White Balance

[McGizmo]

Hi guys,

I figured this would be a good place to post about a simple tool I just came up with out of need. I shoot a lot under water and the water as most of you know is a strong filter on sunlight and its effect changes all the time and varies on the distance the light travels both to the subject as well as then reflected to the camera.

I now use a Nikon D300 which has the ability to calibrate and store a custom white balance setting based on a test shot but this is problematic for me for a couple reasons. One, I have to learn how to do it and go about the calibration with the camera in a housing and me bobbing around in the water. The real problem though is that the calibration would likely be way off because I would need to know the depth of the subject below the water and my angle and distance from it and replicate this geometry in my calibration.

What I have been doing is using the white balance adjustment available in PhotoShop CS4 in post production. Auto White balance seems to only go to 7500k and a 30% tint adjustment which is often not enough due to the loss of the high end of the spectrum. If there is something in the image I know to be white or near white, I can sample that and have the software calibrate. I usually end up manually tweaking the results further to get what I feel is representative of the true colors. However, there is a difference between what I actually saw and perceived and what I would have seen had the subject been bathed in full spectrum sunlight.

The human eye can adjust to the actual circumstances better than the camera unless you give the camera added information. In post production you can go further then the eye and enhance the image greater than the reality allowed you to see.

So what?

Well recently, I noticed that a coral land mark that I use and recognize out on the reef had changed. I came to think of this coral colony as "Big Blue" and there was a nearby colony I considered "Little Blue". The other day I was swimming south and saw Little Blue to my left so I looked to my right to see Big Blue and this also gives me an idea of the visibility that day because sometimes you can't see both colonies if the vis is restricted. Well I saw Big Blue but it wasn't blue?!?! Big Blue has been blue for as long as I remember it and the type of lobe coral it is is a very slow growing coral. I would guess it is older than I am. Something was amiss!!

I have met a number of marine biologists who are presently studying this reef and monitoring it in regards to algae change and coral health. There is a large sanitation plant not too far from here and they use an injection well to pump the processed waste water into with it ultimately seeping into the ocean and out in areas of the reef. The problem is not toxins as I understand it as much as an issue of excessive nutrients which encourage algae growth and bloom.

I contacted one gal who working with others on studying the reef and she is concerned with the images I provided her. Tomorrow I will meet her and possibly others and show them these colonies. I also told her I would be happy to include monitoring these colonies in my near daily swims and photograph them to allow for documentation and following any changes that might come about.

Finally to the point!

I realized that I needed to be consistent in how I rendered the images and color should be based on a better reference than my recall. To wit, I took some MCPET and cut a 3" x 3" square and punched a hole in one corner so I could attach a clip to it. The MCPET is rated at greater than 90% reflectivity of the visible spectrum and I expect it is about as good of a white target as one can come by; especially for under water use since the MCPET is impervious to the water and easy to wipe clean. I mentioned to my friend that I would use this target as part of the image or at least shoot it in the same geometry as that of the colony and use it in post production to adjust the white balance as best I could. Within reason, this should bring some consistency to the images from one day to another and hopefully mostly independent of the oceans filtering characteristics at the time. She thought this was a great idea and also liked the idea that the card, being of a known dimension, could lend some scale to the image. I am going to give some of these to these reef researchers if they want because it might serve them as a viable tool as well.

Below is a shot of "Little Blue" as the camera recorded it with white balance set on sunlight (the setting I use by default).

The coral colony is about 15' down and the sun was directly overhead at the time. The camera is probably 2' off the coral. You can see the MCPET square in my hand.

In PhotoShop CS4, I used the white balance eyedropper in the raw preview screen and clicked it on the MCPET square. The software immediately adjusted the white balance based on this and rendered the following image:

There is a thread in the McGizmo forum about color rendering and the Colorchecker is mentioned. There is much discussion on various light sources and how they effect color rendering.

The further light travels through water from the source to subject and then to viewer or camera, the worse the color rendering becomes. In the past, under water photographers relied heavily on artificial and bright sources of light close to the subject as well as camera close to subject to minimize the loss of spectrum as filtered and dispersed by the water.

The white card can aid correction in post production quite effectively provided the distance the light traveled through the water hasn't filtered the light too much.

On land, if the light source is unknown or there are a number of light sources of varying spectrums, a white card can also be used in post production to "normalize" the image.

I would also imagine that if one preferred a cool or warm tint that they could use an off white card to consistently alter the balance in favor of their preference.

I have read comments from some of you pros that you don't care too much about the CRI of the light source because you can make corrections in post production. This new task I have taken on of monitoring some of the coral colonies has put me in a similar situation in that I have to rely on ambient light which is not consistent because of how the ocean will filter it. In hopes of normalizing the images and color, I think the white card will serve me well and allow for better correction than my guessing would.

So that's my long winded story.

I suggest this thread can cover any and all comments and observations members wish to bring forth in regards to white balance adjustment; either pre or post production and when or why it might be of importance or consideration.

EDIT: I guess I should add a pic of Big Blue and you can see why I have become concerned. There is a fine film of a brownish algae covering the entire colony and what used to be a really striking blue now looks like the color adjusted image below:

You can also see in the shot of Little Blue that there are some areas where this algae is starting to gather and alter the color.Whether this algae presents a threat to the coral or not is still not known to me and that of course is the real concern. Hopefully the experts will view the colonies and tell me I have cried wolf needlessly. :tinfoil:

 

[csshih]

hmm, I've never thought to use a white card in the shot itself! thanks for the info.
can you just use a white piece of paper? those cards are expensive!!!!

about the coral: I'm not going to comment because I have no idea.

 

[McGizmo]

A white piece of paper would be fine. I should point out that you could drape it over the subject and then take your shot with the subject covered and then another shot with the paper out of the pic. If you take note in PhotoShop of what the white balance correction is based on the paper, you can apply the same correction to the image that doesn't have the paper in it.

 

[McGizmo]

I came across a new Hawksbill turtle on the reef today and I used the MCPET card in one image to get the callibration to use in subsequent images taken without the card.

She was down about 20' and the raw shot based on sunlight was all blues and greens. I needed to bump the color temp up to 14750k and the tint up 130% to get what you see above. (incidentally the card does look white after the adjustment and that's always nice reassurance)

I read on a Backscatter article that some company is now selling white swim fins that sounds like a great way to have your white sample handy or footy, as it were.

 

[blasterman]

Yup, the way to do this is with a white/grey patch and shoot RAW.

Years ago when we shot film I had several clients that were weekend divers and would bring me film from all over the world to balance and print. Shooting available light in the ocean involves an extreme loss of red light, and after screwing around a lot the only way I made it work was have the client shoot high speed print film and over-expose by several stops to give me enough film density to correct. The fun part was when they didn't have a neutral target and I had to guess what things were supposed to look like.

Slide film was only useable *if* they were using strobes. No color lattitude with slide film.

Note that the white card only works to an extent because while it aids in balancing, you can't replace colors that aren't there and the image starts to look monochrome. The less water between your target and the camera the better.

I keep a reef tank, and years ago I was much further into the hobby. However, if I saw hard corals that looked that bad along with a coating of algae I'd nuke the tank and start fresh. Yeah, it's usually excessive nitrates or phosphates causing the imbalance. Both are good fertilizer - if kept on land!

Be interesting if you could follow up what happens with this!

 

[McGizmo]

Yup, the way to do this is with a white/grey patch and shoot RAW.

Years ago when we shot film I had several clients that were weekend divers and would bring me film from all over the world to balance and print. Shooting available light in the ocean involves an extreme loss of red light, and after screwing around a lot the only way I made it work was have the client shoot high speed print film and over-expose by several stops to give me enough film density to correct. The fun part was when they didn't have a neutral target and I had to guess what things were supposed to look like.

Slide film was only useable *if* they were using strobes. No color lattitude with slide film.

Note that the white card only works to an extent because while it aids in balancing, you can't replace colors that aren't there and the image starts to look monochrome. The less water between your target and the camera the better.

I keep a reef tank, and years ago I was much further into the hobby. However, if I saw hard corals that looked that bad along with a coating of algae I'd nuke the tank and start fresh. Yeah, it's usually excessive nitrates or phosphates causing the imbalance. Both are good fertilizer - if kept on land!

Be interesting if you could follow up what happens with this!

Bold emphasis by me. Thanks for chiming in! In the image below, my white card was the underside of the spotted eagle ray. I think it is clear if you look at the reef below and then follow it as it gets further from the camera that your color degrades as mentioned to that of a monochromatic image.

Although my example and area of interest expressed here has been UW photography, the white card can be used topsides as well because there are variations in ambient light, natural or artificial that it can aid in for color correction. Again though, the full spectrum needs to be there to some extent or the software can't emphasize and balance it.

In trying for white balance in images I have taken of whales, I have found the software has to pump so much red into the image that it looks terrible and I have to compromise and go back to an overlying blue tint. The digital cameras can record more information than the naked eye can see so you can at least adjust the balance to what you saw and in many cases like the image above, improve on the image with better color than seen by the naked eye.

Again, back on land, I often shoot pics say of flashlights in the shade so I can reduce the contrast and shadow. The light present here is a function of reflected light off buildings, plants and indirect sky illumination. The full spectrum is likely there but not in proportions to direct sunlight. A white card would be quite effective in adjusting the balance.

 

[Archie Cruz]

Couldn't read the whole post but as a former pro shooter, I can pretty much figure out what it said.
Sure PS can kinda neutralize a heavy cast ( cyan/blue) by simply adding red, but we all should realize that nothing, not even WB under water will invent separation of color that you can only achieve through the introduction of artificial light which allows for natural portrayal of colors. I note that your CS color corrected frames are simply monochromatic or nearly so. Underwater or any heavy cast WB is just an exercise in neutralization.
I know you're a pro under water so I await your response.:thumbsup:

 

[McGizmo]

Archie,
Thanks but I am not a pro at anything and certainly don't want to come off as such! :duck:

Sorry you couldn't read the whole post and sorry I couldn't be shorter and more concise in making it.

I don't think I understand what you mean about inventing seperation of color underwater. What I can see is the ability of PhotoShop to take an image that is heavily tinted and near monochromatic as shot and render it with more colors and in some cases go beyond what I actually saw and enhance the colors even more. These colors approach what one would see if the water were replaced by air. If this is an invention based on information recorded but in need of expansion or magnification, so be it.

If I were in need of providing my own artificial light in most of the shots I have taken, quite simply, I never would have got the shot. Any artificial light brought to bear will be filtered and degraded by the water just as much as the sunlight. Granted if you have enough light and it is close to the subject as is the camera, little or less color correction or white balance correction is required.

As a former pro shooter, perhaps you can take on the subject of white balance and either support, clarify or illuminate the rest of us on when or if it is something to consider in our photographs. I am under the impression that it is a tool available and having merit.

 

[blasterman]

Is it just me, or is there any growth at all in the corals the ray is swimming over? Looks pretty dead....but it might just be the color.

 

[Wattnot]

+1 for those suggesting shooting in RAW mode. Take a shot of your white or grey reference card when you're on site, then you should have unlimited control over your white balance. Editing the jpgs the camera makes is limited. Raw is unlimited.

 

[McGizmo]

Is it just me, or is there any growth at all in the corals the ray is swimming over? Looks pretty dead....but it might just be the color.

I am no expert but I would guess that about 50% of the reef in the image is live coral and the balance is either bleached or now covered in algae. The competition for space is always going on and unfortunately the added nutrients from runoff and injection wells give the algae an advantage.

I am also on the band wagon for shooting in RAW.

 

[Archie Cruz]

Sure. Here goes.
Rule #1. Once a color cast has been introduced into the scene it has forever crushed the chroma ( color scale ). Hence the 'CAST'.
Photographing a scene with a cast 'records' the scene with the color gamut diminishing cast on it.
Filtering a cast with a filter of complimentary color (Red for Cyan, Yellow for Blue, Magenta for Green) will only NEUTRALIZE the cast, not remove it. Note that your before pictures of the Ray, start out cyan cast and your filter (actually white balancing) simply neutralizes it to a brownish hue. So all that you've done is to trade one cast for a more pleasant ( less offensive ) one instead.
The full color gamut however, is gone forever. In addition to being a former shooter, I am a pro Photoshopper (no brag, just fact). I've been paid richly to colorize images and that stems from my old habit of hand coloring prints with oil -(before computers).
But make no mistake. I'm not retrieving lost colors when I colorize, I'm simply adding them in artificially.
I started shooting some underwater pics for a test of the Flip Mino HD and got what I'd predicted- cyan cast. I know I can help some by adding a red filter but it will NOT retrieve the lost color absorbed by the body of water between my camera and the subject.
If I use an underwater HID or LED that's powerful enough, I can essentially cancel out the cast with daylight or a tweaked daylight.

This is not my theory by the way, the rule has been in place since optics and sensors of all types have been in use.

Google underwater pictures or videos for some real examples.


Oh. on the subject of white balancing. This is a function of the electronic system of a video camera ( still digital is a just a frame). It is a calibration of a specific white in the scene to eliminate bias. It delivers a REFERENCE color that is theoretically devoid of influence ( casts ) . Doing this is critical before all shooting ( off the white card ) so as to calibrate the sensor with reference to the scene. I actually have devised a trick to cheat the sensor into calibrating itself to somewhat warmer rendition than you would get off a pure-white card. Underwater, all it does is key the sensor to deliver- GRAY or warm GRAY.

NOTE:Forget the screw-on filters that purport to "white balance" a scene- all those do is average out the scene! white balancing is for the specific subject at a specific location WITHIN the scene.

Do this test. Take an underwater shot of a color chart before and after white balancing. The before will be a color chart with a a cyan cast and the after will have a no cyan in the whites but will be de-saturated in the rest of the chart. Now. Take that SAME chart out of the water and shoot it at the same distance after white balanncing and WOW! Now you see all the color differentiation pop out at ya!

Archie,
Thanks but I am not a pro at anything and certainly don't want to come off as such! :duck:

Sorry you couldn't read the whole post and sorry I couldn't be shorter and more concise in making it.

I don't think I understand what you mean about inventing seperation of color underwater. What I can see is the ability of PhotoShop to take an image that is heavily tinted and near monochromatic as shot and render it with more colors and in some cases go beyond what I actually saw and enhance the colors even more. These colors approach what one would see if the water were replaced by air. If this is an invention based on information recorded but in need of expansion or magnification, so be it.

If I were in need of providing my own artificial light in most of the shots I have taken, quite simply, I never would have got the shot. Any artificial light brought to bear will be filtered and degraded by the water just as much as the sunlight. Granted if you have enough light and it is close to the subject as is the camera, little or less color correction or white balance correction is required.

As a former pro shooter, perhaps you can take on the subject of white balance and either support, clarify or illuminate the rest of us on when or if it is something to consider in our photographs. I am under the impression that it is a tool available and having merit.

 

[Archie Cruz]

Some tips for a brother of mine that's going to use my Mini HD for UW video. Housing is Aquapac #420.
Hope you all find this useful...

 

[McGizmo]

Archie,
Thank you for your response. You are the expert and you have the proper terms to describe what is happening. I realize that when I shoot underwater that I do not have a full spectrum of ambient light available. The closer to the surface and the closer I am to the subject, the better the lighting. Sun overhead and clear water also makes a big difference in the spectral fullness. Often the high contrast between that directly lit by the sun and that in shadow or illuminated by bounced and scattered light is problematic in its own right. An artificial full spectrum source close to the subject and with minimal distance to travel after reflecting off the subject will certainly provide for a more full spectrum recorded image. If the source is intense enough, it can significantly override the cast present from the ambient.

I don't dispute your comments but I want to point out that I am free diving and swimming with a sea anchor of a UW camera housing hanging about my neck. I am often tasked with as much speed as I can muster and attempting to photograph small critters up close or large critters at greater distances. I don't have a pair of large strobes on arms attached to the housing and choose not to because the added drag would really impair my mobility. The strobes would in some cases annoy the critters and reduce my contact time with them and in cases where the critter is relatively large and at some distance, the backscatter would add unpleasant and distracting noise to the image.

You say the full gamut is gone forever and being uncertain of the term, I won't dispute this. I will however claim that there are reds and yellows and greens present in some images that to the naked eye seem monochromatic because of the strong cyan cast. It has been my experience that using the white balance in photoshop, post production on a raw image, can reduce the cyan bias and bring out these lesser in magnitude colors. The result is an image that is more pleasing to the eye because it is more in keeping with what you see (saw).

In the image below, the sun was pretty much overhead and the water was relatively shallow. Additionally, the landscape (sand) was reasonably reflective and provided indirect lighting for the undersides of the rays; diminishing the shadow.

The image above was white corrected in Photo Shop using the white underside of the rays as the white sample. The shot was taken with the camera set at a sunlight white balance (5000k) and prior to the white balance in post production, a near monochromatic cyan cast was what you saw. Had I used artificial illumination in this shot, I would have backscatter to contend with, the subjects would be lit from an unnatural angle and I would still have the distant background monochromatic relative to the foreground and subjects. Below is the shot recorded by the camera prior to white balance adjustment:

The point I was trying to make in the initial post is that by using white balance in Photo Shop, one can enhance the image significantly in terms of bring out some colors and a color balance more in keeping with what one would see under more ideal, full spectrum illumination conditions. I still feel this is a reasonable proposal and for me, the comparisons of the two images here give justification to this "claim". I won't disagree that a better image could be had with the introduction of additional and artificial light but impracticalities aside, does that diminish the merit of post production white balance as illustrated in this example?

I would also like to point out that most under water photography and video is shot with super wide angle lenses allowing one to get as close to the subject as possible and minimize the water between the camera and subject. This inherently distorts the view and angles noticeably and contrary to what one actually sees! It is not cinema verite or photo verite. :shrug:

The rays are probably a bad example because they have so little color themselves. Below is a different example and one where the sun has traveled about 20' down through the water before reflecting off the frog fish and myself. The frog fish is a yellow and you can see in the uncorrected image that the yellow is still present and not lost. However, the cyan cast you mention has certainly put a strong bias on the image. In the version where I have corrected for white balance, I believe the colors are more in keeping with what one would see as our eyes correct for white balance as well.

Camera set at manual white balance of 5000k:

Image white balance corrected in Photo Shop based on corrections taken from an additional shot of a white card:

Again, the point I wanted to make with this post is that of considering white balance correction as a means of basically correcting the white balance! I believe it has merit in certain circumstances and situations. :shrug:

.........

....Take that SAME chart out of the water and shoot it at the same distance after white balanncing and WOW! Now you see all the color differentiation pop out at ya!

I am glad you made this comment because the thrust of this thread was to encompass more than just under water photography. I used UW images as an example but there are many cases where one is shooting in indirect light from the sun where the indirect light has been biased by the objects it has bounced off of, prior to reflecting off the subject of interest.

If you don't have a full and balanced spectrum of light present, your subject will not reflect light and be recorded in an image as if it were illuminated by full spectrum light. I believe that is your point? I concede this, at any rate. You are an expert at Photo Shop. Out of curiousity, do you not use white balance correction in Photo Shop? Do you ever use a white card as a target for white balance correction? To correct the camera prior to the shot? To correct the image, post production?

I should add that I don't bother with a white balance correction in the camera itself because I never know ahead of time, what the depth or what the distance from the subject is going to be.

EDIT: You have stated that the ideal solution is to introduce artificial, full spectrum light to the subject. I agree this is ideal if you want to show what the subject would look like under full spectrum light and provided you have the means of providing significant artificial, full spectrum light, on the subject. Such an ideal is by its nature not realistic nor is it often in practical terms, realistic.

I recall a comment once made by a friend looking at photos in a Scuba Diving magazine. The comment was that they had never seen the reef or fish look like the image. The reason was simple. The friend had never dove with an intense, full spectrum light source in hand. By the same token, unless an artificial light source is used or there is some post production correction, a recorded image will not look like what a person sees either. The recorded image is too washed out of color.

 

[blasterman]

I should add that I don't bother with a white balance correction in the camera itself because I never know ahead of time, what the depth or what the distance from the subject is going to be.

I hate to keep repeating myself, but I need to state again that color temp can be ignored in camera and adjusted after the fact in any number of RAW converters. Shooting JPEG and correcting in Photoshop is a very clunky way to accomplish this. We are no longer shooting film or using Grandpa's 1987 VHS camcorder

It should also be noted that while adding a full spectrum light source makes the color pop out better and draw more attention, we are in reality adding color spectra that doesn't exist at that depth of water and only doing so for our aethestic needs. The corals and marine life are doing fine without it.

So, on one hand it's kind of cool to see corals and marine fish light up in day-glo colors when hit with artificial light sources. The reality is that they're producing these insane colors because there's no evolutionary adaptation to something that doesn't exist in their environment.

 

[McGizmo]

Blasterman,
I don't know if you are agreeing with my comments or taking exception to them. To be more clear at my end, for starters, I am shooting in Raw and agree with you completely on its advantages. When I do white balance correction in Photo Shop (CS4), both the color temp and "tint" are adjusted.

Someone in one of threads took one of my UW images and did an auto white adjustment to it in Photo Shop. I have found that this can work in some cases but it seems that it has a limit or ceiling of adjusting the color temp up to but not beyond 7500k and the tint up to I think it was 35%. I have found that when I balance off the white card, the color temp is pushed up to 13000 and beyond and the tint can approach 150%. At depths below 25' or when it is cloudy or murky underwater, the attempt to balance to the white card falls short and obvious color problems arise. At this point, I believe the available light is just too deficient in terms of the full spectrum, for a white balance correction to give you a viable or believable result. The white card was simply not white based on the available light and pushing it to white in the image is beyond the fidelity of the raw image, if I can get away with using that term.

 

[IMSabbel]

I still dont understand why you do the whitepoint setting in photoshop.

Any good raw converter will allow you to set the whitebalance directly, which gives you another 2-3 bits of dynamic range to get some more colours out of.
Even if you are working in 16bit /channel in PS4, the Raw converter can bias the demosaic process, so it still has an advantage.

 

[Mike V]

If you are trying to get consistent colour then a white card isn't enough.

You should shoot a white, black and 18 % grey cards.
There are lots of commercial versions.
A funky solution for example is the Datacolor SpyderCube, but just three patches on a card work fine too.

Then in Photoshop you can use the white, black and mid-grey eyedroppers on the cards and your white point, black point and contrast is normalised.

 

[LuxLuthor]

Archie, what a gift your posts were! :bow: :bow: :bow:

I have already :bow: to McGizmo!

 

[McGizmo]

I still dont understand why you do the whitepoint setting in photoshop.

.....

Actually I use the white balance adjustment in "Camera Raw" which I get to from Adobe Bridge and prior to opening the image in Photo Shop. Perhaps my bad for just saying it was done in PhotoShop. I think this may be what you are commenting on?

MikeV,

Thanks for the tips! I'll look into these. I thought about buying one of the small ColorCheckers but I doubt they are intended for submersion in salt water. :green:

Frankly, I am very satisfied with the enhancment I am getting just using the white balance but it would be nice to get even better results by tempering the adjustment with the black and grey as well. Of course as Archie has pointed out, if the colors aren't there to start with because of the cast, no correction will be viable.