We had 2 Lowel Tota-lights http://www.lowel.com/tota/ set-up for a video shoot, so I thought here is a chance to see what a Canon DSLR would do with 3200k lighting. Light on left bounced into white ceiling, lamp on right bounced off white wall.

Camera was in Program mode with auto white balance, just like many people would do. I could have done this all manually, but that's not how most people shoot in the real world.

Shot on left is .jpg straight from the camera, no correction. Shot on right is RAW file color corrected in Bibble Pro using click white eyedropper, no other corrections. Any RAW converter should work as well.

That's why I switched to RAW, even for studio stuff.

FWIW - Canon AWB is known to be awful, especially in tungsten lighting. Did you try it set to "tugsten" WB?

What I find interesting in the chair example is how little noise there is, even in the shadows of the blue channel which is typically the worst in any light. What camera and ISO did you use?

cgardner wrote:
FWIW - Canon AWB is known to be awful, especially in tungsten lighting. Did you try it set to "tugsten" WB?

I know how to do it. I could have done it. But I was more interested in showing the average shooter a before and after shot as a learning tool. The HD video I was shooting with this lighting set-up was, of course, set for 3200k in camera.

What I find interesting in the chair example is how little noise there is, even in the shadows of the blue channel which is typically the worst in any light. What camera and ISO did you use?


Nothing fancey, just a Canon 40D with an 85/1.8. The ISO was bumped to 400 so I could hand hold the shot. This was lit for the video, no changes made for the stills. I shot 8 pix, then we tore down the lights/grip/camera with teleprompter and went home. This was a total afterthought, not planned at all.

The nice thing with the chair shot is that there is a wide range from pure white to black. The chair and the blue bowl were already in the conference room, a good bit of luck .

c.d.embrey wrote:

Camera was in Program mode with auto white balance, just like many people would do. I could have done this all manually, but that's not how most people shoot in the real world.



I don't know anybody that shoots auto WB. (shrug)

I realize you know what you are doing. I was just interested in knowing if you had shot a baseline comparison in tungsten WB.

The first thing I do with any new camera or flash is just put it in the most automatic "idiot" mode like P or the green box (the meaning of which escapes me) with all the AF points enabled and go around and shoot stuff. All cameras, even different models of the same brand, handle metering and AF differently and letting the camera do all the "thinking" in fastest way I've found to get a baseline understanding of how smart it really is. I then put the camera in Av and repeat the random shooting, then finally shoot in M mode and try to out think it. The comparison of results is always interesting and enlightening.

Every digital camera I've owned has required adapting my workflow to how it works. There are pitfalls using techniques which worked previously, like focus / recompose with a system which expects something different. Focus / recompose can totally confuse Canon's default evaluative flash metering in Av mode because it expects the ambient exposure lock to be done in the same place as the shutter is fully pressed triggering the pre-flash and evaluative ambient vs flash metering process. Another similar example is failing to realize metering is done off the viewfinder and if the face isn't covering the eye piece, such as on a tripod, it will also skew exposure.

One of the first things I found out with my baseline testing is that AWB with indoor ambient sucks.... But interestingly the AWB setting, when used with a newer EX flash, works very well because the flash and camera communicate WB information. I don't have a clue what they discuss, but I have tested AWB with flash with a gray card reference and found it to be quite accurate, even if use a diffuser on the flash which warms the light by about 600K.

When I got my 580ex flashes one of the first things I did was put the camera in P, the flash in ETTL and point it from inside out an open door as a baseline test of how well ETTL worked. Darned if it didn't expose the outside through the door and inside wall of the house nearly perfectly. I did the same thing recently with a high speed flash test. I set the ambient exposure for a backlit scene, then just turned on the flash in ETTL mode at FEC = 0 and it matched the lighting on the shadow side perfectly.

The fact flash works so well is probably why in low light situations when shooting in "idiot" mode the flash pops up automatically when the camera detects the ambient levels are too low.

k7xd wrote:
I don't know anybody that shoots auto WB. (shrug)


There are thousand of people lurking on this site, and other photography forums who use auto white balance. You can tell from the questions asked, that many people on the forums are not professionals and do use auto white balance. Many of these people don't shoot RAW either (and wouldn't know what to do with it if they did). Doesn't make them bad people, just uninformed.

cgardner wrote:
I realize you know what you are doing. I was just interested in knowing if you had shot a baseline comparison in tungsten WB.

As I said, this was an after thought. I remembered that you and several other people were discussing shooting under tungsten light, in a thread earlier this week. So seeing that I had a 3200k lighting set-up, for the video, I thought I'd take five minuets and shoot a couple of shots. And post them for your perusal. That's all, sorry no formal test

I like auto white balance! It's very useful! It's not for indoors under studio lighting for sure tho.

White balance does affect RAW afaik though. And temperature/tint sliders are linear - meaning you're removing or "destroying" image information. You can set ACR to load your jpegs and produce the exact same results in jpg as you produce with the RAW - regarding temperature and tint. For color balancing there is almost no advantage to RAW over JPeg. Hue is different. Hue is nonlinear (360 rotational) and non-destructive. Maybe that's a tip for people trying to correct really bad color casting? - Use Hue first to get it as close as possible and then "temp/tint" correct those results.

I don't understand this test though. Two things freak me out. One you're bouncing the light. That would be OK (for this test) if it was 100% pure white. But no one in their right minds paints their ceilings and walls 100% white even if you could find the paint to do so. Most of it's not even close to 100% but more like beige or very light blue. So by bouncing it you just blew away the one constant you did have in the test. The other thing is the AWB. I could understand testing how good the canon WB presets were. That might be of interest. Why AWB though?

Because of this thread http://www.fredmiranda.com/forum/topic/710888 I made the mistake of thinking some people would be interested in the pix.

The problem with tungsten vs daylight and flash is that the camera captures images with red, green, and blue filters over sensor sites. Tungsten light is rich in red and green (red+green = yellow) but almost devoid of blue. So to get proper WB, regardless of how it is set, the camera must amplify the signal of the blue sensor sites much more than the red and green ones. In the process of amplifying the signal the base noise level is also amplified. That's why tungsten-lit images are tend to have more noise, especially in the shadows where there is less signal, especially at higher ISOs which also boosts the amplification.

Having shot under tungsten light since the 1970s, I sorta disagreed with his assumptio.
This poster usually post pix to back up his position. But in this case he did not.

Sometime later I was shooting a video for pay (not to do FM tests). But at the end of the day I shot some pix under tungsten lighting (video is commonly shot with tungsten light).

I thought these pix may be of interest to some of the beginning photographers (NOT PROs). They were taken using AWB because some of these beginning Photographers use AWB (and it is about the worse case scenario).

Bifurcator wrote
I don't understand this test though. Two things freak me out. One you're bouncing the light. That would be OK (for this test) if it was 100% pure white. But no one in their right minds paints their ceilings and walls 100% white even if you could find the paint to do so.

Bouncing light off of walls and ceilings is common in the movie biz. Film color timers don't have a problem with this, neither do video colorists. The producer I'm shooting for like the results. Bouncing light this way may not be perfect, but it works.

Sorry I posted this thread, all it seems to have done is confuse people. Five minuets wasted from my life

c'est la vie