=0;i

SRB Photo wrote:

That depends on what you're trying to do.  Mostly it's for the Smart Object crowd, though I'm sure some enterprising folks will find more uses for it.  If you're doing "splits" so as to heal / clone onto the separated layers, this doesn't have a lot of value.  I'm told that it's also easy to script with, though as I'm not much of a script writer myself, and the person who told me that hasn't actually implemented it, I can't verify that fact.

Thanks

Liddell Photography wrote:
I have a few questions:

1) How do you gauge how much blur to use for the separation. I've tried blurring a lot and blurring a little. I know that the separation looks the same either way but how does it effect retouching the image once its separated? I read something about you blur only what you want to keep in the end. ???

It's a visual thing. The details that get lost during the blur will be applied to the HF layer after the separation process. So yes, whatever you blur out will be kept on the HF layer. It takes a bit of practice. Keep in mind that there will always be some color information in the HF layer and there will always be some detail in the LF layer. It's the nature of the beast.

Liddell Photography wrote:
2) I find that sometimes I have to retouch the same things twice. Once for the texture (high frequency) and twice for the color (low frequency). Is there anyway to do both at the same time? The answer to #1 may answer #2.

Not to my knowledge. And mostly because the two layers have different purposes. HF for working on details and not bothering the underlying tones. LF for working on the tones without affecting the details. If you want to work on both at the same time, it would be the same as using the healing brush/clone stamp/etc. on a non-separated layer. I believe the point to separation is so they can be worked on individually.

Liddell Photography wrote:
3) I see posts and discussions about using a multi-frequency separation. I've downloaded a few actions but am not sure what's what. What is the benefit of a multi-frequency separation and how do you retouch with it?

I've personally rarely used multiple separations on a single image. It can be useful if there are different areas with different sized "details" that you wish to separate out to make it easier to work on. Example: Skin pores would be a relatively high spatial frequency (small detail) and a pattern on a sweater would be somewhat lower (bigger detail) and then the underlying tones would of course be on the LF layer. If you try to separate based on the skin pores only, then try to work on the sweater pattern, you'd have too much of the detail left on the LF layer which would make it more difficult. If you try to separate based on the sweater pattern, you'd have too much of the tones left around the skin pores and again, it would make the job more difficult.

Liddell Photography wrote:
4) I would love to hear what other's workflows are like when using separations, especially non-destructive ones.

This question could be a whole thread by itself. In fact, it has been a few times.

I honestly don't do much fancy with my separations. Place a blank layer between them to paint tones so I don't paint on the LF layer itself.

SRB Photo wrote:
Below is a very, very early version of a Skin Smoothing / IHP / FS engine.  I'm not going to take the time to document it too thoroughly just yet as a lot is likely to change - this is the product of a few hours work the last couple of nights and still requires polish.  Just the same, it is operational and as some folks are using it, I did want to get a few updates out into the 'wild' before anyone started saving files which might later be made incompatible.

In short, this is a script, so download / save it to the scripts directory you have configured in PS (typically, but not always [Photoshop Root Directory]/Presets/Scripts), restart Photoshop, and then select it from File->Scripts->openSkin.  Subsequent versions will come with a launch action which you can install and configure to a keypress so that you can use it from your keyboard, programmable mouse, Wacom, etc. without needing to use the keyboard at all.  For now, though, you either need to roll your own or do it manually.  Not having it now is the price of being an early adopter .

Usage is pretty straightforward.  You can use this on an 8bit or 16bit image and either will give you the bit-depth respective "Apply Image" / "HPS" quality results.  When the dialog presents itself, it will ask you for the radii which you want to use for GB and for HP.  You can choose to Invert the result if you'd like to remove that frequency band from the image, or by default it will emphasize the selected range.

If you start the script with a selection active, you get the option to use that selection as the basis for your new layer in much the way that Portraiture avoids 'color bleeding' in the skin areas.

If you have an openSkin layer selected when you run the script, it will allow you to:

  1.) Update that layer's current values with the new ones you select.
  2.) Create a new layer based on the same SO as the current, using whichever settings are entered into the dialog [YES, this means you can use the same SO throughout a workflow, make changes to it later, and have the same corrected IHP / FS adjustments applied to it automatically -- at least so long as the rest of your workflow is non-destructive].
  3.) Create a new layer based on the image as it currently appears (default).

Known issues:
  - The UI is ugly as sin.  I'll work on that, but I'm ultimately going to keep this JS-only so as to retain backwards compatibility as far back as we can go with it.  I'm not a Jobsian Flash hater, but I've seen how many people are being excluded from using other products because they can't afford to keep up Adobe's upgrade pace.
  - The code is ugly.  I'm in a critical-bugs-only development freeze while I think through a partial rewrite to streamline things a bit and hopefully to bring more functionality into this.
  - There is no option to automatically add a mask to the generated layer(s).  This will be forthcoming.
  - It hasn't been tested on anything except Windows Vista + CS5.  I'm sure that there are going to be version-based issues to work through.

Future additions to look forward to:
  - Better, more stable codebase - that usually makes everyone happy
  - Better selection integration
  - Saved 'preferences' between uses
  - Slightly improved UI
  - Automatic mask addition
  - "Target Size" + "Range" dialog option for removing frequency ranges without having to deal with radii
  - Whatever else you ask for (within reason)

Please let me know if you have problems with it.  As I said, this is very, very, very early in the development process, so please don't make up your mind about it forever right now.  I'd love to hear your feedback in all cases, though.

With thanks to Photons2Pixels and the folks over at ps-scripts.com for all the code they give the world, I present: openSkin!

openSkin FAQ:

  1.) Why are you writing this?

     - A community desire for such a tool has been clearly demonstrated by similar PS extensions recently, and not everyone can use the Flash panel engine, so this provides a Javascript-only solution to that problem.

  2.) When do you expect to complete the script?

     - It will likely never be 'done' in the traditional sense, as I hope to continue updating it in accordance with feature and functionality requests from the community.  If you have any, please contact me!  That said, I do expect to have a '1.0' out the door by the end of November 2010.

  3.) What makes this different from similar products?

     - Primarily the architecture - this is a Javascript-only application vs. being a Flash-dependent hybrid.  I didn't honestly appreciate how slow the community-wide upgrade process is until I was watching the feedback with how many people were finding themselves unable to use the Flash solutions.  Other differences include:
           - Predictable accuracy (more discussion after 1.0)
           - Ability to use the same Smart Object repeatedly
           - Lack of a real-time option in openSkin (coming after 1.0)
           - Ability to view the full history of what's been done - openSkin is not a 'black box'
           - Responsiveness to community requests
           - Open Source - see everything which goes on behind the scenes; make your own variations; etc.

  4.) Why the name change?

     - openSkin was always the intended name for this application, but development went a lot faster than I thought it would and I ended up releasing a couple of personal test builds to the public which were based around my initials.  Normally I do that with actions and scripts I release since so many actions have such similar names, but with openSkin I wanted to be able to hand off the program to another developer if the time and attention I can give it were to become constrained.  In that situation I wouldn't want the new developer feeling slighted that they were distributing a product with my name on it while they were the one putting in all the work.  And I use my initials because my name is Sean Robert Baker and I stand behind what I write .

  5.) More to come...

More to come please lol waiting lol .......

SRB Photo wrote:
Below is a very, very early version of a Skin Smoothing / IHP / FS engine.  I'm not going to take the time to document it too thoroughly just yet as a lot is likely to change - this is the product of a few hours work the last couple of nights and still requires polish.  Just the same, it is operational and as some folks are using it, I did want to get a few updates out into the 'wild' before anyone started saving files which might later be made incompatible.

In short, this is a script, so download / save it to the scripts directory you have configured in PS (typically, but not always [Photoshop Root Directory]/Presets/Scripts), restart Photoshop, and then select it from File->Scripts->openSkin.  Subsequent versions will come with a launch action which you can install and configure to a keypress so that you can use it from your keyboard, programmable mouse, Wacom, etc. without needing to use the keyboard at all.  For now, though, you either need to roll your own or do it manually.  Not having it now is the price of being an early adopter .

Usage is pretty straightforward.  You can use this on an 8bit or 16bit image and either will give you the bit-depth respective "Apply Image" / "HPS" quality results.  When the dialog presents itself, it will ask you for the radii which you want to use for GB and for HP.  You can choose to Invert the result if you'd like to remove that frequency band from the image, or by default it will emphasize the selected range.

If you start the script with a selection active, you get the option to use that selection as the basis for your new layer in much the way that Portraiture avoids 'color bleeding' in the skin areas.

If you have an openSkin layer selected when you run the script, it will allow you to:

  1.) Update that layer's current values with the new ones you select.
  2.) Create a new layer based on the same SO as the current, using whichever settings are entered into the dialog [YES, this means you can use the same SO throughout a workflow, make changes to it later, and have the same corrected IHP / FS adjustments applied to it automatically -- at least so long as the rest of your workflow is non-destructive].
  3.) Create a new layer based on the image as it currently appears (default).

Known issues:
  - The UI is ugly as sin.  I'll work on that, but I'm ultimately going to keep this JS-only so as to retain backwards compatibility as far back as we can go with it.  I'm not a Jobsian Flash hater, but I've seen how many people are being excluded from using other products because they can't afford to keep up Adobe's upgrade pace.
  - The code is ugly.  I'm in a critical-bugs-only development freeze while I think through a partial rewrite to streamline things a bit and hopefully to bring more functionality into this.
  - There is no option to automatically add a mask to the generated layer(s).  This will be forthcoming.
  - It hasn't been tested on anything except Windows Vista + CS5.  I'm sure that there are going to be version-based issues to work through.

Future additions to look forward to:
  - Better, more stable codebase - that usually makes everyone happy
  - Better selection integration
  - Saved 'preferences' between uses
  - Slightly improved UI
  - Automatic mask addition
  - "Target Size" + "Range" dialog option for removing frequency ranges without having to deal with radii
  - Whatever else you ask for (within reason)

Please let me know if you have problems with it.  As I said, this is very, very, very early in the development process, so please don't make up your mind about it forever right now.  I'd love to hear your feedback in all cases, though.

With thanks to Photons2Pixels and the folks over at ps-scripts.com for all the code they give the world, I present: openSkin!

openSkin FAQ:

  1.) Why are you writing this?

     - A community desire for such a tool has been clearly demonstrated by similar PS extensions recently, and not everyone can use the Flash panel engine, so this provides a Javascript-only solution to that problem.

  2.) When do you expect to complete the script?

     - It will likely never be 'done' in the traditional sense, as I hope to continue updating it in accordance with feature and functionality requests from the community.  If you have any, please contact me!  That said, I do expect to have a '1.0' out the door by the end of November 2010.

  3.) What makes this different from similar products?

     - Primarily the architecture - this is a Javascript-only application vs. being a Flash-dependent hybrid.  I didn't honestly appreciate how slow the community-wide upgrade process is until I was watching the feedback with how many people were finding themselves unable to use the Flash solutions.  Other differences include:
           - Predictable accuracy (more discussion after 1.0)
           - Ability to use the same Smart Object repeatedly
           - Lack of a real-time option in openSkin (coming after 1.0)
           - Ability to view the full history of what's been done - openSkin is not a 'black box'
           - Responsiveness to community requests
           - Open Source - see everything which goes on behind the scenes; make your own variations; etc.

  4.) Why the name change?

     - openSkin was always the intended name for this application, but development went a lot faster than I thought it would and I ended up releasing a couple of personal test builds to the public which were based around my initials.  Normally I do that with actions and scripts I release since so many actions have such similar names, but with openSkin I wanted to be able to hand off the program to another developer if the time and attention I can give it were to become constrained.  In that situation I wouldn't want the new developer feeling slighted that they were distributing a product with my name on it while they were the one putting in all the work.  And I use my initials because my name is Sean Robert Baker and I stand behind what I write .

  5.) More to come...

Thank you Sean!  Cool script.

I ran it on a finished image and clicked off the top custom layer, ran the smart object layer opacity down to 25% and WOW! (see image below)  I ran it on another image and did the same, but this time I ran the smart object opacity down to 0 and then started up until I got a really cool look.  Obviously, you can set the Radius and Blur to your liking for the particular image.

Can't wait to use and see more.

Julian Marsalis wrote:
More to come please lol waiting lol .......

I'll get back to it someday, but my day job has taken off lately and I've had little time to put towards it (read: none).  This is, of course, why I open-source'd it in the first place .  Anyone wishing to take up the mantle is free to, as I clearly won't have a lot of time in the near future.

SRB Photo wrote:

I'll get back to it someday, but my day job has taken off lately and I've had little time to put towards it (read: none).  This is, of course, why I open-source'd it in the first place .  Anyone wishing to take up the mantle is free to, as I clearly won't have a lot of time in the near future.

Cool cool I feel ya and thanks for what you have done and maybe one day I will get off my ass and get back to scripting just programming all day then at night just leads to burnout tho lol.

Julian Marsalis wrote:
A very interesting read in Russian thank goodness for google translator:

http://zhur74.livejournal.com/998.html

Nice find
Thanks

Julian Marsalis wrote:
A very interesting read in Russian thank goodness for google translator:

http://zhur74.livejournal.com/998.html

Neat article, but for someone capable if citing Fourier, I would have expected a more robust conclusion about why the HP filter can give poor results (especially as he also cites the recommendation to halve contrast). Still, an apparently much more intuitive explanation than that which I originally offered.

SRB Retouch wrote:
Neat article, but for someone capable if citing Fourier, I would have expected a more robust conclusion about why the HP filter can give poor results (especially as he also cites the recommendation to halve contrast). Still, an apparently much more intuitive explanation than that which I originally offered.

Sean it's only because you speak in fluent super geek, but photon's always comes to the rescue and translates.

Seriously you have provided and shared so much useful information in this thread, that has influenced and affected so many peoples work-flows.

Opening up the numerous permutations and varied applications that split frequency allows for in practical uses.

Least not forget all the work that has gone into scripts and actions, that you, photon's and others have made available here.

Julian Marsalis wrote:
http://www.screencast.com/users/Shift_S … 352a580feb

You can do the same thing with skin FWIW - and may get better results using something other than GB as the base for the separation, but that depends on the source material (and on whether you preselect + move to a new layer).

Sean Baker Photo wrote:
Wow.

That does look easy.

Maybe even easy enough that a caveman like me could do it.  I'll look into it.

Thank you!!

When I see MM I'm starting to think "Matrix Mayhem".

rdallasPhotography wrote:

When I see MM I'm starting to think "Matrix Mayhem".

Now this was good.

Bo Brinch wrote:
Hi Sean
Am I expecting too much from your technique or does it look like I've done something wrong?

Or am I just looking at it in the wrong way?

Best regards
Bo

little edit: read your post wrong.

The question was not aimed at me, but I felt like responding.

The frequency split is just that. A splitting of your image in 2 frequency bands.
A high and a low.
The high contains most of the detail(and a little color)
While the low contains most of the color and very little detail.

The fastest way to see result is to blur the low frequency layer and see what happens.

As you can read in this thread there are many ways you could take advantage of this splitting technique.

e.g. Heal/Clone on the high frequency layer to even out texture.
or
Paint/Color on the low layer(or a new layer between the High and the Low layer) to even out tones.
etc.

Lanenga wrote:

The question was not aimed at me, but it looks like you removed the low frequency layer.

The frequency split is just that. A splitting of your image in 2 frequency bands.
A high and a low.
The high contains most of the detail(and a little color)
While the low contains most of the color and very little detail.

The fastest way to see result is to blur the low frequency layer and see what happens.

As you can read in this thread there are many ways you could take advantage of this splitting technique.

i.e. Heal/Clone on the high frequency layer to even out texture.
or
Paint/Color on the low(or a new layer between the High and the Low layer) layer to even out tones.
etc.

Very good explanation, Lanenga.

You should right a book called Split Frequency for Dummies.
You would make a fortune.

P.S.
It wasn't aimed any anyone personally.

Bo Brinch wrote:
As far as I've understood your technique is supposed to give both sharper images and less artefacts but I find that there's only less artefacts.

Am I expecting too much from your technique or does it look like I've done something wrong?

Or am I just looking at it in the wrong way?

I would assess things as being some of all of the above, but in the order you've asked...

1.  The technique isn't itself designed to necessarily give sharper images directly, though it's designed to do everything with less artifact.  And, as there's less artifact, you can (at least theoretically) sharpen more with fewer problems.  Primarily, though, the technique's advantages lie in the amount of control you gain.  As Lanenga points out, the ability to heal or to clone on the different frequency bands, to apply curves to one or the other independently (steep 'S' curves on the HF will give you a kind of sharpening which few are accustomed to controlling), etc.  Which leads into #2...

2.  You're not expecting too much, just looking at it...

3.  ...from a slightly different perspective than that which post people using it daily have.  It's a complex subject and one not easily grasped - e.g. none of us knew what exactly was the issue with the HP filter when this thread started - don't feel you need to understand it all in one read .

Krunoslav-Stifter wrote:

Very good explanation, Lanenga.

You should right a book called Split Frequency for Dummies.
You would make a fortune.

P.S.
It wasn't aimed any anyone personally.

I'll need the Cliff's Notes for the Dummies' book.

Robert Randall wrote:

When you've hung out in here long enough, you come to realize there are a large number of people that can't accomplish much in life. They need some form of reinforcement in their lives, and they seem to have chosen the art card for help with that. They can't make art, but they can hide behind the art card. Their arguments use the ploy of deflection to acheive credibility, much the same that OJ did during his murder trial. Everyone knows they are talentless hacks, but when they produce a snap shot of their girlfriend tugging at her soiled panties while sticking her ass out and giving the cell phone that come hither look that reminds you of a deer about to meet it's maker, they argue the picture is their art, and that somhow validates everything, and they are no longer a talentless hack screwing a toothless hag, they am be ARTEESTS!

Of course in this particular case, I'm sure the respondent is a highly talented artist that has a firm belief in the soul of his work. 

Or not.

All to true Rob. That is another reason you left, and rarely visit. Also why I am teetering on the notion. I too not needing to think with my crotch are looking for greener pastures.

Photons 2 Pixels Images wrote:

I'll need the Cliff's Notes for the Dummies' book.

Yea, well now that my photoshop is corrupted with the bad actions posted here and other forums I am at a crossroads.