Topic: Lab Bauble: ab by Lum (Page 1 of 1) Pages that link to <a href="https://ozoneasylum.com/backlink?for=28789" title="Pages that link to Topic: Lab Bauble: ab by Lum (Page 1 of 1)" rel="nofollow" >Topic: Lab Bauble: ab by Lum <span class="small">(Page 1 of 1)</span>\

 
warjournal
Maniac (V) Mad Scientist

From:
Insane since: Aug 2000

posted posted 01-03-2007 08:02

This is one of those things that has been on my mind for awhile. So I'm sitting here working on my vectors and this thing just popped into my head. Just popped in to say Hello. I guess the part of my brain that had been working on this finally figured it out.

I've done quite a bit of chatting about masking using Luminosity. You know, grab the highlights, mask, and operate.

I've also talked about using Linear Light and Curves in a wave kind of fashion. Where you use a channel to add/subtract from itself.

The thing that has been on my mind is offsetting ab by using Lum values. Not in the mask and tweak method, but the Curves LL method.

- if L is high, then tweak ab like so
- if L is mid, then tweak ab like this
- if L in low, then tweak ab like that

I want to be able to do that without masking.

Gonna repeat the important part with emphasis: without masking.

While some of the tools in Photoshop do allow for that, the built-in HML just won't cut it. How dare Photoshop tell me what a midtone is! I want to be able to do it in a custom, definable, uber tweakage, WJ-style and manner.

The answer is so easy. Why, I can't believe I didn't see it before. It's uber duh!

- Start with some random photograph.
- Go into Lab mode.
- Copy photo to another layer.
- On the copy, to go Channels palette and C&P L into a and b

That last step is the magic part. Now just gotta finish setting up the rig.

- Change the blending mode to Linear Light so we can add/subtract values.
- Blending Options and turn off L channel.
- Clip a Curves Adjustment Layer to this copy.
- In the Curves, 'zero out' the a and b curves.

Once it's all set, get busy with the tweakage.

You can now very quickly and easily adjust colour balance based on L without masking. That is, based on any L value(s) with perfect blending. Touch-up with whatever... like... maybe a layer mask... hehehe.

Damn.

Stick that in your tea and stir it.

warjournal
Maniac (V) Mad Scientist

From:
Insane since: Aug 2000

posted posted 01-04-2007 02:49

I walked into work and had an epiphany. Turned around and came right back home. I went in 2 hours early on accident (don't ask), so it's alright. The epiphany has to do with the vector thing I'm working on. But 2 hours isn't enough time to thoroughly test my epiphany, so I want to chat about this Lab thing for a bit. Just my thoughts on it and why I had it tucked away in my brain.

Many moons ago I started tearing into colour cast. Steve's Curves tutorial was tremenduous in helping me with colour cast - not just fixing, but tearing apart. How is it manifested in the data? How do you shift the data to fix? When I started getting into HSL and HsY, Steve's tutorial allowed me to look closer at before and after.

Why didn't anybody tell me more about colour cast back then? You Asylumites should all be able to read my mind by now. After all, a decent chunk of it has been uploaded.

Alright, colour cast. Most y'all know what it is. While y'all prolly don't see it the way I do, you do know about it and fixes. Leap of faith? You bet. So, I'm not gonna rehash it.

Lab allows for some awesome colour cast fixing. So does HSL, HSB, and HsL if you can get past the polar thing. And, as Steve has shown, even RGB can be used. But, for this kind of colour work, I prefer Lab because of its connection to our brain and eyeballs.

One thing I've noticed about colour cast is that it's hardly ever uniform. Sometimes you have blue shadows, but the rest will be fine. Sometimes you have orange highlights, but the shadows will be fine. The amount of cast invariably varies. There is kind of a wiggle to it. Sometimes uber, sometimes not uber.

Just for giggles, use Colour Range > HML and Average to see this for yourself. If the highs and mids are uber cast, why use the same adjustment for lows and end up throwing them out of whack? Well, that is where masking comes in, which is what I'm looking to get rid of.

Imagine a cube in space. Imagine it is sliced into a bunch of planes. Those planes are a and b in Lab. Each ab plane lies at very L value. Offsettng ab values on those ab planes is what I'm after. Like drawing a spline right down center of all planes, and then giving the spline some wiggle to offsett ab values.

That is what I'm doing in the above technique. When you use the above technique, you are giving that 3d spline some wiggle in La and Lb in turn.

Now, if you leave L turned on, you can move things around in a 3d fashion. You can even adjust the fall-off. Kind of like defining a sphere in Lab space and moving it around. If you think about, it can be a very powerful shortcut for ChOps and/or channel mixing (even in RGB et al... sort of).

Colour spaces are 3d. Never forget that. Or don't think about it. Whatever floats for you. All I know for sure is that thinking in 3d has helped me tremenduously. Even if said 3d is one 2d plane at a time.

So, fixing variable colour cast (variably) based on L without hit-or-miss masking and visual, dynamic feedback. That's for me.

I've got more thoughts on how colour cast is manifested. Not just in the data, but how it happens in the first place. I suspect logs are involved, but this particular thing is currently in the back of my mind. Have you ever seen exponential functions like gamma in 3d space? Hmm... tasty. But for another day.

play.fiddle.learn



Post Reply
 
Your User Name:
Your Password:
Login Options:
 
Your Text:
Loading...
Options:


« BackwardsOnwards »

Show Forum Drop Down Menu