DON'T MISS: Italian conceptual portrait photographer Sara Lando is coming to the US to teach in Atlanta (8/16) and Baltimore (8/23). Highly recommended.

Tuesday, October 06, 2009

A Simple Idea to Improve Flash Photography

This is something that has been bouncing around in my mind for the last six months. It's simple, yet could improve the flash photos of every pro (and serious amateur) shooter.

The beauty is, it even could be done retroactively for pro and prosumer cameras already in circulation via a simple firmware upgrade. And it could help Nikon flash shooters, whether they use SBs, ABs, Profotos or whatever.

How to integrate white balance and flash, inside.
__________


So, Here's the Idea

Both Nikon and Canon (and Sony, Pentax, Olympus, Holga and Lomo, for all I know) do a very good job of manual white balance, based on the ambient environment. If you are shooting available light, you just shoot a white (or grey) card and set a new white balance to match your mystery ambient light. That'll get you pretty darn close.

But with flash you 're screwed if the weird ambient is not daylight, incandescent or "30CC green" fluorescent. Because whatever weird white balance you shift to is gonna leave your flash out in the cold. Or the warm. Or the too cyan-ish magenta. (You get the picture.)

It shouldn't have to be that way. Since the camera can balance in just about any color of light, it knows the exact difference between white light and your ambient environment. Wouldn't that be a handy little piece of info to have at your disposal?

It would be a simple, in-camera calculation to convert that offset into a color-correction (CC) filter pack. Then the light coming from your flash would be color-matched to the weird, ambient environment. Now, your camera's white balance corrects for everything. And Rosco already makes that gel pack. (It's about $45 for enough material to last you forever with speedlights.)

Imagine walking into a mystery-vapor high school gym, doing a quick white balance and being able to gel the strobes to exactly match the ambient color. I'd be all over that. And this would be especially sweet, now that fluorescent light color temperatures have all gone to hell in a handbasket. Adding 30CCs of magenta doesn't correct for jack anymore.

So, Nikon -- you already have the ambient offset color information available in the camera, and it works great. How about you just give us a downstream menu option to know how to CC gel our flashes so we can match the ambient without buying an expensive flash color meter?

I dunno what we would have to do to get this noticed by anyone at Nikon Japan (plus-sized strip-o-gram, maybe?) But if you are onboard for the idea, please leave a comment, tweet it, blog it -- whatever.

Maybe we'll get their attention somehow.


__________

Brand new to Strobist, or lighting? Start here.
Or, jump right into our free Lighting 101 course.
Connect: Discussion Threads | Reader Photos | Twitter

139 Comments:

Blogger Mitch said...

So Cool!

October 06, 2009 3:37 PM  
Blogger Scott said...

This would be so incredibly useful I can't even begin to describe how it would change the process of how I work with my flashes.

October 06, 2009 3:43 PM  
Blogger Andrew said...

From the coding side, I can't imagine this being that much work for a firmware upgrade. Nothing too new, like David said, just utilizing in a new way what is basically already there.

Dear Canon/Nikon,

Fix this immediately.

Thank you.

October 06, 2009 3:46 PM  
Blogger bobby said...

I second that! Literally... cause The Light commando said it first...

October 06, 2009 3:47 PM  
Blogger TC said...

Great idea! Here's hoping somebody picks it up... would make them a darling of the 60.000+ strobist members.

(the http://www.rosco.com/us/filters/calcolour.asp gives me a 404 error)

October 06, 2009 3:49 PM  
Blogger Stephen Crowers said...

Mightn't this be easier to do if gels were named by color temp (converted to) rather than color? e.g. - full CTO would be named 3200K. Most cameras already tell you the temperature they come up with when setting the custom white balance (I think).

October 06, 2009 3:58 PM  
Blogger Jim Cutler said...

Superb idea. Seems all the necessary ingredients are already there. Just tie it together for us Nikon, please! I agree that getting Nikon to hear about this is gonna be tough.

October 06, 2009 4:01 PM  
Blogger Noah Bennett said...

Fantastic idea, this makes too much sense to ignore.

The relevant data could be displayed right after you set the custom white balance by shooting a white/grey card. even better, would be if that info was integrated into the metadata of the photo, so when you look at that photo in camera, the filter data will be accessible even after shooting other photos.

In the same vein, you should be able to set preferences so that if you want a particular effect, say a slightly warmer key, you could set it that way, and when it gave you info about how to gel, you would automatically get the color just how you want it.

October 06, 2009 4:03 PM  
Blogger Paul said...

An excellent idea - so good in fact you wonder why it hasn't already been done!

October 06, 2009 4:10 PM  
Blogger perfectblog said...

Actually, it can be an app, iPhone app, for example. You balance your White Balance (oh, pun..) way you like, enter data (or even make a photo of display) and app gives you right type of filter on your flash. Something like that.

October 06, 2009 4:15 PM  
Blogger Tony said...

David

Or... What about the OTHER way around... Have the offset printed on the Rosco filters...?

Tony

October 06, 2009 4:15 PM  
Blogger Christopher said...

That is a great idea indeed. Maybe it could be a little difficult to include the other manufactures (Lee is very popular in Germany e. g.)...

Btw: The link to the Rosco product does not work

October 06, 2009 4:18 PM  
OpenID wetzelphoto said...

Great Idea! I would love to have this feature in my D300...

And while you are at it Nikon, can you please upgrade the firmware of my D300 to allow me to set my bracketing steps to 2 EV so I can shoot -2, 0 +2 for HDR???

That would be nice too!

October 06, 2009 4:23 PM  
Blogger Dean said...

FANTASTIC IDEA!!!!

October 06, 2009 4:26 PM  
Blogger Tonton said...

My guess is, if Nikon ever does this, that then they won't support the Rosco's but their own gel system which is quite proprietary and expensive.

However, would be better than nothing and a great idea though.

October 06, 2009 4:33 PM  
Blogger Jay said...

That would go so far in helping seasoned pros and newbs to get those balances matched. I'd be on board for signing a petition, sending an email, making a phone call or whatever it takes to get this idea out there.

I guess the only question would be how to get each flash, from old but functional clunkers to the newest model to agree. Perhaps an ambient measurement, a strobe measurement, then let the body tell you the correction required for that particular flash?

October 06, 2009 4:36 PM  
Blogger GGcadc said...

Woot! i'm always for a free smart update to something that already exists that adds useful functionality to it.

October 06, 2009 4:37 PM  
Blogger Craig Maunder said...

This is an amazing idea... unfortunately Nikon's attention to what end users are asking for is pretty pathetic at the moment. More middle of the road point and shoots anyone?

October 06, 2009 4:37 PM  
Blogger Alex said...

Absolutely, the more information available for the photographer to enable the better. You can choose to use it or choose not to, but there can clearly be nothing bad with this. Of course I see Nikon standardizing with one brand of gels, and Canon standardizing with another brand of gels. Now just to get everyone to stick to one standard would be fabulous.

October 06, 2009 4:40 PM  
Blogger mbrakes said...

Fantastic idea. I have been wishing for something like this for a long time.

October 06, 2009 4:52 PM  
Blogger Dan said...

Brilliant! Now we just need Nikon and/or to make it available. Since neither, especially Nikon, seem to listen to the customers much, maybe we should come up with a plan B.

Could we use the camera's auto WB or a custom WB and a test shot to get us a kelvin number and then do a bit of gel math?

October 06, 2009 5:01 PM  
Blogger Kerwin said...

Correct me if i'm wrong...
But couldn't this be yet another iPhone (or similar) app?
I mean, they're not the greatest light/color meters, but they do it better than me...

October 06, 2009 5:14 PM  
Blogger SAL-e said...

David,
I think we are going to have better luck with one of the custom firmwares like CHDK for Canon. I don't know any hacked firmware for Nikon.

October 06, 2009 5:32 PM  
Blogger RJurden said...

Wow... What a genius idea!

I love this idea. I've been toying around with getting serious about color matching my flash for a couple weeks now and am going to pick up the Strobist Kit from Midwest this weekend (I love living in Columbus!)

This would make it so much easier!

BTW, your link to Rosco is dead.

October 06, 2009 5:37 PM  
Blogger Steve Korn said...

Hey David, I nice step in that direction is that Broncolor has adjustable color temperature on their new lights, maybe the technology will work it's way down at some point.

October 06, 2009 5:44 PM  
Blogger Zoomfreaky said...

Oeee that's a great idea, I always f.. up my lighting when it gets a bit complicated... I'm on board.

October 06, 2009 5:53 PM  
Blogger Michael said...

Fantastic idea.

But, since I'm so darn lazy, I'd never bother with the gels.

So I've got one more step of refinement for you... why a gel pack instead of a small, one-huge-pixel CCD, mounted inside the flash that sets the 'gel' accordingly? This would act like a projector does in changing the colors.

(Wow, think of the amazing effects you could do with this if you had more than one pixel...)

October 06, 2009 6:02 PM  
Blogger Yuli said...

That'd be cool, but for those that can't wait, I believe Sekonic's Prodigi Color C-500 does it already.

October 06, 2009 6:10 PM  
Blogger Don said...

Fly that one by Joe Mcnally chief Nikon cls evangelist and have him stuff the suggestion box.

It's all about who you know right?

October 06, 2009 6:21 PM  
Blogger pkphoto said...

My father (a colour scientist) has written a program that will do this already. We have done a few experiments (using a spectrometer to get the initial readings) which I was going to send to you when they were ready. It is not quite as simple as it seems. You are right that fluorescent lights are all sorts of colours these days, and the initial results are not bad. The problem is that because specific gels have not been made for a lot of the sources that we are trying to balance then you need to use a combination of gels. For a tricky light like sodium (which has a very narrow spectrum) we are having trouble finding a good enough match that doesn't cut out 95% of the light from the strobe.

I believe that we were planning on releasing the program free under a CC license, we hadn’t thought of building it into a camera…

Perhaps the peeps at Stanford would be interested if Canon is not.

October 06, 2009 6:33 PM  
Blogger sonchai said...

David, this is truly brilliant.
I have had some serious trouble with these dreadful "energy saving" lamps. This problem will be getting more prominent in the future, as the European Union already has started to ban (!) incandescants.

On the other hand, i see two (minor) problems: An imprint (on the gel) of the colour temperature alone wouldn't be sufficient enough, sadly. Colour is a two-dimensional space (as you can see in Davids picture of the Nikon(?)-Display). And a light with, say, 4000K can still be either "yellowish" or "magentaish" (cf. http://en.wikipedia.org/wiki/CIE_1960_color_space for more information, it's sort of hard to explain in words...).
In essence one would need to have the uv-offset being displayed by the camera and imprinted on the gels. I admit, this is rather a standardisation problem. But you know how it's like with these standards...
The other problem (as Jay mentioned) would be the variety of colour temperatures provided by the variety of flashes. One would need to measure the flash alone and take its offset into account, as well.

Anyway, great idea. I'll spend you a Mountain Dew when this will have its break through. ;)

October 06, 2009 7:04 PM  
Blogger George Natis said...

You know David,

Sony is the only one to help you in this matter. This is the only downside on the two big players. Sony cams incorporate a nice feature on the manual WB. When you pick out your grey card and do a manual WB, the camera instantly tells you which Kelvin its adjusting the WB to. Now how nice is that. A free colormeter built in. Canon and Nikon are too proud to copy this feature which is truly sad.

October 06, 2009 7:08 PM  
Blogger Adam Szpruta said...

Fantastic idea. It's far over due.

October 06, 2009 7:08 PM  
Blogger sonchai said...

@pkphoto: Correct me if i'm wrong, but you simply cannot correct for any colour without taking away the "unnecessary" spectrum. Hence, when correcting for a narrow band source you must take away everything else, resulting in what you call a "95% cut-out" (not sure about the percentage).

October 06, 2009 7:15 PM  
Blogger platypus studios said...

I think we can do this without a software mod!
0. Set color balance for daylight.
1. Photograph a white card under the light of interest.
2. View the color histogram. The there should be one peak in each color.
3. Determine the relative position of each peak. (My Canon has divides up the x-axis (brightness) in one-stop intervals))

Categorize your filters by photographing your white card with each filter. Record relative position of peaks in your RGB histogram. Pick the filter that has the same relative offsets in R, G, and B peaks. Remember that 1 stop is equivalent to 30cc units.

This method will match your ambient light with your gelled flash. After you pick the appropriate gel, you can switch to auto white balance or RAW to take the picture.

Hope this makes sense.

Bob

October 06, 2009 7:30 PM  
Blogger John said...

Amen. Balancing flashes with unknown ambient colors is a pain, most of the time you can wing it or even "cheat it", but its still a PAIN.

Nikon jump on board! ...pretty please.

October 06, 2009 7:31 PM  
Blogger Matthew Kerr said...

Posted on my blog www.kerrphotography.ca/blog and tweeted @kerrmatt and on facebook. I think this is a fascinating idea.

October 06, 2009 8:19 PM  
Blogger Mark said...

You see... this is why we love you David.

October 06, 2009 8:33 PM  
Blogger Jason Sheesley said...

For a manual workaround (at least on Nikon) you can turn on live view and adjust the WB manually in Kelvin in real time.

October 06, 2009 8:36 PM  
Blogger John Fowler said...

What a great idea!!!!

October 06, 2009 8:51 PM  
Blogger Jason Wallis said...

I just had this exact conversation. I have been looking for an external light meter that would do that very thing. I will do whatever it takes to aid you in making this happen. Only if you promise to go as hard after Canon....

October 06, 2009 8:56 PM  
Blogger Steve Thurow said...

I don't know why they don't do this. I recommended this to Canon back when the 1D came out and got the canned thanks we'll look into it. Fast forward and they still haven't looked into it.

Take this one step further now with cheeper LCDs and the camera could set the proper white balance, flash with LCD integrated in the head changes to the correct opposite white balance, fire flash through and you have white light all around.

While they're at it, throw in the ability to generate any filter needed or desired in the LCD flash cover.

What? It could happen. If they can build an LCD projector into a small digital camera they can turn a digital camera into a color meter and put an LCD panel in a flash head for unlimited filtration.

October 06, 2009 9:35 PM  
Blogger Wade said...

This sounds fantastic. Actually any development in the firmware realm other than super critical bug fixes would make me feel like I made an investment instead of bought another version of a disposable camera

October 06, 2009 9:36 PM  
Blogger Patey North said...

@pkphoto-

Yes! Please do release it under a CC or GPL license. This is just the sort of thing that photogs with laptops/netbooks all over the world need.

Where can I go for news and updates about your software?

October 06, 2009 10:09 PM  
Blogger JVL said...

I think you'd have to get Canon to admit their flashes need gels, or any modifier, first.

October 06, 2009 10:28 PM  
Blogger Phat Baby Photographer said...

You could take this idea one step further and just have a flash with two bulbs in it: one "cool" (~3000K) and one "hot" (~5500K). The camera just sends the WB info over as part of the IR signal and then the flash could decide how much power to give each bulb to get the right WB. It wouldn't be exact since WB isn't entirely linear but wouldn't be any worse then using gels.

It'd be an expensive flash and probably a bit bigger but would let you work a lot faster, be more convenient when your flash isn't easily accessible or in changing lighting conditions and allow you to do all types of other effects (e.g. imagine alternative warm/cold flashes).

October 06, 2009 11:29 PM  
Blogger Arthur McLean said...

Dear David,

While you're fixing Nikon's flashes, can us poor, HS football shooters get a rangefinder function?

Nikon supposedly uses distance information fed from its lenses to help control flash output. (At least they used to, I'm guessing the capability is still there).

I want to be able to focus on something at mid-field, set my flash power for a good exposure there and have the camera dial down the flash power as the action comes closer to me so I'm not under exposing between the hashes and overexposing on the sidelines.
That'd be great, thanks.

Your friend, Arthur

October 06, 2009 11:46 PM  
Blogger SAM OKSNER said...

I would love this! It would sooo easy, since WB is already calculated in terms of Warm-Cool and Magenta-Green. You do into custom white balance and select FLASH balancing
and that would be it! It gives you a lift of the most compatible gels... done deal! This would be a huge sales point to flash photographers.

Having the Kelvin values on the gels isn't that useful because that just deals with warm v. cool but not magenta v. green. Sodium lamps are very green for example.

Brilliant!

October 06, 2009 11:48 PM  
Blogger Tim said...

Pure Genius!

October 07, 2009 12:04 AM  
Blogger bradj said...

I'm with Steve. Let's have an LCD (or similar tech) over the flash that will change colour to match the ambient/available light. Until it gets that integrated, I think it's all stop-gap measures.

October 07, 2009 12:08 AM  
Blogger rick said...

i've wondered about that - if you look at the metadata you can see white balance correction.
If you could shoot tethered, or you had time to do a white-balance exposure you could measure it on the computer, and then setup for it. But having the camera report it somewhere would sure make it easier!

October 07, 2009 12:19 AM  
Blogger Edward said...

Ok, this is a very neat idea, but...

I don't see it being implememnted in any but the highest end cameras. Think about it. You want to take the functionality of a $1k light meter like the Sekonic C-500, and incorporate it for free in a camera.

Not gonna happen anytime soon.

Think about it. What is preventing the camera/flash companies from incorporating RF wireless flash triggering built into all of our flashes and cameras?

October 07, 2009 12:37 AM  
Blogger Alex said...

Very interesting idea, but as one commentator above notes, you would have to change the naming of the gels to meet this need. The Rosco designations seem to be standard here on this blog, and perhaps in the US, but its that the case in other markets? (Not a rhetorical question--I do not have the answer to this.) In order to make it work, would we not need gel color standardization to make this work? At the very least, K color temp. rating for viable gels would be a must. Then the camera manufacturers can correlate the readings from their sensors with a matrix of color-temp names (numbers) that matches the gels.

October 07, 2009 12:53 AM  
Blogger andy said...

Fantastic idea!!

canon/nikon please please please please....?

October 07, 2009 12:54 AM  
Blogger JMart said...

@Baby Phat - This would be worse than using gels. For example, what if the ambient light was heavy on the green? The two bulbs would not be able to match it with any combination of "hot" and "cool" while a specific gel could provide a near perfect match.

I like Michael's idea of having a single color changing element in front of the bulb though. :)

October 07, 2009 1:24 AM  
Blogger me said...

Sounds like a pretty good idea. I also like the idea some people have mentioned of flashes with built in "gels".

October 07, 2009 1:41 AM  
Blogger Jacob said...

This is a great idea - but wouldn't it be possible to make some kind of table that could be printed on a small card that you could bring along with your gels/flashes? I would be happy to assist in the layout of these cards, however i'm not so math oriented.

Of course it would be wicked cool if it could be automated by firmware updating the camera, however i doubt that neither Canon or Nikon will develop new firmwares for older cameras like the EOS 40D / 5D mk I.

October 07, 2009 1:50 AM  
Blogger Jason said...

I'd have to say that would be a fantastic feature to have considering how difficult matching the flash to random ambient lighting is now.

October 07, 2009 2:03 AM  
Blogger Lux Artist said...

This IS a neat idea! Reminds me of adjusting the filter pack for printing color photographs BD (before digital). Being able to control the ambient color temperature would be a big creative boost. Sadly I don't expect to see this on Canon cameras soon--we're still waiting on mirror lock-up that's not buried deep in a menu.

October 07, 2009 2:04 AM  
Blogger svnsn said...

sure man, let's get it on!
shouldn't be a problem for the canon-guys either...
hope this works!
cheers - sven

October 07, 2009 2:33 AM  
Blogger Paps said...

Clever thinking.

Want one! Now! Or tomorrow.. or next week... But I want this option available in my cams. Maybe these guys can hack this into G9 firmware: http://chdk.wikia.com/wiki/CHDK.

A new topic has been started. All register and vote here :) http://chdk.setepontos.com/index.php/topic,4328.new.html


(10 hours later wordpress is "strobed" to death by huge number of visitors ;) )

October 07, 2009 2:36 AM  
Blogger abey said...

Great idea!

I'm not sure about the technical details, but assuming they can be worked out, I would guess that it's easier to get Rosco to write a iPhone app than Nikon to modify their firmware (let alone Canon). This is btw partly understandable as camera manufacturer would not like to commit to a specific filter company...

I, for one, would buy the iPhone app. Especially if I get 10% off the related filter pack.

October 07, 2009 2:56 AM  
Blogger jeffegg2 said...

A flash with three bulbs behind gels red, green, blue. And power control for each would do it. Then have the camera control the color to match the custom color balance. I think this would not be that difficult to acheive.

October 07, 2009 4:36 AM  
Blogger Raul Kling said...

Excellent idea, David. I hope the guys at Nikon will hear you (and us, the members of the Strobist grup). Keep us posted!

October 07, 2009 4:43 AM  
Blogger Mich said...

! postponed reading article to note the idea and want to share it here -

How To Select Proper Gel For Ambient:

- go to manual WB and set it for your current ambient;
- put your gels in front of you and take a shot of them. The grey one is your first choice.

p.s: excuse me for my (yet) lame english (in case of) 8}

October 07, 2009 4:53 AM  
Blogger Ken Davidson said...

I'm quite intrigued by the cludges folk are coming up with, rather than addressing the problem head on.

What is the problem? We may compansate for ambient by shooting a grey card, but we can't meter for ambient+flash (which will change the total light colour in the scene).

If we could meter the total light colour *at the moment of flash*, and then dial that into the camera, why use gels at all?

By metering at the instant of flash we would also be taking into account strobe degradation over time, rather than relying a fixed spec at time of manufacture.

Haha! Now, is there a technology out there that can deal metering in the instant of flash? Computationally, this doesn't even have to be done in real time, but some time after a critical exposure had been memorised.

Hang on. Clearly I'm thinking out loud here. What's wrong with just shooting a grey card in the scene, flash + ambient, and correcting for the aggregate colour after a test? Why do we need to match the flash to ambient with gels?

Then, if we need to modify say a single strobe to add a red cast for a special effect, we meter first the aggregate colour sans gel, dial that into the camera, then add the gel. The /difference/ will still be recorded.

Hey, don't flame me. I'm just trying to think it through.

Someone, no doubt, will correct me - as it can't really be that simple can it?

October 07, 2009 4:55 AM  
Blogger Lars said...

Well I think that it's a great idea. Might even force some of us to buy a new camera..

But are we forgetting that someone will have to sit down and create a "look-up" table to match a white balance off-set with a single filter or maybe a combination of more filters. Or do the filter manufactures already know which off-sets match which filter?
(will this work in all color spaces?)

And what if your white balance off-set is just on the border between two filters? Do we just combine those two or? I think a fair bit of testing is involved. I wouldn't mind getting involved with testing. Just make them send me a test camera with beta firmware!! :-)

October 07, 2009 5:23 AM  
Blogger Adrian said...

yes, come on ... it's genius

October 07, 2009 6:54 AM  
Blogger Simon J said...

It would be a start if the Custom White Balence actually told you what color temp it was reading...Canon's don't, unsure on a NIkon.

October 07, 2009 6:55 AM  
Blogger Philippe and Patti said...

Sweet idea.

@platypus studios: Would love to hear more on this.

Whichever solution works, it's needed.

October 07, 2009 7:47 AM  
Blogger greg said...

I now this might be a pain in the butt, couldn't you take a photo of a white card under the ambient light and measure it in lightroom. Then you could use the chart that comes with Rosco color correcting gels to get the right gel?

October 07, 2009 8:08 AM  
Blogger TC said...

.... want to add a bit..

Two ideas:

1. Add a LCD matrix in front of the flash head and make that adjust the temperature of the flash. Presets and free RGB adjustments please. Add data to the x-ttl signal, so we have ttl with full color balance...

2. If the camera makes gel suggestions based on readings... make it give alternatives ranked on how well it matches, like:

90%, ½CTB+1/8CTM
70%, ½CTM+1/4CTO

So if you have a full swatch of gels, you may be able to hit a very good match, but if you only have a basic set you'll still be able to match fairly well..

October 07, 2009 8:22 AM  
Blogger Andy said...

Well, indeed this would be a fantastic improvement!

October 07, 2009 8:50 AM  
Blogger Gabe P said...

Right on David... I've been wondering for years why I can get at least a kelvin temp number out of the camera. It's there in camera raw software on the computer, but by then it's too late for the shoot.

October 07, 2009 9:45 AM  
Blogger SkyTang said...

superb! what a great idea. I'm in.

October 07, 2009 11:01 AM  
Blogger Nick Strocchia said...

Yes, this is a phenomenal idea!

October 07, 2009 11:02 AM  
Blogger GeoWulf said...

Call Chase Jarvis, have his crew write an iPhone app, and after it sells millions, have him arrange a meeting with nikon and bada bing, Your GOLDEN!

October 07, 2009 11:18 AM  
Blogger Kevin said...

"Rent-a-McNally"
in all seriousness that would be awesome. Good concept and im crossing my fingers hoping it will happen.

October 07, 2009 12:08 PM  
Blogger Dawn said...

Sounds like a great idea !

October 07, 2009 12:27 PM  
Blogger Bryan said...

High end video cameras have reported the color temp when you do a custom white balance for many years! Great concept to have enthusiast & higher level still cameras to include that info!! Please listen to the users!!!

October 07, 2009 1:05 PM  
Blogger marco said...

Sharp idea Dave. May I go a little further and walk the sci-fi territory just a second?

I'd like to have the flash automatically set its "white balance" using a properly colored light, instead of the usual white one, to match the ambient. That info could be just another part of the TTL signal. Shooting a grey card would set the correct WB both for the camera and the flashes, and you can of course manually correct all the WB in "stop" increments.

You could ask - "fine, and how do you think you can color the flash's light?" And the answer is simple: I'm not a camera manufacturer, ask them :)

I think it's time to create a color temperature standard, so that lighting manufacturers (flashes, gels, etc.) can build consistent devices.

October 07, 2009 1:49 PM  
Blogger hudson said...

If anyone is interested in implementing this for Canon's cameras, CHDK or my project, Magic Lantern, would be the place to start.

I've made some initial attempts to figure out how to do custom white balancing on the fly to reduce the amount of effort to white balance on set.

I'm also looking into how to adjust the Picture Styles parameters on the fly since that processing happens with the 12 or 14-bit raw sensor values rather than the subsampled 8-bit h.264 stream.

For what you want to do however, it is much easier. You would be able to read the YUV data from the raw image buffer of the gray card and apply what ever functions you wanted to the color space data.

October 07, 2009 2:05 PM  
Blogger Roz Todaro Photography said...

Right on!

October 07, 2009 2:34 PM  
Blogger James said...

Nice idea, but I'd like to add a bit to it. What I'd like to be able to do is tweak the custom white balance. So, if you'd like it a touch warmer than neutral, just do the normal custom white balance (using an Expodisc or whatever) to set the neutral point and then dial it a little warm/cool from there. (This is a lot better than using the "Portrait" Expodisc, which is just an expodisc with a blue cast that fools the camera into warming the light...). This would apply to flash & non-flash situations.

October 07, 2009 3:01 PM  
Blogger captaindash said...

The entire camera world is still completely in the dark ages. Reminds me of the auto industry. The fact that there's no common, universal way to mount all gels without mods is absolutely insane. This is where one of the margin players like Pentax or Olympus could really gain some ground. Offer the crap we've been making ourselves thanks to home depot and the craft section at walmart. The fact we still have to 'think' about white balance for every flash photo is really sad. Unless there's more than one ambient source, it should be automatic. I agree with the lcd built into the flash idea. Keep bugging the dinosaurs. They don't seem to get it yet. The market for a company with its head firmly OUTSIDE it's hindquarters could make a killing.

October 07, 2009 3:11 PM  
Blogger Greg said...

Why bother with gels. A pro camera like a D3 or D700 should be able to automatically correct color on the ambient, then suspend that during the duration of the flash which is thousands of times brighter than ambient per milisecond. It should all be automatic in camera and could be done if they wanted. Then open in NX2 or photoshop and tweak the color slightly.

October 07, 2009 3:39 PM  
Blogger Thingomy said...

Awesome idea -- any bets on how long til it's live?

Now with regard to the rosco strobist pack (that I just got some spam about) what's with the lack of 1/8th cto?

October 07, 2009 4:13 PM  
Blogger vahed4u said...

no one answered the solution (question) put forward by kendavidson?(where the camera computes WB with the ambient+flash factored in) i was thinking along the same lines...still waiting for some informed comment on it...(

October 07, 2009 5:29 PM  
Blogger charles said...

I guess I'm a dinosaur (a professional dinosaur at that). I shoot digital just like I shoot film. I don't chimp, don't look at histograms, use an incident light meter. If color were ever an issue I'd just get a ColorTemp meter, push the button and get a reading. Real simple.

Apparently your mileage does differ 8-]

October 07, 2009 6:32 PM  
Blogger buffaloz said...

Great idea. It would be really nice to have access to a color meter in the camera. Why hasn't it already been done?

October 07, 2009 8:22 PM  
Blogger Darren Whitley said...

I agree.

October 07, 2009 10:49 PM  
Blogger platypus studios said...

Phillipe and Patti, here is some further clarification. If you look at Greg's comment (it follows yours chronologically), he mentions analyzing a photograph of a white card using a photo editor. This can be done using the RGB histogram on your camera. If the white card completely fills your field of view, then there is only one color in the picture. You can figure that color out by looking at the RGB histogram. The position of the peaks in Red, Green, and Blue should be at the same value (remember that the height of the peak is the relative number of pixels at that value, and the position on the x-axis is the relative intensity). In actuality, the peaks will not all occur in the same x-axis position which is indicative that the color balance is not neutral on your white card.

If you did the same exercise exposing your white card with a series of different cc filters, you can "read" off the corrections made by the filter on "daylight" balanced white light (i.e. flash light).

Bob

October 07, 2009 10:50 PM  
Blogger Jeff said...

Genius, and would make my life so much easier. Come on Nikon, hook a brother up!

October 07, 2009 10:59 PM  
Blogger Aaron Brown said...

Sterling idea!

October 08, 2009 12:32 AM  
Blogger Dennis said...

Nice

October 08, 2009 4:29 AM  
Blogger David said...

One thing that I have found frustrating with my Nikon D300 is that although you can do a manual or custom white balance, the camera does not tell you what Kelvin that WB is. When I take my images into post, the program knows what that number is, so it must be in some Exif or custom metadata tag. I would like to be able to know what the WB number is on the camera. For one thing, it would help me gain an intuition for what the ambient colors are in different situation. And if we could get gels in Kelvin, that would be sweet. So I second your idea.

October 08, 2009 5:05 AM  
Blogger Martin said...

I dont really understand how the camera would be able to tell the difference of flashlight from ambient.

The camera only registers light uniformly, though different light sources offcourse appears in different colors. Perhaps the camera could be told to change all red colors into the current preset WB setting. Lets say you have a RED gelled flash and the camera changes that red burst into i.e white. That would cause a problem with skintones, red clothes, red items in backgroun, ie.

I dont see how that would be differend when flash produces a 5500K burst. Changing that color into the scenes ambient, would sure change other colors in the scene aswell?

October 08, 2009 8:20 AM  
Blogger Thingomy said...

I know the original idea here was a quick and cheap way to balance things -- but it it tecnically posible to build a flashgun that has electronic gels built in for colour correction.

The technology already works in high end stage/TV lighting -- there it works by using a disc that has a grad CTO around it, rotated to the approprate position.

I would perfer for our uses to have multiple bulbs with diferent filters (5 of them?), with mixing between the diferent colours. That can't be all that hard can it?

October 08, 2009 10:20 AM  
Blogger Michael said...

Eh-- Fix it in post.

October 08, 2009 10:58 AM  
Blogger Jose said...

I think David´s idea is fantastic, and the base concept for building a truly simple and automated method to white balancing..

In my modest oppinion, I also think that Ken Davidson took this concept a step further, adressing the root of the problem by considering "ambient light + flash" the "total light" to be white balanced. If this were integrated in the cameras, we could just shoot a gray card with ambient + flash, and let the camera do the math to correct for "that" total light in subsequent photos. No gels required, except for coloring on purpose..

The only problem I can think of would be "where" to place the gray card, because I understand that different flash-object distances would give different proportions of ambient light vs flash.. (the Inverse Square Law thing)

Thanks David for your really awesome work and sharing of your knowledge.

October 08, 2009 11:36 AM  
Blogger Sam said...

This is deeply cunning.

But this is the 21st century. If the camera/flash can work out the ambient colour temperature then I also want them to magically mimic it with prisms and micromirrors rather than telling me which bit of plastic to stick on the front.

And I would also like some hoverboots.

October 08, 2009 11:44 AM  
Blogger Stephen said...

Simple or not, this is a great idea. It would sure beat having to buy a expensive color meter that I would have to find space for in my kit.

October 08, 2009 2:23 PM  
Blogger Christopher said...

I'm frankly shocked that cameras can't already do this. I have actually found myself taking a picture of a grey card and manually balancing it in Lightroom to see what the temp/tint is and then gelling for that before.

Having the camera tell me CC units would be even better, but I'd be happy with kelvin and green numbers, which I'm sure the camera has, it just won't give them to me.

October 08, 2009 3:27 PM  
Blogger Luke Shelton said...

I have though about this a lot myself.
What if you shot your white card in the ambient, downloaded your white card shot, printed it on a transparency, and placed it in front of your strobe? Would that not create a perfect gel? Assuming of course that your printer is accurate in color perception.

Obviously you cant always leave to print a gel. But it's just a thought.

October 08, 2009 3:32 PM  
Blogger Stephen Crowers said...

Ok, since we're talking about a totally in camera solution: it could be done the same way that the new Sony point and shoots handle noise reduction- by taking more than one shot and combining them in-camera afterwards. The first shot would be sans-flash at the set shutter speed and would calculate the ambient WB; the second would be with the flash at a high sync speed (to remove the ambient) and use the daylight setting (adjusted as appropriate for the manufacturers' flash). The lit pixels from shot #2 are then overlain on image #1 (or combined heuristically) to create the integrated and balanced final image. You'd lose some responsiveness and action-stopping, but it would work for portrait shooters and casual photography.

hmm...perhaps I should be filing my patent paperwork now...

October 08, 2009 4:41 PM  
Blogger Kevin said...

Extremely interesting idea!

I doubt that most people (eg my mom) will take the time to set up the camera, dig out a gel, add it, then take their photo. Most people I know go from power on, snap, power off in 5 seconds or less. For widespread adoption, it will have to be some completely automated system.

LCD is not practical, because it only works with polarized light. You'd lose a huge amount of light to both the polarization plus the fact that the LCD is a huge mask; very little actual light (maybe 10%?) makes it through an LCD. Aging of the color filters would be a huge problem too; flashes are incredibly intense and break down anything that they illuminate over time.

However, there are a number of other possibilities:

Start small with cell phone cameras that have LED lights. Tt would be trivial to replace the single while LED with a set of three.

There's been some interesting work done recently with iridescent reflective displays: http://www.mirasoldisplays.com That seems like it's plausibly efficient enough to act as a "colorizer" for a flash.

A motorized "color wheel" in front of the flash may be a simple first generation flash "colorizer". Rotate to the desired color, then fire the flash. This has the unfortunate problem of being limited in the number of gels available, but 4 (three plus a clear daylight gel) would be a start.

A potential refinement would be to spin the color wheel quickly (like on a DLP projector) and fire the flash multiple times through different gels. This may cause weird color shadowing on moving objects or unstable hand-held shots, but would still be worth exploring.

October 08, 2009 4:52 PM  
Blogger Patrick F said...

@Ken Davidson - brilliant idea!
So what you're suggesting is that the camera effectively does a double exposure and then composites them - the first (or second, depending on whether front or rear curtain sync is used I guess) with ambient only, the second at the 'moment' of flash (which could be something like 1/64,000 sec).

This makes me wonder if any sensors can be 'sensitive' for such a short burst of time.
I also think another commenter pointed out that this method might have problems if the total flash power isn't higher than the ambient, or isn't higher across the entire image (think a person in front of a bright background). Admittedly, though, such a method would still prove useful in a myriad of situations.
I'm thinking there must be a technological barrier, though, as I've never heard of shutter speeds as high as would be necessary for this to work at all but full power.

To the folks who are suggesting an LCD panel over the flash to automatically control WB, this would be amazing, but I'm not entirely sure an LCD can simulate the full spectrum of light like a gel can (maybe they can, I don' know). But I wonder if an LCD wouldn't just put out too 'pure' of a color (like, say, nothing but 540nm) and thus not simulate the full range of frequencies of a black body, like a real flash does (and like a CTO-ed flash does).

October 08, 2009 8:00 PM  
Blogger bill said...

http://translate.google.com/translate?hl=pl&sl=pl&tl=en&u=http%3A%2F%2Fltt.com.pl%2Findex.php%3Foption%3Dcom_content%26task%3Dview%26id%3D376%26Itemid%3D318

October 08, 2009 8:13 PM  
Blogger Adrian Henson Photgraphy, Inc. said...

Canon and Nikon, please just display the color temp and g/m offset anytime one performs a custom WB, is it too much to ask? Sony apparently does it. Make it a firmware update for all recent cams.
I have been saying for a long time, stop dazzling us with mega pixels and give us features we can fall in love with. Like time synced to the atomic time clock in Colorado, my $50 watch does it. Anyone who shoots events with more than one camera would love to always have their time perfectly in sync. It's $50 technology.
Give us user-definable prefixes in our image numbers so that there is no chance of any of my cameras outputting the same # at a job.
Give us optional image # lengths 4, 5 or 6 digits so that our camera does not roll the odometer every 3 months.
Touch screen lcd that cover the entire camera back.
Through the lens laser aiming for shooting with your camera away from your eye.

October 08, 2009 11:17 PM  
Blogger Thingomy said...

@adrian henson

"Through the lens laser aiming for shooting with your camera away from your eye."

interesting -- it sounds the most barmy on the list, but a few months ago I was doing macro photography, and fell on an idea of shining a torch in the viewfinder, and focusing using the projected image of the viewfinder screen etching marks on the target. The wierd thing is -- it worked. My torch wasn't bright enough to do it usefully, and it depends on having a light object or putting a piece of paper next to the object.

Sony used laser AF iluminators on some of thier high end compact cameras a few years ago, they were so far ahead of anything we have now it's not funny (accurate focusing of a zero detail object at night 30 feet away if I recall) but they had to shelve the idea because of unfortunate issues with people getting lasers in thier eyes, and people shouting "sniper".

October 09, 2009 5:21 AM  
Blogger ylhainen said...

This is superb idea. Even if Nikon cameras would tell you which of the sb-900 gels was closest to the ambient lighting would help a lot.

October 09, 2009 6:18 AM  
Blogger Andy said...

Fab! Just been shooting set-ups in a number of shops - all with different lighting combos. The first 5 mins of each shoot were taken up getting the lights/strobes (roughly right) - often getting the lights turned off - and thena bit of post work in CS3.

Nikon - c'mon!!!

October 09, 2009 8:00 AM  
Blogger Dan said...

I don't understand! Why wouldn't Jose's approach: ambient + strobe measurement in camera with, say, an Expodisc type device and set the custom WB?

October 09, 2009 11:31 AM  
Blogger TEKeez said...

David, you're onto something. Years ago Cokin made some of these for artsy purposes. They called them color back filters. One for the lens and one oversized for the flash. It would correct for the subject shot with flash but the background would go Dali surreal. Should not be a problem for the BIG DOGS to do with a real result.

October 09, 2009 3:49 PM  
Blogger David C. Pearson, M.D. said...

Dave: funny how people think alike. I came up with this idea last November (code named the "flux capacitor" from Back to the Future fame). I even worked on a design to bolt over an existing flash with a color LCD gel that would sense ambient light color temp and then adjust the LCD filter to color the flash output appropriately. So clever did I think my idea was, I started to type up a provisional patent application....UNTIL I DISCOVERED THAT HEWLETT PACKARD ALREADY RECEIVED A PATENT FOR THIS IN JUNE 2004. Arggghhh!!! It's U.S. Patent 6,755,555 if you're curious. Sadly, they've never incorporated it into a product!

October 10, 2009 9:30 PM  
Blogger See-ming Lee said...

This is indeed an awesome idea!!!

October 11, 2009 1:00 PM  
Blogger Sean said...

I agree that the cam folks should have this already. Too bad they don't.

First, this is not something that is 'broken'. It is, however, and awesome out-of-the-box idea David came up with.

Second, I also think that it should be GPL if possible.

And lastly, my note to Canon/Nikon:

Please fix this ASAP.

Thanks,
Sean

October 11, 2009 6:54 PM  
Blogger rexyinc said...

ok guys

- your first wish has been granted -

I sent this link to a close friend of mine who works on a major camera makers development team.. so now at least someone who works on the firmware design and other bits now knows about this idea and is in the position to actually think about it and follow it up if it has merit..

he said if the development doesn't cost a mint he'll think about it a bit.. :( I'll keep on his back about it cos I actually thought about something like this years ago but did nothing about it ..

I've already seen colour temp changing LED video lights in China. same tech could work here ...

- but - this would need to be looked into in depth " UNTIL I DISCOVERED THAT HEWLETT PACKARD ALREADY RECEIVED A PATENT FOR THIS IN JUNE 2004. Arggghhh!!! It's U.S. Patent 6,755,555 if you're curious. " -as noted by David C. Pearson, M.D.

Later
-Tony

October 12, 2009 6:56 AM  
Blogger Richard said...

Err, as no one has yet let on to Jose, Ken & pals why we need to match flash to ambient, allow me:

Your 'total light' assumptions are based on a completely even flash and ambient illumination! This never happens in the field, unless you are taking a picture of a totally flat surface illuminated with no graduations by flash and ambient. Real photographic subjects are *unevenly* lit by the flash and ambient, so different mixing occurs between the two. Imagine a person lit on one side by a street light, and on the other side with a flash. If the flash is not gel balanced to the street lighting, then the person will *always* be a different colour on one side to the other. There is no such thing as 'total light' as you call it - we're not integrating light colour across the whole scene - and so different surfaces in the scene will always reflect more or less of one light source than another (unless both ambient and flash sources are in the same position, and pointing in the same direction. Which happens precisely never).

For further information, I would recommend starting with the 'balancing light' section of the Strobist tutorials.

Hope this has cleared this up for you.

October 12, 2009 10:18 AM  
Blogger Richard said...

p.s.

Platypus studio's answer is spot on correct. It is the simplest and most effective thing we can do *right now* to improve our flash balancing, and not waiting a decade to see expensive in camera/flash versions.

However, it will require a little effort on your part - you must not only record the RGB peak positions of your gels, but also as many combinations of those gels as possible (and there are *a lot* of combinations ;) Of course, some combos are more likely to be useful than others. The method is near perfect, but, I can't think of anything more soul destroying...!

October 12, 2009 10:26 AM  
OpenID BigLittleWhutevR said...

Barry White?!?!?! hahahaha...

-Brandon D.

October 12, 2009 11:03 AM  
Blogger Paul said...

Platypus Studio's answer seems the best.

Mich (October 07, 2009 4:53 AM)'s suggestion also seems worth trying:
- go to manual WB and set it for your current ambient;
- put your gels in front of you and take a shot of them. The grey one is your first choice.

October 12, 2009 3:22 PM  
Blogger Kyle Durigan said...

i agree with the dude above. fix please.

October 13, 2009 6:53 AM  
Blogger Richard said...

@Greg - I really, really, really like your idea for solving flash/ambient colour corrections. Essentially your talking about splitting the total exposure in two parts - one for the longer period of ambient exposure and one for the very short period of flash exposure, both individually white balanced. I wonder why this cannot be achieved with current camera technology? I'm thinking it may be a buffer issue.

If it is technically possible, we could go much further, and split the total exposure into as many exposure periods as we want - no more gels at all...

...Want to balance an ambient incandescent fill with a neutral flash keylight, and also have a cooler flash rim light? No problem - you would split your exposure into 3 parts and program the camera to use the correct custom WB for each part. No ambient light contaminates the two flash parts because they are so freakin' short, and a miniscule delay between the two flash firings allows you to WB each separately. Bink-bink-bink in the press of a few buttons and no gels needed! What a truly genius idea!

All it requires is an accurate electronic shutter (why do Nik/Can-ons still have physical shutters anyway?), and a large and speedy buffer to store the raw data of all the part-exposures before the chip applies the custom WB's.

p.s. sorry for the essay, and you should patent this before I (or David) does...

October 13, 2009 7:08 AM  
Blogger Thingomy said...

@richard -- I've done events in the past at night, with many spectators point and shoot cameras flashes would enter the 1/3 second exposures I was taking -- I've seen shots with 5+ diferent flases in them:

http://www.thingomy.co.uk/pics/index.cgi?mode=image&album=/Beltane&image=DSC00539.jpg

and what's needed to sort all that out is basically ultra high speed video at very high ISO -- pull together the frames that you want -- tweek the frames that are a bit odd, and dump the rest. the niose will be reduced by using pleanty of frames -- this also lets you control exposure time in post, and lets you get perfect control over motion blur etc.

Unfortunatly this is way beyond all but the most sophisicated of modern profesional kit -- and i don't think the resolution on that stuff even goes beyond HD -- give me a 120FPS 10MP camera, and I will be a very happy person.

The other problem is dynamic range -- high iso usually leads to low DR -- this is a problem, because I'd like to merge frames with and without flash. the only thing that has the potientila is fovian, and I don't think they have made a proper one of those yet, have they?

October 13, 2009 9:28 AM  
Blogger joshua said...

I'm definitely with you on this.

October 13, 2009 8:43 PM  
Blogger Patrick said...

Brilliant, and seamingly simple solution to be included in a firmware update. Please consider this Nikon! And others.. Also, leave us a comment so we know if you've heard us. :)

October 13, 2009 10:14 PM  
Blogger fjblau said...

GAK!

Can we STOP with the idea that an iPhone app is the platform for ALL THINGS??

Some of us think that the iPhone is an expensive, terrible phone with miserable call quality and coverate... and DO NOT HAVE ONE...

It is just ONE mobile platform... and if you think Nikon and Canon are proprietary... just try Apple. :)

I would MUCH prefer to see this in my freaking camera.

October 14, 2009 11:39 AM  
Blogger Luiyauga said...

Like this idea very much and agree to all the others: shouldn't be such a coding problem!

As some politicician said:

"We can" - so just do it!!

October 15, 2009 11:20 AM  
Blogger Luiyauga said...

I like the idea and as a lot of people metnion it should be very easy.

as Obama said

We can


so just do it!!

October 15, 2009 11:29 AM  
Blogger Eça da Palmeira said...

Great post David, it would be a lot easy even for us amateurs to make a color correct photo with flash. But there's something puzzling me. Every time I click the picture of the post it links me back to the Barry White wikipedia page. Is it on purpose ? I love Barry White, it's been a trustworthy companion since many years and never left me down ;D . One might wonder if that mystery-vapor high school gym you mention could be the girls locker?
Francisco Monteiro

October 15, 2009 2:35 PM  
Blogger David said...

Just a little "White" balancing easter egg.

October 15, 2009 5:17 PM  
Blogger lindes said...

Nice idea!

And funny, I've already been working on an app for doing color correction (CC) calculations, but for color darkroom purposes. Taking the idea you present here to the iPhone could be a perfect extension of that work, though... With a little help from you, I might just be convinced to do the work for that.

And Canon and Nikon: please, make sure you at the least give numeric color temperature and tint information on-screen for any future firmwares (be it new cameras, or updates)... and if you were to obviate the need for an iPhone (or whatever) app that used that information by providing a filter pack listing, all the better.

(Of course, with the iPhone app version, you could correct for known differences in your strobe color output, and have dial-a-correction for warmer or cooler light, etc. etc.)

Nice idea, David!

October 24, 2009 6:25 AM  
Blogger Kevin said...

Hopefully Canon & Nikon are paying attention. Get this out there!

October 27, 2009 12:51 PM  
Blogger scott-c said...

One of the above posts suggested a color LCD to be integrated into the head instead of correction filters may have seen the future. A recent prototype lens featured on the Australian show 'The New Inventors' uses a B&W LCD over the image sensor. With a nifty bit of software the lens detects bright areas (read sky) and turns on the LCD in those areas (read aligned pixels) to act like a neutral density filter. There goes HDR photography.

see Episode 42 at http://www.abc.net.au/tv/newinventors/video/video.htm

November 19, 2009 11:39 PM  
Blogger Bob said...

I was just checking out my D200 and while it can correct for a custom WB, it never shows it anywhere. It would be nice if they showed the offset like you said or at least the temp it picked. If I missed where the D200 shows the selected temp, please comment.

November 25, 2009 6:55 PM  
Blogger RBJ1948 said...

Dave the colour temperature is available on the raw photo file, but you have to open the picture in Bridge or similar software. When you open the raw file it states the colour temp, I think. My problem is that info is not available at the camera so it means shooting under the ambient light check the samples and correct next time there.
I am shooting with a Nikon D90 and it would be nice to be able to add a strobe or two to 'arena' shots.

January 01, 2010 5:12 PM  
Blogger James said...

Someone with more know how might be able to write the code needed using Nikon's software development kits (SDKs)that you can apply for at(https://sdk.nikonimaging.com/apply/).

There is a claim that someone accessed the video feed using this program. There is a version listed for the following cameras: D3, D3X, D40, D60, D80, D90, D200, D300, D300S, D700, and D5000.

I do not know how to do this type of thing, but maybe you can, or maybe you know someone who can.

January 10, 2010 8:41 PM  
Blogger BallardFamily said...

NewB here, or at least someone with only a little strobist experience.

Why not just flash the white/gray card while doing an in-camera custom white balance? I was not even sure my D90 would fire the flash while doing this, but it did.

January 21, 2011 11:45 AM  

Post a Comment

<< Home