RetouchPRO

Go Back   RetouchPRO > Technique > Photo Retouching
Register Blogs FAQ Site Nav Search Today's Posts Mark Forums Read


Photo Retouching "Improving" photos, post-production, correction, etc.

Dan Margulis and his curves Philosophy...need help

Reply
 
Thread Tools
  #1  
Old 04-13-2012, 07:30 PM
Andymania Andymania is offline
Member
 
Join Date: Sep 2011
Posts: 49
Dan Margulis and his curves Philosophy...need help

I am 75% done reading Professional Photoshop by Dan Margulis. It opened up alot of new stuff to me. However, I am still confused to what he mentioned about curves. He says that you should avoid the master curve since it raises/lowers all RGB values at the same time. He mentioned that it is better to adjust each channel individually. I tried messing around with this but I also noticed that not only do you darken and lighten each channel individually but you also change the hue as well. So I still don't know what is the advantage of individually adjusting each channel when you are radically altering its hue as well. Maybe someone can clearly explain to me what is the advantage to adjusting each channel seperately as opposed to using the master curve...

Thanks!

-Andy
Reply With Quote top
  #2  
Old 04-13-2012, 08:13 PM
kav kav is offline
Senior Member
 
Join Date: May 2009
Posts: 574
Re: Dan Margulis and his curves Philosophy...need

You really are over thinking this. It just gives you more control. Of course it changes the hue when you adjust one differently from another. It gives you the ability to set that however you like, and quite often the value that looks right when something is darkened or lightened doesn't actually correspond to what looked correct before. Anyway it's more difficult to manage, but you can really fine tune the thing if you wish. That is the whole advantage. Some people use this to help develop the right look for the color of the image.

I'm kind of hoping this doesn't spawn more change curve to luminosity posts. That's another commonly misunderstood thing. Luminosity can still make the color look incorrect. It just shoves color corrections onto a different step if they're behaving really weird. This is more common if you're adding a lot of contrast via one or two curves, but I wouldn't recommend that for the vast majority of things anyway.
Reply With Quote top
  #3  
Old 04-14-2012, 12:03 AM
redcrown redcrown is offline
Member
 
Join Date: Jan 2006
Posts: 87
Re: Dan Margulis and his curves Philosophy...need

There are definitely times to adjust each RGB channel separately in Curves, but those are when you want to change color balance, not tones. Fixing a color cast, modifying skin color, etc.

But to just adjust contrast (shadows, highlights, midtones) stick with the composite channel. It's much easier.

And I kind of agree with KAV, that puting a Curves adjustment in luminosity mode is rarely needed. Maybe only in cases where the Curves adjustment is so strong that it drives colors crazy.

Years ago I read a facinating story somwehere about the development of "curves" in the early days of Photoshop. As I recall, the initial "normal" version was really luminosity only. But the developers didn't like the way it looked. Something was missing. I recall that Nack and Knoll were involved in the debate, and the developers had to do a lot of extra development to get color "right". The result being that the final curves "normal" adjustment involves a lot of heavy math. The initial and simpler method is what we now have as luminosity mode.
Reply With Quote top
  #4  
Old 04-14-2012, 06:16 AM
Andymania Andymania is offline
Member
 
Join Date: Sep 2011
Posts: 49
Re: Dan Margulis and his curves Philosophy...need

LOL Hey Kav. We meet again. I'm just trying to understand some of the stuff I read. I'm not trying to over think anything. This is for my own personal knowledge and understanding.

Redcrown, I hear what you are saying since it made more sense to stick with the composite curve after messing around.
Reply With Quote top
  #5  
Old 04-14-2012, 10:12 AM
edgework edgework is offline
Senior Member
 
Join Date: Jul 2004
Location: Brooklyn
Posts: 494
Re: Dan Margulis and his curves Philosophy...need

Keep in mind that Dan's techniques evolved in the days when CMYK was the only game in town. Pulling the CMYK master curve will definitely throw a cast into your image, since C, M and Y are not of equal strength (that's why it takes 10% more C to balance M and Y into a neutral.) In theory, R, G, and B values are of equal strength and so the master curve should simply darken and lighten, without a cast. In theory. Result may vary.

Another of Dan's maxims is that if it's wrong in one part of an image, it's wrong in the image as a whole. Hence, his emphasis on global moves, and paying particular attention to balancing skin tones and obvious neutral areas. They are reliable measuring points to see where the image is out of balance. Getting them right will do much to help the entire image and that takes individual curve moves, not master curves.
Reply With Quote top
  #6  
Old 04-14-2012, 10:15 AM
andrewrodney's Avatar
andrewrodney andrewrodney is offline
Senior Member
 
Join Date: May 2010
Location: Santa Fe
Posts: 1,091
Re: Dan Margulis and his curves Philosophy...need

Quote:
Originally Posted by Andymania View Post
So I still don't know what is the advantage of individually adjusting each channel when you are radically altering its hue as well. Maybe someone can clearly explain to me what is the advantage to adjusting each channel seperately as opposed to using the master curve...
Your analysis is good!

Dan is famous for making a big deal out of his so called master curve issue. First off, it was designed this way for good reason (testing by Thomas Knoll years and years ago, before Dan even knew what Photoshop was showed most people preferred having the hue alter with a curve move). Without this move, images appear to be a bit ‘flat’.

Next, there are plenty of provisions for making a curve without this move if you so desire. For example, fade luminosity. As others have pointed out and other workarounds is often not what you will desire although in Dan’s mind, this is the ‘correct’ way to adjust a curve (individual curve adjustment).

Use whatever technique provides the appearance you desire, be it in an Adobe raw converter or Photoshop. Much of the functionality in Photoshop was created by really smart people like Thomas after doing a lot of testing and asking users what they prefer. And there is often an alternative approach if so desired. Dan makes it sound like the current and designed approach is unconscionable.

Dan further goes on about how egregious the tone curve effects in ACR are, using the same ‘logic’.

If you really want a reality check on all this made up (by Dan) controversy, I suggest you read this:

http://www.luminous-landscape.com/essays/Curves.shtml
Reply With Quote top
  #7  
Old 04-16-2012, 07:47 PM
Andymania Andymania is offline
Member
 
Join Date: Sep 2011
Posts: 49
Re: Dan Margulis and his curves Philosophy...need

Thanks guys for the input. But I am still fumbled about this. Dan in chapter 2 of "Professional Photoshop" talks about contrast. He shows several before and after pics as well as the different individual RGB curves that he used to make the after version. Now he used 3 different S curves for each individual channel. I noticed that there wasn't a hue shift either. But those 3 individual channel curves would probably produce the same result as one modified curve in the master composite. So why bother? ( I know you guys mentioned he is from the CMYK generation, but why use the same principle for RGB?)

He also talks about white point and how the camera automatically sets the white point to specular highlights when the white point should be the lightest significant part of the image. Now I wonder again. If I am doing color correction, shouldn't technically the whitest part of the image be the specular highlight since I KNOW that for a fact that speculars should equal R=255 G=255 B=255? I tried setting white point with the white eye dropper tool in curves to something a single tone darker than the specular and it just blew out my highlights. So I have no idea what Dan is talking about in this case.
Reply With Quote top
  #8  
Old 04-16-2012, 08:12 PM
andrewrodney's Avatar
andrewrodney andrewrodney is offline
Senior Member
 
Join Date: May 2010
Location: Santa Fe
Posts: 1,091
Re: Dan Margulis and his curves Philosophy...need

Quote:
Originally Posted by Andymania View Post
So why bother? ( I know you guys mentioned he is from the CMYK generation, but why use the same principle for RGB?)
Exactly. He’s making a big deal out of nothing really. The behavior was designed to work this way for a reason, he doesn’t agree. And yet, there are easier work-arounds if you want a different process. Dan likes to make stuff far too complicated (much like trying to decipher his writings).

Quote:
He also talks about white point and how the camera automatically sets the white point to specular highlights when the white point should be the lightest significant part of the image.
That just isn’t true, certainly not with raw data and probably not true for all JPEG camera captures. Dan isn’t a photographer by a long shot! In fact, he’s made a calling by taking butt-ugly originals that no photographer in his right mind would produce and polishing that turd. That is a useful technique to know; making lemonade out of lemons. But if you don’t work with lemons, you can dismiss about 90% of what he is writing about. GIGO:Garbage In Garbage Out!

Look at how ridiculous his statement is from the perspective of someone what understands photography: a camera capturing raw or JPEG simply is not going to clip a highlight to anything if the photographer expose so it doesn’t clip. To say the camera automatically sets the white point to specular highlights comes from someone who has no idea about photography and exposure 101. Under expose an image 1 stop (or for a scene that has a dynamic range wider than the sensor can capture and expose for the shadows instead of the highlights), the highlights simply don’t clip to white. Or anywhere near it.

Quote:
If I am doing color correction, shouldn't technically the whitest part of the image be the specular highlight since I KNOW that for a fact that speculars should equal R=255 G=255 B=255?
Well in a well behaved RGB working space (sRGB, ProPhoto, Adobe RGB etc) yes, that is neutral. But there is no tone there, it is a true specular highlight. At least when white balancing raw data, you’d never use that area to set WB. You’d use something white but with a bit of tone (something like the 2nd white patch on a Macbeth Color Checker).

Quote:
I tried setting white point with the white eye dropper tool in curves to something a single tone darker than the specular and it just blew out my highlights.
Because in Photoshop, that is what the tool does (clips to those values). You can alter the values lower than 255 by double clicking on the eyedropper.

You’d use this tone to WB in a raw converter and again, not on a specular white. In Photoshop, you are working on gamma corrected data, you’d gray balance in an attempt to neutralize a cast. Raw data is linear encoded. Half of all the data is in the first lightest stop. So it is in this area with this data you white balance.
Reply With Quote top
  #9  
Old 04-16-2012, 09:30 PM
kav kav is offline
Senior Member
 
Join Date: May 2009
Posts: 574
Re: Dan Margulis and his curves Philosophy...need

Quote:
Originally Posted by Andymania View Post

He also talks about white point and how the camera automatically sets the white point to specular highlights when the white point should be the lightest significant part of the image. Now I wonder again. If I am doing color correction, shouldn't technically the whitest part of the image be the specular highlight since I KNOW that for a fact that speculars should equal R=255 G=255 B=255? I tried setting white point with the white eye dropper tool in curves to something a single tone darker than the specular and it just blew out my highlights. So I have no idea what Dan is talking about in this case.
I disagree that specular highlights should go to the edge. You really don't need a single 255,255,255 pixel in the photo to make it look good. Adjusting channel by channel has been used successfully by some people to get certain looks. I mentioned I use it sometimes for creating curves to be used in lightening or darkening areas for retouching purposes, but it's for difficult blending there where details would break or go grey and bleh. Much of the time if you're blending a surface, the shadowed portions really aren't just a darker version of the same color, so I do set things separately because i'm referencing colors.

I haven't read his book, so I'm not sure how to respond beyond that. Even my previous paragraph makes more sense as a visual explanation.
Reply With Quote top
  #10  
Old 04-18-2012, 10:29 AM
Andymania Andymania is offline
Member
 
Join Date: Sep 2011
Posts: 49
Re: Dan Margulis and his curves Philosophy...need

You’d use this tone to WB in a raw converter and again, not on a specular white. In Photoshop, you are working on gamma corrected data, you’d gray balance in an attempt to neutralize a cast. Raw data is linear encoded. Half of all the data is in the first lightest stop. So it is in this area with this data you white balance.[/QUOTE]

Andrew,

You gave lots of great info! But can you elaborate in simpler terms to what you mean in the above? What is meant by gamma corrected and linear encoded?

Also since I was talking about color correction, Dan talks about going by the numbers in the info pallete in regards to getting accurate color. I am practicing some of these concepts he mentions on some professionaly shot raw images. The problem is is that since I was not actually present at the photo shoot, I could not tell you what the right RGB values should be for the model's skin tone. Obviously, if there is significant color cast I wil spot it. Now if I know for a fact that something is supposed to be neutral gray, I can then alter those RGB values via curves. But in lots of these images there are no neutral grays. So how would I go about doing this? I am assuming I have to rely on either my own taste/judgment or the art director's taste/judgement.
Reply With Quote top
Reply

  RetouchPRO > Technique > Photo Retouching


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -6. The time now is 06:21 AM.


Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
SEO by vBSEO ©2011, Crawlability, Inc.
Copyright © 2016 Doug Nelson. All Rights Reserved