"asymmetric" frequency separation?

[drode]

While I'm no expert on the technique, I use frequency separation regularly bu I have no clue what asymmetric frequency separation is or how I might make use of it.

Could anyone explain it to me or point me to a reference? Google and a local search here didn't turn up much that helped.

Thanks in advance,

Dan

[KTG]

There was a post about this a day or two ago. MisterMonday had a very good explanation and instructions on how to set it up. Look back in the posts for this week.
k

[KTG]

The tread was Smudging when using Frequency Separation-- look under the Retouching section.

[drode]

D'oh. I was reading that thread. I don't know how I missed it.

Thank you.

[Boneappetit]

This could help: http://www.computerarts.co.uk/tutori...ncy-separation

The results on google might be because the correct theme should be: Spatial frequency separation (Photoshop)

[kav]

Originally Posted by Boneappetit

This could help: http://www.computerarts.co.uk/tutori...ncy-separation

The results on google might be because the correct theme should be: Spatial frequency separation (Photoshop)

I read the article for fun. Here's what you must realize. Everything he did there can be accomplished just as well without creating all of these "frequency" layers. The author is just seeking page views and nothing more. If you can't get similar results when you try to repeat it, it's not your fault, as the "split" crap had nothing to do with the end result. Also an S curve is a very bad way of finding rough spots. An oscillating curve works a bit better for some stuff. In the end, you want to use things that allow you clean results when used properly and a lot of control. This is just going to give you longer save times for little to nothing in return.

I've visited this site for a very long time. People often come here looking for advice on their work, so advocating stuff like this can really just propagate bad work.

[Boneappetit]

Sorry for trying to help !!!

[kav]

Originally Posted by Boneappetit

Sorry for trying to help !!!

This is a reason I don't post as much anymore. I unintentionally come off as abrasive. You always seem to try to be helpful. I just thought the article was trash journalism :P. It's the kind of thing where people will wonder why their results don't match, and it's because the technique described had no real influence on the end result like I mentioned. Don't take it as a personal attack.

I posted on here a very very long time ago (early 2000s) regarding how to deal with facial hair on women in photoshop. The results were primarily suggestions that I blur it out. I didn't do so for obvious reasons. I'm not necessarily the best, but I don't like to point people toward things that will push their learning curve in the wrong direction. It's like when people interested in photography discover hdr programs and creating cartoony stuff. It looks cool to them for a short while because it's new, but it hurts the pace of their long term growth.

[Der_W]

Asymmetrical spatial frequency separation (full terms ) is not the usual symmetric spatial frequency separation (hence the "a" in front of symmetrical, although nobody would ever call it like that).
The whole idea of using an "asymmetric" split was (I guess) introduced by Murray and referred to the idea of having all the color information on the low pass layer instead of having most color on the low pass and some on the high pass.
This is actually a pretty simple technique that works by first duplicating your low pass layer, blurring the lower one and setting the top one to "Color" mode. Now merge these two layers and perform the usual split with the high pass layer.

I guess the earliest information about it was here: http://www.retouchpro.com/forums/pho...lated-rgb.html
And after a while Murray created a thread about it that I unfortunately can't find right now.

Edit: Found it I guess although it was not a thread but a post in the very thread I just linked: http://www.retouchpro.com/forums/pho...tml#post274376

[drode]

Thanks!

When I ask for help here I invariably get a number of great answers. However, I often find that I don't fully understand the answers so I have to go off and read lots of other material, practice and then maybe ask a follow up question that starts the cycle again.

This is one of those cases. I understand the general concepts but figuring out how and when I might use it is tough.

[kav]

Originally Posted by Der_W

Asymmetrical spatial frequency separation (full terms ) is not the usual symmetric spatial frequency separation (hence the "a" in front of symmetrical, although nobody would ever call it like that).
The whole idea of using an "asymmetric" split was (I guess) introduced by Murray and referred to the idea of having all the color information on the low pass layer instead of having most color on the low pass and some on the high pass.
This is actually a pretty simple technique that works by first duplicating your low pass layer, blurring the lower one and setting the top one to "Color" mode. Now merge these two layers and perform the usual split with the high pass layer.

I guess the earliest information about it was here: http://www.retouchpro.com/forums/pho...lated-rgb.html
And after a while Murray created a thread about it that I unfortunately can't find right now.

Edit: Found it I guess although it was not a thread but a post in the very thread I just linked: http://www.retouchpro.com/forums/pho...tml#post274376

I think I was in a bad mood yesterday, but if anyone wishes to take me up on it, I would bet I can do everything enabled by that technique in a simpler manner with a smaller file size . I've tested so many ways of doing things, and the simple ones often work the best. Too many odd layers often just creates preview bugs making the image more difficult to complete.

[Der_W]

That'd be nice :-).
I'm always in for improving my workflow and techniques.

Though we should probably discuss this in a different thread not to confuse the OP even more?

[ShadowLight]

... I'm not sure about the benefits of that approach
pull the luminosity only (example img 1)
or desaturate (different thing, so - example img 2)

and see how much HF detail will be dumped on the LF