Elite Retouch wrote:
Yeah. I think the next major thing Adobe should implement into PS is the ability to have TOTAL CONTROL of that actions pallet. It's so constricting when it comes to changing SLIGHT mistakes.

I've been requesting that since version 7. (I skipped v6, and while I loved Actions--still do--the particular implementation is irksome. Modularity makes actions very fragile; the lack of control over the palettes, etc.)

Tell Adobe. Formally. If they get more complaints requests, they're more likely to change it.

(Oops. /hijack.)

Getzphoto wrote:
I have a question for Sean regarding this thread. I am trying to accomplish a certain look and am missing one piece of the puzzle and thought you may be able to help. I have been trying to work out in addition to many other techniques using Luminosity Masks, calculations in many different color spaces to create a digital zone system with one image. It is very long winded, but have hit a wall with the final piece of the puzzle. I am trying to get a heighten level of contrast and sharpness and thought that it maybe accomplished through Low frequency & High frequency techniques but wanted your take on this. I was wondering if you knew how Robert Randall was getting the ultra sharpness and contrast. I believe he originally was involved in this thread or a similar thread, but he has not participated in things in a while.

address to website: http://www.robert-randall.com

He also wrote way back about local curves adjustments and then making Highlight masks, Shadow masks, and then a difference mask and putting curves on them. I was wondering if you or anyone knew how the highlight , shadow & difference masks were made. I make masks using Luminosity techniques of intersecting channels, adding  & subtracting of various channel selections. By combining my seamless masks I will be able to totally control sharpness and contrast of various areas while providing detail & contrast without blowing out highlights or blocking up shadows.
I am not trying to copy another persons style at all and I only singled him out as he was involved in a similar thread to this. There are a ton of photographers that are doing this and my style is every different. I Want to incorporate this ultra sharp/contrast as a small part of a broader scheme. Outside of High Pass is there another way and if so can you explain and if High pass what would be the best approach.

Unless a mod comes in to rule otherwise, neither of us can publicly discuss Bob's techniques beyond what he previously stated here in the forums (site rules).  I will try to offer a few clarifications to the best of my ability to do so, though:

  - Bob has stated that he has been using the sharpening techniques discussed in this thread - particularly 'HP' separations with a curves layer clipped to the 'HF' data, directly increasing the contrast of the higher frequency components.

  - The highlight, shadow, and difference masks which Bob referred to are very similar to the luminosity masks which you already mentioned.  In the examples he gave he was using the black channel out of a CMYK copy of the original image to create those masks.  It gives a bit different result than a luminosity mask, sometimes preferable, sometimes not, but in all cases an analogous process to what you already know how to do (you could just as easily do the whole thing with the red channel, yellow channel, etc).

HTH.

Alstare Photography wrote:
Out of curiosity...

I been following this thread since page 1. And I am using the original action that just separates the frequencies for sharpening. the re-loading the high pass to my original 16bit image for the effect on linear light mode.

Is there an action version that does just that?  I know one was posted around page 4-ish i believe but it didnt seem to work.

Is not a big deal having to copy the layer back over to the original image,  but an all in one step action would be nice, without the extra d&b layers, etc...!

Thanks!

The issue comes from how many layers currently exist, their names, what layer(s) are selected when the action is run, etc.  It'd be possible to write a script to do all that, though sadly that's not really my forte .  The easiest thing to do (IMO) would be to decide how you want to work through your workflow, and wherever in the process you want to do your separations, record your own action to create two snapshots of your image, perform the steps written out in the OP, and then delete the LF data before ending.  That way it's tuned to the way that you work and won't fall apart because everyone handles their layers differently.

Alstare Photography wrote:
Out of curiosity...

I been following this thread since page 1. And I am using the original action that just separates the frequencies for sharpening. the re-loading the high pass to my original 16bit image for the effect on linear light mode.

Is there an action version that does just that?  I know one was posted around page 4-ish i believe but it didnt seem to work.

Is not a big deal having to copy the layer back over to the original image,  but an all in one step action would be nice, without the extra d&b layers, etc...!

Thanks!

I could come up with something for you but I'm not quite sure what your final setup looks like. Do you want to be left with the original background layer and the HF layer on top of that with nothing else?

Alstare Photography wrote:
Thanks for the responses...

I should of gave more detail, I am running the separation at the end of my work-flow.  So I will be starting with nothing but the original "Background" Layer in my stack. 

Currently I am just copying the the top freq layer from the new image the script creates then pasting it back on my original document with the Linear Light blending mode. 

I tried to add to the action to do what I needed but am not a technical master when it comes to making actions.  Ideally I would like to end up with a Background layer and the HF layer in the correct mode, in my final stack.

No problem.  I'm guessing 16 bit mode? I'll get one for you when I get home from work if I remember. I'm forgetful. remind me and I'll definitely get to it.
I forgot to ask if you want a clipped curves on the HF layer.

Alstare Photography wrote:
Thanks for the responses...

I should of gave more detail, I am running the separation at the end of my work-flow.  So I will be starting with nothing but the original "Background" Layer in my stack. 

Currently I am just copying the the top freq layer from the new image the script creates then pasting it back on my original document with the Linear Light blending mode. 

I tried to add to the action to do what I needed but am not a technical master when it comes to making actions.  Ideally I would like to end up with a Background layer and the HF layer in the correct mode, in my final stack.

OK. Uploaded.

http://www.nunuvyer.biz/Photoshop/Separation/

Click on the "Frequency" link to download the action file. The name of the action you want is HF Only 16bit. I did leave the clipped curves layer in, but all you're left with is the Background and High Frequency (with clipped curves).

Getzphoto wrote:
Thanks Sean for clearing some things up. Do you feel that using apply image techniques that you describe previously in this thread to be better than HP for your HF layer? I find that HP can cause haloing as most sharpening techniques do if not done right. I do know that using a midtone contrast layer through blending modes works well when you restrict the blend in the layer style dialogue box. I was just trying to find the best way to to add contrast /ultra sharp without haloing while still keeping the image some what natural.

There's no question that the Apply Image technique is technically superior than either the HP directly or even it's 1/2 contrast'd workaround - it's truer to the data.  For 99.9% of usages, though, either workaround technique is sufficient and avoids the filter's problematic issues.

Haloing can be just as much an issue with or without the workarounds and has more to do with radii selected and how they interact with the original image than with the particular implementation used.

HTH.

DerW wrote:
I'm just having a discussion with some folks in Germany about this topic so I rethought about it once again :-)

One thing that I still don't really understand is why this technique should be superior in sharpening images towards a standard unsharp mask.
We know, that the blending mode "Linear Light" works C=B+2*A-1 (otherwise we wouldn't have to use the scale of 2 and could relay on the standard highpass).
So I tested some things.

I first duplicated my background layer two times, applied a Highpass of 5Px to the bottom one and set its blending mode to "Linear Light".
Then I selected the top layer and ran USM with a strength of 200% (remember: 2*A) and the same radius as in the Highpass. Switching the blending mode to "Difference" I saw none (although I'd have to test it some more).
And now... we're just using a Highpass that is scaled by two, so the ultimate blending would be C=B+2*(A/2)-1 which equals C=B+A-1 and that's exactly what USM does if I'm not mislead (of course I did some tests and there actually is a difference but I'm not sure whether that is because of the remaining inaccuracy).

So maybe I'm totally mislead on this, if so, please correct me :-)

(Corrected - vs the filter) Highpass sharpening isn't superior in itself.  What it offers (and which some quite enjoy) is a greater degree of control over the result than is achievable by the USM filter.  That is, by pinning a curve or level to the HP data, one can amplify the highlights, shadows, or both selectively to accent those aspects of the image, imparting a greater perceptual sharpness in either range.

But you're quite right that a default HP sharpen & a USM should look identical - they are .

nathan touchup wrote:
lol never thought I see the day you would say High Pass sucked

Why?

Elite Retouch wrote:
I have a new, refined version of my 3 part action.

Link? And Thanks

DerW wrote:
I just can't get enough of this thread :-)

So once again, I thought about it and came up with a way (which still needs improvement!) to separate the image into frequencies and intermediate frequencies.
Some people said, they'd use multiple frequencies, every time building up new ones after flattening the old ones, so I thought that could as well be done more easily :-)

First things first, start with the separation as described in the first post of this wonderful thread.
After that, duplicate the HIGH layer and blur it by a radius of x pixels. Then go to the duplicate of the HIGH layer and choose "Filter"-"Other"-"High Pass" and enter the exact same amount of pixels. Put this layer on "Linear Light" mode and again, this is your original image, just with some intermediate frequencies.

You could as well use the "Apply Image" command again, but this time use the appropriate blending mode for the bit depth but enter a scale of 1 and an offset of 128 in 8Bit or -128 in 16Bit (although as far as I know that's what high pass does anyway).


However, this seems to be quite accurate, but the shadows and highlights again are clipped (although by far not as badly as they would in the initial separation with the normal high pass filter).

And that's exactly what I couldn't manage yet, how to prevent the clipping.
My first thought was to use the scale of 2 again, but this seems not to give the desired results, the "higher" high layer doesn't contain the necessary details anymore (the result is way to soft).

Just curious because I'm pretty sure it's easily solved, but I just can't wrap my head around how at the moment...

Scripts to do this: here and here.

^^ Should I link those from the OP - is there enough interest?

DerW wrote:
Stupid me!
Thanks again, Sean!

Not at all. You're innovating and sharing - no one can ask for more. I just need to update the OP with more of the thread resources.

SRB Photo wrote:

Scripts to do this: here and here.

^^ Should I link those from the OP - is there enough interest?

Sean, if you want I can make a repository where anyone can upload their own actions/scripts and even examples to share. It would be much easier than going through the thread and updating the OP each time. It would be simple to set up.

Photons 2 Pixels Images wrote:

Sean, if you want I can make a repository where anyone can upload their own actions/scripts and even examples to share. It would be much easier than going through the thread and updating the OP each time. It would be simple to set up.

That would be fantastic, though I would prefer it as an adjunct to posts in the thread vs. a replacement - unless there would be mechanism for discussion and announcement therein?

SRB Photo wrote:

That would be fantastic, though I would prefer it as an adjunct to posts in the thread vs. a replacement - unless there would be mechanism for discussion and announcement therein?

I was referring to the former. It would mainly be a place for people here to upload files (actions, scripts, examples) so others can access them. They would all be in one central location so people can simply go there to get what they want and see if there are any updates. They can link to examples within their posts, also.

I have a simple php script to do this now, but I'll look around a bit more to see if I can find something with more features.

It'll give me something to do the rest of the day.

Photons 2 Pixels Images wrote:

I was referring to the former. It would mainly be a place for people here to upload files (actions, scripts, examples) so others can access them. They would all be in one central location so people can simply go there to get what they want and see if there are any updates. They can link to examples within their posts, also.

I have a simple php script to do this now, but I'll look around a bit more to see if I can find something with more features.

It'll give me something to do the rest of the day.

I'm a fan of the idea, and like the way you think sir!

Now if someone would just tell me more about this 'banding' thing, I'd have something to work on the rest of the day .

SRB Photo wrote:

I'm a fan of the idea, and like the way you think sir!

Now if someone would just tell me more about this 'banding' thing, I'd have something to work on the rest of the day .

I'm not happy. It appears we'll have to go with simple for now. I've downloaded 2 repository scripts so far and neither works as advertised.

I'll get it set up and let you know.

Photons 2 Pixels Images wrote:
I'm not happy. It appears we'll have to go with simple for now. I've downloaded 2 repository scripts so far and neither works as advertised.

I'll get it set up and let you know.

OK. I got it set up. It's very simple for now. Hopefully I'll find something that will work better or I'll have to dig in and recode this one. It's been awhile since I've done PHP.

http://www.model-citizens.com/FreqSep/

Elite Retouch wrote:
I have a new, refined version of my 3 part action.

And???? I haven't played with it in a bit....

This is the best thread on the site, IMO...

E Martin Photo wrote:
Very cool idea.  Thank you for sharing your technique!

To the naysayers, I fail to understand the idea that post processing is somehow degrading the artistic value of an image.  Have you never shot film and done your own development?

Lights, props, lenses, and cameras are just tools for creating an image.  Photoshop and other processing software is no different.  It should be viewed as a tool just as essential as your camera for creating your final image.  Just my $0.02 for what it's worth.

We're not dealing with film anymore, digital is a whole different animal...

DarkSlide wrote:

You mean I spent all those years in a darkroom making certain the D-76, Acufine, and Rodinal were at the precise temperature, making certain the development times were spot on, the agitation just so, the dryer at the correct temperature all for nothing?  I was messing with the art- huh? - by practicing precision and by testing my Tri-X in various chemistries until I found which worked best when I needed to pull or push my film speed.

What about the print process?  Paper choices; tone and amount of silver, developer; secret brew or Dektol? toning chemicals? times? blah blah blah -- not artistic concerns cause they deal with chemistry and math?

And what about all that f-stop, shutter speed, iso, dof, lens angle, blah blah blah -- not artistic cause it involves math?

I presume you use a computer in your camera and on your desktop to acquire and make your images publishable on this site. Yes?  That's all math baby.  Nothing but 1s and zeros dancing to the programmers music.

Damn good point. Excellent I should say.

SRB Photo wrote:
To use this technique for HP-type sharpening, copy the top layer generated by the action back to the top of your original document.  The effect should be immediately apparent.  And as suggested to others, you can apply a clipped curves layer to it to heighten the effect, run SS on it, or anything else which enters your mind.  The uses for the separation are only limited by what we choose to create with it.

aaaaaaaaaargh! I had missed that.

Love it! Thanx.

ACPhotography wrote:
This is the best thread on the site, IMO...

+1 Love it!

SRB Photo wrote:
Speaking of this, I hope to have a new video for y'all tomorrow or Sunday.

Awesome!

I gotta say, Sean, you've peeked my interest in post processing with Photoshop more than anyone else has and I've been pushed into using my atrophied brain cells once again. For that I thank you, sir.

SRB Photo wrote:
Speaking of this, I hope to have a new video for y'all tomorrow or Sunday.

what's your method for tutorial capture, by the by?