Home · Register · Search · View Winners · Software · Hosting · Software · Join Upload & Sell

Moderated by: Fred Miranda
Username   Password

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Canon Forum | Join Upload & Sell

1       2      
3
       4       end
  

Archive 2013 · Shooting in golden hour WB question
  
 
jcolwell
Online
• • • • • • •
Upload & Sell: On
p.3 #1 · Shooting in golden hour WB question


gschlact wrote:
...Perception, although it can be influenced by different variables, Will remain constant with the same variable for the same person....


I doubt that very much. Tomorrow, I might agree. Subject to lighting, and opinion.

Like two arrows meeting in mid-air. Impossible, yet inevitible.



May 13, 2013 at 03:45 AM
gschlact
Offline
• • •
Upload & Sell: On
p.3 #2 · Shooting in golden hour WB question


Monito,
Your notion of subjective alteration is the exact reason I described the method for reverse WB to apply the alteration in non subjective fashion.

I disagree regarding perception consistency - take example of director at the show sitting in same director chair with same full control of lighting, costumes and actors. He snaps photo and now wants to reproduce the lighting cast in print, but not by guessing whether he post processes the pink cast properly on the white short.

Besides my reverse WB method, the only other way would be to do spectral measurement in the field and manually produce that cast again on am neutralized version of photo, much like my Reverse WB method describes.





May 13, 2013 at 03:49 AM
jcolwell
Online
• • • • • • •
Upload & Sell: On
p.3 #3 · Shooting in golden hour WB question


gschlact wrote:
Monito,
Your notion of subjective alteration is the exact reason I described the method for reverse WB to apply the alteration in non subjective fashion.

I disagree regarding perception consistency - take example of director at the show sitting in same director chair with same full control of lighting, costumes and actors. He snaps photo and now wants to reproduce the lighting cast in print, but not by guessing whether he post processes the pink cast properly on the white short.

Besides my reverse WB method, the only other way would be to do spectral measurement in the field and manually produce
...Show more

...or, you could use your own perception.

P.S. I don't think the director "snaps photo". The director has a vision of the end product. The translantion and transposition between what the photographer takes and what the end user sees is only dependant on what the photographer captures, and how she processes it.



May 13, 2013 at 03:59 AM
Monito
Offline
• • • • • •
Upload & Sell: Off
p.3 #4 · Shooting in golden hour WB question


gschlact wrote:
I disagree regarding perception consistency - take example of director at the show sitting in same director chair with same full control of lighting, costumes and actors. He snaps photo and now wants to reproduce the lighting cast in print, but not by guessing whether he post processes the pink cast properly on the white short.


However, we, as photographers, never sit in the same chair with the same lighting unless we are recreating a canned lighting setup in a studio for something like passport photos or catalogue items.

In any case, even if we take your assumptions, the pink cast as perceived will vary depending on many factors, including what else is lit on the stage. If everything is light with the identical pink light, then nobody can know just how pink the paper is. We will know that it is pink because faces will look pink even though they are not neutral gray cards, but we won't know whether it is an orangy pink or magenta pink because faces are not neutral. If we are very careful observers we can get a rough sense of whether a white shirt is in orangy or magenta light even without a reference, but it will only be a rough estimate and subject to a lot of vagaries, including how tired we might be at the time.

We only get a semi-objective sense (as human observers) of the pinkness if there is some other non-pink lighting in our field of view showing a white that we perceive as white. Then the way forward is clear: put the gray card at the position of that white, and set the balance for that white. Then the paper in the pink light will look like we saw it at the time (but not necessarily how we remember it or how we would like to see it).

But that only works if the scene includes a white that we agree is white and that we are able to balance for.



May 13, 2013 at 04:05 AM
Monito
Offline
• • • • • •
Upload & Sell: Off
p.3 #5 · Shooting in golden hour WB question


AJSJones wrote:
We learn a lot by looking at the stars and the spectra of their light - red-shifts for distances, atomic composition, even now atmospheric component gases on exoplanets. Do you think those guys "eyeball" the white balance till it looks nice or do you think they have calibrations and standardization protocols to ensure accurate recording for those studies. SInce we can't even see the stuff they record, there's no eye-brain axis involved, just the data , ma'am. I think that is what you would like for your images - then play to your heart's content later, right? Raw data
...Show more

Star light is only objectively calibratable because we can measure the spacing between the emission lines (bright lines) that correspond to the elements, especially Hydrogen, Helium, Carbon, and Oxygen (there are some other elements that the astrophysicists use for the finest measurements). Likewise we can see the composition of dust clouds and exoplanets because of the absorption lines (dark lines) of the light that passes through them.

Such spectral analysis is not feasible for photographers. Not the least of the problems taking that approach is the fact that the light we need to measure is not generated (as by a star) but is reflected off skin, clothing and other objects. Not only do we not know the spectra of the light source, but we don't know the absorption, reflection, and retransmission spectra of the objects (three different types of spectra in play). There are physical measurements that can get around such problems, but not even for your average pro.

So the astrophysical analogy is not applicable.




May 13, 2013 at 04:13 AM
Monito
Offline
• • • • • •
Upload & Sell: Off
p.3 #6 · Shooting in golden hour WB question


Complicating things is that the "white shorts" cited upthread are never actually white. They age differently, they are made from different fabrics, they have been washed with different combinations of colours, and will have different colours of skin and underwear beneath them. A further complication is that different detergents have different kinds of whiteners which are white substances and special substances that fluoresce in ultraviolet light to retransmit in the visible range to get the "whiter than white" effect the advertisers so love.


May 13, 2013 at 04:15 AM
Gunzorro
Offline
• • • • •
Upload & Sell: On
p.3 #7 · Shooting in golden hour WB question


Fred -- I agree that the simpler approach is the best approach with shooting under those conditions. Also, rapidly changing light and extended exposure times leaves very little time for tinkering with color shifts in the field.

I find AWB usually gets it pretty close for city shots and most open vista landscapes, at least close enough to need only minor PP tweaks.

I'm sure it depends heavily on how precise a photographer or videographer needs to be, but I can afford to approach pretty simply.



May 13, 2013 at 06:22 AM
Eyeball
Online
• • • •
Upload & Sell: On
p.3 #8 · Shooting in golden hour WB question


Guy,

Let me point out some inconsistencies and conflicts in the things you have stated so that maybe you will start to see the difficulty:

First, your stated "reverse white balance" method:


1. Pick daylight WB
2. Shoot gray card / white card in funky light
3. Import gray card shot and scene shot into PS
4. Sample the casted gray card
5. Duplicate layer for scene shot
6. Creat layer and full with sample casted color and make semi opaque.
7. Duplicate basse scene layer under the cast layer
8. Do white Balance on the duplicate base layer to remove cast
9. Adjust transparency of cast only layer.


The inconsistency in your thinking is that in step 1 you are using Daylight as your reference point but in your second post to this thread you said:


your suggestion is understandable but won't produce accurate results as it still make assumptions in the Daylight setting itself.


I would also point out that there is a simpler way to do your reverse white balance process, particularly if shooting raw. You just neutralize the gray card to establish one "end-point" and select a "standard" white balance like Daylight/5500K to establish the other "end-point" and then adjust white balance between the two. It is easy to establish the two end-points in a relatively precise manner. Where the subjectivity comes into play is your step 9 or in my method when I say "adjust between the two".

But your initial post implies that you are not satisfied with establishing two "precise" end-points. You are looking for a "scientifically precise" method of reproducing WHAT YOU OBSERVED, or at least something that will get you automatically to something closer to what YOU OBSERVED than either of the two "precise" end-points.

There are other places in your posts where this conflict becomes apparent in wanting to use the precise measurements of what the camera captured to estimate automatically what YOU OBSERVED. I tried to point one of those out before:


In my case, the face or whatever under the pink light or golden hour light on the hair or summer white dress, or red rocks of the canyon etc etc, non neutralized but instead showing / representing the Cast we observed


In this statement you are explicitly rejecting one of the "precise" end-points, the neutralized one (the one where the gray card is balanced completely to a measurable neutral gray). You are also implicitly rejecting the Daylight end-point because you are after the "cast we observed". So how do you get closer to the "cast we observed" than either of the two end-points? That's the problem. There is no easy, scientifically precise way to determine "what we observed". We have numerical values for what the camera recorded but we don't have numerical values for what you perceived.

As I mentioned earlier, you could do a series of exercises to "profile" your perception but I doubt if it would be worth the effort. You could try to control the variables as much as possible - recording your perception off a gray card, for example - but I doubt the exercise would give you much help for normal shooting.

Finally, let me convert your stage example to a situation that presents the same, exact challenge but in a form that we as photographers usually see on a more common basis: indoor incandescent lighting.

I take a photo of a gray card indoors under incandescent lighting. I can then adjust the white balance of the gray card to produce a neutral gray. There are many times when this is perfectly acceptable but what if I want to convey a bit of warmth that I "observed"/"felt" when I took the image? I can check what the image looks like with a Daylight white balance but that image will likely look TOO warm. The Daylight setting looks too warm because your eye/brain combo did its own white balance to some (unknown) degree. So what do I do? I subjectively adjust between the two points to get something that "feels" right or gives the image the tint that I WANT it to have. Again, I can measure the color temperature of the captured image and I can identify another particular color temperature like Daylight but what I DON'T have available to me is a PRECISE MEASUREMENT OF MY PERCEPTION/FEELING/ARTISTIC VISION.

If you still feel that there is some scientifically precise way to get what you're after, I encourage you to read a few articles on the topic to get a feel for just how difficult an endeavor it is. Here are two articles to get you started. The first talks about some unexpected results with museum lighting and the second talks about the challenge of obtaining "true color" on the Mars Rover missions. That particular article finishes up with the following quote:

“What we’re doing on Mars is really just an estimate, it’s our best guess using our knowledge of the cameras with the calibration target. But whether it is absolutely 100% true, I think it’s going to take people going there to find that out.”

http://www.solux.net/cgi-bin/tlistore/infopages/eyes-response.html
http://www.universetoday.com/11863/true-or-false-color-the-art-of-extraterrestrial-photography/










May 13, 2013 at 11:24 AM
gdanmitchell
Online
• • • • •
Upload & Sell: On
p.3 #9 · Shooting in golden hour WB question


I'm late to the party and I'm not going to directly answer your question - but I'm hopeful that some of this might be a bit useful.

Regarding color balance for scenes that are intensely colorful and especially in which the light's color and/or the overall color of the subject is very strong in one channel, using jpg and a in-camera color setting is not the best approach. The camera is going to attempt to essentially move the overall color average of the scene towards neutral, and this creates problems when the scene is actually far from neutral. Given the variables in the light itself and in how you might interpret it, here it is almost always generally better to shoot in raw and save the decision making for the post-production phase where you have the most control and still have full image data in the raw file.

Determining the "right" color balance is different than determining the "accurate" color balance, and even the latter leaves you with at least two choices.

If you are going for supposedly accurate reproduction - which I do not generally recommend except in a few limited circumstances - you could either go the gray card route and attempt to render something that is known to be gray as an equivalent gray tone or you could presume that your camera's raw mode renders an "accurate" version of the coloration that was there.

There are problems with both of these, and often the best results is somewhere between the two.

The problem with the gray card is that it essentially eliminates the interesting colorations that probably made the scene interesting in the first place! If you shoot in warm "golden hour" light and then shift everything so that your gray card is gray... you just eliminated that lovely, warm light coloration that drew you to photograph the scene in the first place. Yes, you'll have neutral colors - white will be white and gray will be gray and, hopefully, other colors will come along for the right - but it will look like you shot in neutral light rather than warm light.

The problem with going with the "accurate" rendition of raw mode is one that we're all familiar with if we have shot subjects, say, in shadow and illuminated by open sky. In the actual presence of such a thing, the scene looks fine to us - but when we look at the resulting photograph the coloration from the sky or other sources seems overdone and grotesque. (Anyone who has shot snow in shadow is familiar with the actual sky-blue color of snow!) On the scene, our visual system compensates and reassures us that blue snow is actually white, but in a photograph this breaks down.

The "best" outcome is perhaps somewhere between these two extremes. You don't want to neutralize the natural coloration of the light that gave the scene its character, but you also don't usually want blue snow. So, as with some many things in life, the answer is some sort of subjective compromise between the two extremes. In many cases I want to move the color balance away from the supposedly accurate tones of the raw file, but not all the way to the neutralized coloration of the gray card version.

How to get there? There are many techniques that can work, and assessing which one is right is a personal and subjective matter. I have seen people use the gray eyedropper tool in a curves layer and click around the image until finding something that seems like it should be gray. I've been known to add an "average" layer in photoshop, click on that with the gray eyedropper, delete the gray layer, and then adjust the layer opacity to taste. In essence, what one frequently does is move the balance from the raw color toward but not too the gray card value, making careful (and experienced) but subjective judgments about how much is just right. Two photographers will come to different conclusions.

Sorry for the longish answer, but it isn't really a simple issue and it is not susceptible to an automated methodology.

Take care,

Dan

gschlact wrote:
Quick question for the group on what result you would expect to see using the following lighting scenario and WB setting method (custom).

Say you took a photo of a proper gray card under weird lighting such as golden hour on the field, or unicolor stage lighting. In that specific lighting, you then on the Canon Camera set the White Balance (WB) to use that customer setting of the gray card shot you just took.

Now, say under that same lighting (call it pink), if you took a jpeg photo of a pure white sheet of paper and displayed it
...Show more



May 13, 2013 at 01:48 PM
austin.grant
Offline
• •
Upload & Sell: Off
p.3 #10 · Shooting in golden hour WB question


This is precisely why I always carry my hydrogen spectrum tubes, power supply, diffraction grating and h-alpha filters with me on every shoot. I mean, sure, the 5,000 volt power supply is tough to run in the field, but my clients really appreciate that fully-accurate depiction of their special day.

The ever-changing light conditions mean I'll have to have the gas tube in each shot, but most customers don't even notice the bright pink glowing tube in their images. I for one can tell you that I don't mind the extra effort of calibrating each image individually....

Shoot RAW, use AWB and adjust to taste in post. That's all you can reasonably hope for. Even if you did get it "perfect" it wouldn't be that way on a different monitor. Prints won't help, as I may have it under tungsten lighting or in a room with diffuse sunlight. What time of day do I look at the print?




May 13, 2013 at 02:25 PM
 

Search in Used Dept. 



Photon
Offline
• • • • • •
Upload & Sell: On
p.3 #11 · Shooting in golden hour WB question


austin.grant wrote:
This is precisely why I always carry my hydrogen spectrum tubes, power supply, diffraction grating and h-alpha filters with me on every shoot. I mean, sure, the 5,000 volt power supply is tough to run in the field, but my clients really appreciate that fully-accurate depiction of their special day.

The ever-changing light conditions mean I'll have to have the gas tube in each shot, but most customers don't even notice the bright pink glowing tube in their images. I for one can tell you that I don't mind the extra effort of calibrating each image individually....

Shoot RAW, use
...Show more
In the spirit of Austin's post, let me suggest that every photographer who wants to master digital color reproduction should train by shooting exclusively under a mix of sodium and mercury vapor lamps, as well as each alone. Between the incomplete spectra and the rapid cycling of color temperature, one quickly learns how to deal with the problem: by relinquishing all hope of a perfect and repeatable solution.



May 13, 2013 at 02:41 PM
Steven W
Offline

Upload & Sell: Off
p.3 #12 · Shooting in golden hour WB question


I could not read through all these posts, so sorry if this has been mentioned, but I have used the flash white balance setting with decent results. YMMV


May 13, 2013 at 02:46 PM
Fred Miranda
Offline
Admin
Upload & Sell: On
p.3 #13 · Shooting in golden hour WB question


Steven W wrote:
I could not read through all these posts, so sorry if this has been mentioned, but I have used the flash white balance setting with decent results. YMMV


I'm not surprised as flash WB is pretty close to Daylight and the custom 5,500K I use as a preset.
Fred



May 13, 2013 at 04:21 PM
artd
Offline
• • •
Upload & Sell: Off
p.3 #14 · Shooting in golden hour WB question


After reading some of the intricate discussions and convoluted propositions, I guess I would ask:

If you want to make sure you keep the color cast from golden hour what is wrong with just setting your WB to daylight? Perception is always going to be relative, and if you are going to be perceiving a warm color cast, it is going to be relative to what you would perceive during daytime.

If after the shot, you think that your white sheet of paper needs to be more or less pink, because that's how you remember it, then adjust the colors in post production to make the sheet more or less pink.

Chasing forensic accuracy is not really as important as matching your perception. Because our brains do not operate as rigid, forensicly accurate devices. Our brains reconstruct a perception of reality by interpretation.

So in the end, interpretation is always the photographer's job, not the camera's. Adjust the scene until it looks right to you. If it looks right, then it is right.



May 13, 2013 at 05:11 PM
AJSJones
Offline
• • •
Upload & Sell: Off
p.3 #15 · Shooting in golden hour WB question


Monito wrote:
Such spectral analysis is not feasible for photographers. Not the least of the problems taking that approach is the fact that the light we need to measure is not generated (as by a star) but is reflected off skin, clothing and other objects. Not only do we not know the spectra of the light source, but we don't know the absorption, reflection, and retransmission spectra of the objects (three different types of spectra in play). There are physical measurements that can get around such problems, but not even for your average pro.

So the astrophysical analogy is not applicable.


I wasn't in any way suggesting we take pictures with spectrophotometers
It was a scenario in which accuracy of photon recording (regardless of how they were generated or absorbed etc), and not artistic perception and post processing style etc, are paramount, as an illustration of the "forensic" or "objective" recording Guy was interested in, so others could appreciate his question a little better than it seemed they had.



May 13, 2013 at 09:31 PM
anscochrome
Offline
• •
Upload & Sell: Off
p.3 #16 · Shooting in golden hour WB question


I have been handling outdoor WB like this ever since I first owned my digicams. Ia sked myself-what did I like for color balance? The answer was-5500K, just like daylight slide films of the past. So, I always shoot in AWB outdoors, and simply select "daylight" balance in LR, since it defaults to 5500K when "daylight" is selected. It also adds +10 magenta to the tone, so I scale this back to +5 or zero, depending on the scene. Done. Now my "golden hour" shots always look correct for what I am used to with 5500K slide film.

I only use custom WB when shooting under artificial light conditions, and then it usually needs some tweaking in post.



May 14, 2013 at 01:52 AM
ipronoun
Offline

Upload & Sell: Off
p.3 #17 · Shooting in golden hour WB question


Here's an idea that I haven't ever seen suggested (apologies if anyone mentioned it earlier in the thread):

Take a picture of your final output print (or whatever your target output medium is) using the same camera and same WB that you used to take your original image. Now you have two images: image A (original) and image B (picture of print). The color differences between image A and image B that are now recorded in your camera are a measurement of the difference in color between your original scene and your print. So use Photoshop to take the difference in color (average color should suffice) between image A and image B, and apply the opposite of this difference to a copy of image A (called image C). Now make a new output print based on image C. The color balance of this new output image should be close to the color balance of the original scene. In practice, this process would have to be iterated multiple times to get it right.

Once you start thinking about this method, lots of caveats immediately come up. First, this is a attempt at reproducing the objective color balance of the original scene using the available equipment (a camera, not a spectrophotometer) in an output medium. It won't necessarily reproduce anyone's subjective experience of the scene. Second, it's highly sensitive to the output medium (as are all color balancing methods), and if you change, e.g., the lighting of your print, you'll reduce the accuracy. Third, the whole chain of color measurement (camera + lens) must remain the same when recording image A and image B. Fourth, the color gamut of your output medium probably can't match that of your camera, so the adjustment will always be an approximation. There are probably other issues that I haven't thought of.

Would I ever try to do this? No way. I'd rather simply adjust the WB of Raw files until they look nice. It's an interesting thought experiment, though. I am a little curious to see how well the method would work.



May 14, 2013 at 07:58 PM
gdanmitchell
Online
• • • • •
Upload & Sell: On
p.3 #18 · Shooting in golden hour WB question


I think your second paragraph begins to get at why this will not be useful as more than a theoretical concept. But in reality... no. ;-)

And "accurate" reproduction is rarely the real goal, it turns out - with the exception of some things like product photography. If "accurate" reproduction were the measure, much (most?) photography regarded as great would have to be dismissed.

Dan

ipronoun wrote:
Here's an idea that I haven't ever seen suggested (apologies if anyone mentioned it earlier in the thread):

Take a picture of your final output print (or whatever your target output medium is) using the same camera and same WB that you used to take your original image. Now you have two images: image A (original) and image B (picture of print). The color differences between image A and image B that are now recorded in your camera are a measurement of the difference in color between your original scene and your print. So use Photoshop to take the difference
...Show more



May 15, 2013 at 12:36 AM
ipronoun
Offline

Upload & Sell: Off
p.3 #19 · Shooting in golden hour WB question


gdanmitchell wrote:
I think your second paragraph begins to get at why this will not be useful as more than a theoretical concept. But in reality... no. ;-)

And "accurate" reproduction is rarely the real goal, it turns out - with the exception of some things like product photography. If "accurate" reproduction were the measure, much (most?) photography regarded as great would have to be dismissed.

Dan


Yes, I agree that accuracy is rarely a goal. (Beginner that I am, I don't think I've ever taken a picture where I cared about truly accurate color rendition.)

I saw the OP as an interesting puzzle: how to represent the colors at a scene, including color casts due to incident lighting, accurately in a final output, using only ordinary photographic equipment. (Certainly there's scientific equipment that can do the job.) The usual photographic techniques that I'm aware of don't really address this, so I came up with a method that, at least in theory, does address it. Whether it works any better than just shooting with daylight white balance and displaying the result on a well-calibrated output, I have no idea.

Anyhow, thanks to gschlact for asking an interesting question. It got me to register here, so hi everyone!



May 15, 2013 at 03:51 AM
Gunzorro
Offline
• • • • •
Upload & Sell: On
p.3 #20 · Shooting in golden hour WB question


ipronoun -- Welcome to the forum and thanks for your thoughts here and the earlier thread.


May 15, 2013 at 05:10 AM
1       2      
3
       4       end




FM Forums | Canon Forum | Join Upload & Sell

1       2      
3
       4       end
    
 

You are not logged in. Login or Register

Username   Password    Retrive password