Let's get anal about white balance with flash

PacAce wrote in post #4814771
Wilt, although the results are better, I would still caution everybody from making a conclusion based on your latest test images. Why? Look at the window borders of the test shots above and those of the JPEG shots. Unless you can make them all have the same color blue, there's no telling how much of the color discrepancy is due to whatever was changing the colors of the window frames and how much of it was due to the strobeor flash itself. Or is the one window frame darker because that window has been selected while the others have not been?

the one window frame darker because that window has been selected while the others have not been. this is a 'screen capture', with all three photos loaded and viewed at the same time and grabbed into a single JPG, so the active window at the time of the capture is what Windows colors differently

And as good as RAW is, if you really want to be anal about color fidelity, the color balanced shots are still NOT the same as the one illuminated with a properly color balanced light (although I will admit that the last test used more off-balanced lighting due to the Rosco Tough filters over the flash, that would ever be encountered with even cheap electronic flash units!)

Wilt wrote in post #4815199
the one window frame darker because that window has been selected while the others have not been. this is a 'screen capture', with all three photos loaded and viewed at the same time and grabbed into a single JPG, so the active window at the time of the capture is what Windows colors differently

If you don't want the "windows effect", I suggest you develop each in ACR, then open all 3 in PS. The copy them to different layers in one image with them arranged the way you have them.

You second post talked about illumination. I did a quick check yesterday on the jpeg....not in detail...and have not had time to follow on the RAWs. However, running the RGB eyedropper on the borders seemed gto show difference, which I suspect were caused by uneven illumination. Have you seen/checked that Wilt?

Thanks for doing these tests...it adds to the wealth of knowledge...

Wilt, I won't disagree with you here, but you may want to tweak the exposure of those three shots with your RAW converter so it's exactly equal, and see if it helps.

magicmikey wrote in post #4815056
Can you elaborate on "stable"? Are you referring to power output from shot to shot or are you referring to color temperature from shot to shot or both?

Both. I measured magenta shift and color temperature from lowest to highest power and from shot to shot.

From shot to shot, at any given power the color balance remains relatively consistent with pretty much all of the strobes. Opus had some problems with that, but the unit varies by as much as half a stop depending on how long it charged for.

BottomBracket wrote in post #4815155
Pictures? I'm interested in the tests and comparisons.

It was a test for myself. I didn't write it up lab report style because I never thought there was a big audience for it.

It won't change anybody's opinion. A lot of pros shoot with bees and get good colors regardless of a shift of 200k... a lot are not willing to shell out extra 5k to get that consistency. So what's the point of typing?

Curtis N wrote in post #4815031
2) A WhiBal card or other good reference target and a RAW converter are usually required to get truly color-correct images, regardless of the lights.

Not exactly.

The whole "white balance" thing degrades the quality of an image. It might not be significant or noticeable, but it still does.

The attached gif demonstrates that I sold my wacom and was drawing with a mouse. It also shows what the output from the sun and good flash is.

The X axis is the wavelength and the Y axis is the emission spectra of a light source. Most sensors and films have been balanced to daylight. The good flashes mimick that daylight, so that there's no manipulation to the curve which has been applied so carefuly in the labs.

If your light strays from that temperature, expecially if it's a halogen, the output will look like "nalogen" spectra in the picture.

Which means some colors won't get as much excitation electrons as they should, and wouldn't appear the same as they normally would under the sun.

So if you match the "nalogen" lightsource emission spectra to "subject" absorption spectra, you'll see that it's not getting enough excitation electrons in the frequencies which are most important for it.

Hence if you try to color balance such a picture, you'll get noise in the colors. This results in muted and degraded colors.

This is the difference which Broncolor, profoto and elinchrom provide to the professionals... and that's the difference why the big pros don't shoot with bees.

Let me provide an analogy, one side of which is generally accepted by the forumites: It's like the color difference between 35L and some non L zoom. It's not always obvious and you can make good pictures with either... but the 35L is more vivid/neutral/saturate​d/better. Hard to describe, but the difference is there and the difference is noticeable.

I had a similar experience with lights.

Curtis N wrote in post #4815567
Wilt, I won't disagree with you here, but you may want to tweak the exposure of those three shots with your RAW converter so it's exactly equal, and see if it helps.

Curtis, how anal do you think I am?!

I think that is an exercise not worth doing, simply because it is discernable that some colors are not affected at all, and the ones that are affected are those you would expect to be affected by shift in the red-green axis. So changing the exposures to be precisely the same does not alter that perception.

jrsforums wrote in post #4815556
You second post talked about illumination. I did a quick check yesterday on the jpeg....not in detail...and have not had time to follow on the RAWs. However, running the RGB eyedropper on the borders seemed gto show difference, which I suspect were caused by uneven illumination. Have you seen/checked that Wilt?.

Uneven illumination...yes, the sheen of the flash on the card. One would expect sheen would impact two adjacent color patches similarly, not selectively as one can see in the sample shots, though.

DocFrankenstein wrote in post #4815756
The whole "white balance" thing degrades the quality of an image. It might not be significant or noticeable, but it still does... Most sensors and films have been balanced to daylight.

Sensors are not balanced to anything. A RAW file is not an image, and can't be made into a color image without a white balance setting of some sort. Whether the white balance setting is made before exposure in the camera, or after exposure with a RAW converter, there is no single white balance setting that is superior.

I will agree that certain lights such as sodium vapor and some fluorescents have color spectra so terrible that it's near impossible to get good color from them. But the minor differences between electronic flash units do not limit us in this way.

This thread sure has been fun!

And interesting. Lots of great information here, food for thought.

Now we can move on to the real issues of life. Like why, after shooting a white balance in daylight, then shooting an oil painting, I still have to color correct the image to get it to look just like the painting? I'm half tempted to give up photographing artwork - but the artist is my wife, the half that wants it done.

Curtis N wrote in post #4816286
Sensors are not balanced to anything.

Really?

What about the RGB filters in front of the sensors?

Those are balanced to 6000K just like color film.

Even if they weren't, if the emission spectra of the light source is uneven, some parts have to be overamplified in WB correction and that does introduce noise.

Titus213 wrote in post #4816527
And interesting. Lots of great information here, food for thought.

Now we can move on to the real issues of life. Like why, after shooting a white balance in daylight, then shooting an oil painting, I still have to color correct the image to get it to look just like the painting? I'm half tempted to give up photographing artwork - but the artist is my wife, the half that wants it done.

I can't go into details to explain it because I don't fully understand the physics behind it myself but I do know that the color of certain objects behave differently under under different types of light. I'll see if I can find the article that discusses this if you're interested.

The other reason why the image may look different from the actual painting is because you are white balancing the photographed image but in reality, you are viewing the painting under a non-daylight balanced light (assuming you aren't taking the painting outside to photograph).

DocFrankenstein wrote in post #4816657
Really?
What about the RGB filters in front of the sensors?
Those are balanced to 6000K just like color film.
Even if they weren't, if the emission spectra of the light source is uneven, some parts have to be overamplified in WB correction and that does introduce noise.

Each photocell turns either red, green or blue light (light of various wavelengths) into voltage. The voltage of each is recorded. Extraneous voltage that does not actually come from light creates the noise. This is the RAW data. And it's only a pile of data until the computer or image processor applies its white balance setting - that is it factors in the relative intensities of various wavelengths of the light source.

BottomBracket wrote in post #4815155
Pictures? I'm interested in the tests and comparisons.

Hey Pio- Robert and I will be taking an posting images this weekend. I'm in agreement that Bron are the best but I didn't want to put out the cash to buy one for our tests. Likewise the only Elinchrom light in our test will be the cheapest of the cheap light which is only slightly more than Alien Bees. I have no doubt that if we tested their RX and BX lines that we would get better results than the D-Lites.

We've got to grab Mark and go out shooting.

DocFrankenstein wrote in post #4814887
Some people equate baul buff to Jesus or Muhammad ibn ‘Abd Allāh and are ready to defend it to death.

This is a spit up your drink funny line!