White Balance slightly off anyone else?

[JFirestone]

I just did a test shoot with my RED, a new CineAlta 900R and an EX1. We were using tungsten lights, and while the color looked correct for the other two cameras, my RED looked off, when set to 3200. I white balanced with a card, and the camera registered that it was about 3800. Other shooting has shown the color temperature to be off by a bit. Since the camera records raw, I'm not too worried, and it should be an easy fix, but I was wondering if anyone is running into this, or if it is a setting somewhere on my camera that is causing this.

Jon
#232

[donatello b]

i haven't compared it to other camera's ..
but when i've tried the white balance the results ( viewing LCD or external monitor) have always been on the cool side - so i either switched to tungsten or daylight ...

just did a white balance in liv room ...
red = 2881 ..
minolta color meter II 2450
minolta color meter III 2480

[JFirestone]

Thanks donatello, that's along the lines of what I am seeing as well. It looks like it's just a little off.

[Rich Schaefer]

How dose it look if you adjust the WB manually?

[JFirestone]

How dose it look if you adjust the WB manually?

When I white balance to a card, it looks fine. The camera shows you what the current balance is, and the camera was reading about 3800, when I was white balanced to a 3200 source. The other cameras looked correct at 3200. My Red looked too blue. At 3800 mine looked right.

[Seth Larney]

Hey JFirestone,

Does it look the same when you bring it into Redcine/Redalert ? I wonder if it's the LCD on your camera ?

Athough the screen on your computer may not be much of a better indication unless it's calibrated.

Cheers,
S.

[JFirestone]

Hey JFirestone,

Does it look the same when you bring it into Redcine/Redalert ? I wonder if it's the LCD on your camera ?

Athough the screen on your computer may not be much of a better indication unless it's calibrated.

Cheers,
S.

It does look the same in Redcine. It's not just an observational thing, it's a measurable thing. When using 3200 balanced lights, and a white card, and using the auto white balance feature of the camera, the camera reads the card at about 3800. donatello's post also shows that the camera reads significantly different than his color meters.

[Manfred Lopez]

Does anyone know if at higher temperatures the camera is also off by 600 degrees (in other words, is it a linear thing?), or does it exponentially go out of wack, OR does it exponentially get corrected the closer it gets to the sensors native color temperature rating??

Would Red's color temp reading be a curve, a parallel or an acute angle if graphed against real temperature?

[JFirestone]

Does anyone know if at higher temperatures the camera is also off by 600 degrees (in other words, is it a linear thing?), or does it exponentially go out of wack, OR does it exponentially get corrected the closer it gets to the sensors native color temperature rating??

Would Red's color temp reading be a curve, a parallel or an acute angle if graphed against real temperature?

I don't know, I'm off to Sundance in the morning, when I get back I'll test it if no one else has. Also I will upgrade to the new Beta build and see if anything has changed. The good news is that if my observations are correct, the blue channel is being boosted more than it should be to get proper white balance, and that is adding to the noisier blue channel. So calibrating the color temperature should also reduce the apparent noise we get in the blue channel at 3200.

[Manfred Lopez]

I don't know, I'm off to Sundance in the morning, when I get back I'll test it if no one else has. Also I will upgrade to the new Beta build and see if anything has changed. The good news is that if my observations are correct, the blue channel is being boosted more than it should be to get proper white balance, and that is adding to the noisier blue channel. So calibrating the color temperature should also reduce the apparent noise we get in the blue channel at 3200.

Wow, so this is (possibly) the reason for those noisy shots that were posted earlier about the blue channel. It all now starts to make sense.

[JFirestone]

Wow, so this is (possibly) the reason for those noisy shots that were posted earlier about the blue channel. It all now starts to make sense.

Don't get me wrong, the blue channel will always be noisier when shot at 3200. I'm just saying that it shouldn't be quite as noisy if it is calibrated correctly. That is if my findings are accurate. But even calibrated correctly it is still having to amplify the blue channel to balance to 3200.

[jbeale]

The in-camera display of color temperature after an auto-white balance may not match a calibrated meter, but that does not necessarily mean the final image as exported from Redcine, etc. actually has a color cast. The camera's reported color temp. is just a number (ideally, it should match the standard), but probably what you care most about is the color balance of the image data.

The camera's reported color temperature and tint offset (based on a truly neutral target image) will match an external color meter reading under all lighting conditions, if and only if the camera's R,G,B spectral sensitivity curves match up with whatever sensors are used in your color meter.

My guess is those curves don't match. If so, the in-camera reported color temp cannot be calibrated to match a meter under all lighting conditions, although it may come close.