DAN MARGULIS APPLIED COLOR THEORY

Sunrise, Sunset, Reflections

From: Dan Margulis

Date: May 13, 2013 10:27:39 AM AST

Subject: [colortheory] Sunrise, Sunset, Reflections

When buyers of the book register to use the video resources they get automated follow-up e-mails from me every few weeks. Understanding that these are not personal messages, many readers have replied to me with comments or questions. From them, it certainly seems that the book has been well received.

Some of the questions that come in, I now realize, would do well as discussion points for the group. And since I can’t answer them as quickly as one might like, or where my own answer might not be to everybody’s liking, I’d welcome hearing other people’s POV.

Dan Margulis

Begin forwarded message:

From:

Date: May 12, 2013 5:07:14 PM AST

To: Dan Margulis

Dear Dan,

I had been familiar with the PPW from your Lab Color book and from Kelby videos, but the new book and the associated resources have definitely helped me take the next step. I’m currently reading and digesting Chapter 10. Thank you for writing the book and for leading the PPW panel effort (and for distributing it at the right price!)

One question I have is about color evaluation at sunrise and sunset, especially when there is an ocean or a lake involved. There are a variety of warm colors that can be present, and they can create a situation where there is no blue sky and there are no white or gray clouds to use as a reference. Yet it’s still possible that there is a color cast that needs attention. The normal color rules about sky (and other things) seem to be not quite right, though I can imagine, for example, that there might be a yellow cast that should be dealt with. But, how would you tell? Sky, clouds, rocks, sand, and other things will likely be taking on hues that are warmer than normal for the observer who is there. Do you have any guidelines for this class of situation?

The other thing I wonder about is reflective surfaces. For example, consider the SUV in the desert in the “Initial Color Evaluation” video for Ch. 3. When I look at it, I can see the reflection of mountains and sky just below the crease in the side that runs below the door handles. So, right under that crease, I’d expect the color to have a blue (reflected sky) component, changed of course by however the SUV paint differs from neutral. Clearly if the surface were really neutral and mirror-like, we’d worry as much about the color of the reflected scene as about the color of the SUV’s paint. But at what point do we stop worrying about reflections?

Another example: if somebody holds a known white target so that the reflected green/yellow light of the leaves on nearby trees makes the paper in the image negative in the a channel and positive in the b channel, do we try to make the target white, or do we allow the green/yellow to reflect off the white?

In general: how much should I worry about reflections?

Thank you for any advice.

___________________________________________________________________________

From: Frederick Yocum

Date: May 13, 2013 3:43:39 PM AST

Subject: Re: [colortheory] Sunrise, Sunset, Reflections

Dan

This could be a good use for the group, though often you will get divergent views which might be equally confusing.

Were you thinking, you would raise the question, a discussion would happen and you would synthesize a response from the points raised, or something more direct? A stackoverflow.com for color theory?

Frederick

___________________________________________________________________________

From: Dan Margulis

Date: May 13, 2013 7:54:16 PM AST

Subject: Re: [colortheory] Sunrise, Sunset, Reflections

Frederick writes,

Dan

This could be a good use for the group, though often you will get divergent views which might be equally confusing.

Were you thinking, you would raise the question, a discussion would happen and you would synthesize a response from the points raised, or something more direct?

I never know what the group will say or do. Recently we have had many threads specifically about the new book or specifically about the recent Adobe announcement. These are of interest to many of us but not necessarily to others.

The questions posed by the reader may have been in the *context* of the book but they had nothing to do with the PPW or any of its tools, they are straight color correction issues that have been a problem from the beginning of time. Sunset images are notoriously difficult to work with and many of us have spent more time than we would like to admit trying to make them look good. Ditto his question about colored light reflecting off near-neutral objects.

So, since color correction is what we’re here for, I thought I’d shoot the line out there and see if there are any nibbles.

Dan Margulis

___________________________________________________________________________

From: “dlruckus”

Date: May 14, 2013 12:48:59 AM AST

Subject: [colortheory] Re: Sunrise, Sunset, Reflections

From the perspective of a non guru, I have to say a, not very satisfactory, --”it depends”--. The closest thing to a rule of thumb I can come up with is that the final image has to be believable by it’s intended audience. That can cover an extreme range of color correction not infrequently including deliberate mis correction.

An example would be something like a white barn at late sunset where the sun side is very red while the opposite side is very blue.Film will record that but people often dislike red and blue appearing white barns while they will accept a warm color on the sun side only or blue in shadows if after dusk etc. On the other hand some people like screaming color and will be happy with whatever, as witness some of the HDR effect examples that abound these days.

As far as sunsets are concerned I would say if you like the color and others do also then have at it. I spend most evenings for 6 months of the year watching sunsets from my home with a panoramic view by a small lake. I have never seen any two alike in over a decade there.

Neutrals will follow the same reasoning as the white barn. Likewise no one wants to see green skin except perhaps in Oz so light under a tree canopy may require masking skin and corrections. It also will make a big difference as to whether or not a reflection is just a general lighting cast or if it is an actual image that can be seen in reflection.

Dan has stated, I believe, that deep shadows should be neutralized so one could use that as a general practice perhaps and apply the correction overall to get a starting point of sorts.

Not much help here I guess but there is no way to fix an immutable rule for these things in my HMO.

Regards,

Duane Ruck

___________________________________________________________________________

From: Stephen Marsh

Date: May 14, 2013 12:55:35 AM AST

Subject: [colortheory] Re: Sunrise, Sunset, Reflections

One question I have is about color evaluation at sunrise and sunset, especially when there is an ocean or a lake involved.

This is a basic artistic decision in my opinion. I believe that Lab colour readings would probably be helpful in deciding whether one should go warmer/colder or more yellow or more blue or more magenta or more green. If one Googles “sunset photo” there is such a range of images returned that it is impossible to formulate any rules, it is highly dependent on the image content and the artistic mood that one wishes to impart.

The other thing I wonder about is reflective surfaces.

Cameras can often record things differently to the human observer. This is not just for highly reflective surfaces, colour “spill over” can happen on skin from bright clothing and I often find that I am correcting this in personal photos. In these cases, I believe that the human observer viewpoint is correct and generally speaking I am not looking for an artistic effect in most of these cases.

Stephen Marsh

___________________________________________________________________________

From: Doug Schafer

Date: May 14, 2013 11:00:45 AM AST

Subject: [colortheory] Re: Sunrise, Sunset, Reflections

I would welcome questions from others, a discussion here, and a conclusion summary answer from you.

I’d think we would all learn new stuff or confirm/solidify what we know.

Doug Schafer

___________________________________________________________________________

From: Doug Schafer

Date: May 14, 2013 11:23:21 AM AST

Subject: [colortheory] Re: Sunrise, Sunset, Reflections

slightly OT;

You might be amazed at how many 3D render algorithms are written to correctly add-in this effect to help create reality result images.

One program I have calls it “color bleed” with a control slider of 0-100%. Other programs have different names and varying controls. But it is important, and needs to be controlled correctly.

Next time you watch a movie or print ad with 3D graphics special effects rendered for the scene, especially in animated characters (people & skin), you will see reflected/included colors that add to the realism....if done correctly.

Doug Schafer

--- In colortheory@yahoogroups.com, “marshyswamp71” <samarsh@...> wrote:

....snip....for highly reflective surfaces, colour “spill over” can happen on skin from bright clothing ....snip>

Stephen Marsh

___________________________________________________________________________

From: J Walton

Date: May 14, 2013 11:59:49 AM AST

Subject: Re: [colortheory] Sunrise, Sunset, Reflections

I’ll share something that has helped me with those kinds of scenes. In the

case of any saturated color, focus on the contaminant (or as I think Dan

puts it, the “unwanted” color, another color expert called it the

“corrective” color). Basically the cyan in reds and, the magenta in greens,

the yellow in blues, etc. So in the case of a sunset I’d watch my LAB

numbers but I’d pay particular attention to my CMYK numbers, even if the

image is in RGB or LAB. If there is any cyan in a sunset shot, I’m using

that as my “highlight” queue and color balancing off of that.

A super-saturated shot (like a sunset) will hide a lot of problems, but

the contaminating color will reveal queues that can help you get rid of

what clearly doesn’t belong. The rest is art, not science, so enjoy it.

J Walton

___________________________________________________________________________

From: Francis Corvin

Date: May 15, 2013 6:52:36 AM AST

Subject: Re: [colortheory] Sunrise, Sunset, Reflections

Here are my contributions, from a photographer’s perspective, FWIW.

One question I have is about color evaluation at sunrise and sunset,

especially when there is an ocean or a lake involved. (...) Do you

have any guidelines for this class of situation?

This is where working with the photographer really pays off IMHO. They

have access to the RAW file and can set the white balance to 5500K,

which would be correct most of the time, if slightly warmish. More

importantly, it would give consistent results across a series of shots.

Trying to find absolutes in these circumstances seems futile to me: the

colours vary with the progress of sunrise, sometimes in a couple of

minutes; the ambiant humidity will also affect the sky colour; and

particles in the atmosphere, even distant, will tend to produce more

colourful sunsets/sunrises (volcanic eruptions work well for this).

<http://www.google.co.uk/search?num=100&cr=countryUK%7CcountryGB&newwindow=1&tbs=ctr:countryUK%7CcountryGB&q=eyjafjallajokull&spell=1&sa=X&ei=0meTUY31KdSjhgezpoDgDg&ved=0CCwQvwUoAA>

The other thing I wonder about is reflective surfaces. (...) But at

what point do we stop worrying about reflections?

There are plug-ins that average a colour in a selection. In similar

circumstances, I have made a selection that attempts to capture sky and

landscape 50/50 (because that is how the car is illuminated, if in the

shade), averaged out, and used the result as a guide to colour balance.

There is always going to be an element of personal taste after this,

particularly when highlights and shadows get reflections from

differently coloured scenery.

Another example: if somebody holds a known white target so that the

reflected green/yellow light of the leaves on nearby trees makes the

paper in the image negative in the a channel and positive in the b

channel, do we try to make the target white, or do we allow the

green/yellow to reflect off the white?

That depends on how exact you want to be. I personally strongly prefer

to eliminate green casts, especially on portraits as they are so

unpleasant. But the issue is sometimes that the green cast is

directional and you have to use the brush to only eliminate the green

cast in certain areas. “Good” photographers don’t wear loud shirts. ;)

Francis Corvin

___________________________________________________________________________

From: Dan Margulis

Date: May 14, 2013 6:36:52 PM AST

Subject: Re: [colortheory] Re: Sunrise, Sunset, Reflections

Stephen writes,

One question I have is about color evaluation at sunrise and sunset, especially when there is an ocean or a lake involved.

This is a basic artistic decision in my opinion. I believe that Lab colour readings would probably be helpful in deciding whether one should go warmer/colder or more yellow or more blue or more magenta or more green. If one Googles “sunset photo” there is such a range of images returned that it is impossible to formulate any rules, it is highly dependent on the image content and the artistic mood that one wishes to impart.

I agree completely with that, as far as it goes. However, IMHO the most important feature of a sunrise/sunset is that it is both bright and (often) brilliantly colored. Unfortunately, it isn’t possible to combine these two attributes perfectly. The brightest color we can produce is 255r255g255b or 0c0m0y, but that is a white; if we make it yellow by reducing blue/increasing yellow that also darkens it slightly; if we make it orange by also reducing green/adding magenta, that makes it darker quicker. This leaves us between the sword and the wall with respect to the sun, which is supposed to be both brilliant and orange.

The solution of painters (who have faced the same problem since time began) is historically to have a very small sun, completely blank, surrounded by a rapid transition to strong color. The easiest way to arrange that in a digital file is by going into LAB and inserting a strong cast, limited by mask to the lightest colored area that isn’t blank.

However there are also other possibilities, such as deliberately making something blue that in principle shouldn’t be so, with the idea of emphasizing orange elsewhere.

The other thing I wonder about is reflective surfaces.

Cameras can often record things differently to the human observer. This is not just for highly reflective surfaces, colour “spill over” can happen on skin from bright clothing and I often find that I am correcting this in personal photos. In these cases, I believe that the human observer viewpoint is correct and generally speaking I am not looking for an artistic effect in most of these cases.

That sums up my opinion, too. The human visual system rejects these casts unless they are very strong. The example the OP gave, as I recall, was a strong blue sky reflecting off a silvery car. In that case a human observer might pick up the blueness, but not to the extent that the camera would. So there a certain amount of judgment is needed.

Light that is slightly green because it’s reflected off foliage and so contaminating a face, that’s another story, it should be taken out IMHO.

Dan Margulis

___________________________________________________________________________

From: “twlane”

Date: May 15, 2013 3:23:29 PM AST

Subject: [colortheory] Re: Sunrise, Sunset, Reflections

--- In colortheory@yahoogroups.com, J Walton wrote:

...So in the case of a sunset I’d watch my LAB

numbers but I’d pay particular attention to my CMYK numbers, even if the

image is in RGB or LAB. If there is any cyan in a sunset shot, I’m using

that as my “highlight” queue and color balancing off of that...

This sounds interesting - could you provide more detailed steps/instructions?

Thanks,

Ted Lane

___________________________________________________________________________

From: J Walton

Date: May 15, 2013 9:22:56 PM AST

Subject: Re: [colortheory] Re: Sunrise, Sunset, Reflections

It’s much easier to speak in the language of images. If someone has a

sunset shot we’d have an easier time of it.

But anyway having cyan in a brilliant red-yellow flash of light is a

mistake, or at least it will limit your potential saturation. The only cost

of removing it completely is shape, which isn’t a factor in a sunset,

especially towards the light source. If it were a blue light I’d be looking

for yellow. If it were a green light I’d be looking for magenta. Just set

your info palette to CMYK and your eyedropper radius to 11 (or so).

Sometimes these clues are enough to tell me what to do with my secondary

colors. But that’s as close to a rule as I can get, the rest is art. So

have fun with it!

J Walton