To 85 or not to 85 ????

[Guest]

Hey everyone,
Has anyone had real experience with shooting outside with the tungsten preset & an 85 filter ? I remember something about the CCD's being tungsten by nature. Hmmm...

Oh, and any thoughts on coral filters over 85's ? Brings back the ARRI days, I miss film. Sob.

Thanks in Advance , Steve

 

[Barry_S]

Re: To 85 or not to 85 ????

White balancing your camera to tungsten light and using an 85 filter to warm daylight to the color temperature of tungsten makes absolutely no sense.  CCDs are not natively balanced for tungsten or any other color for that matter.  There are three CCDs--each with a red, green or blue filter.  The R, G, and B signals are recorded and reassembled into a color RGB signal.

White balancing equalizes the voltages from each CCD while your camera is pointed at a white reference.  In an RGB system, white is always made up from equal proportions of R, G, and B.

Imagine a situation where you're shooting under bluish light--like open shade at noon on a cloudless day.  You point your camera at a white card and push the white balance button.  The camera checks the voltage from each CCD and finds that the blue filtered CCD voltage is say, twice as high as the red and green CCD voltages.  The camera thinks, "My master says this card should be white so I have to drop the output of the blue CCD so it's exactly equal to the green and red CCD voltages."  Poof--the camera equalizes all three CCDs and outputs equal values of R, G, and B to give you pure white.

You like coral?  Use a coral filter.  You like a warm coral effect?  Balance against a bluish card or drop the color temperature in the menu and use your coral filter.

 

[mastermind]

Re: To 85 or not to 85 ????

I've actually read somewhere that CCDs actaully have better daylight rendition so it is better to shoot DV balanced for daylight. Meaning that if you're shooting indoors, gell your tungsten to maintain a 5600k.
not a fact...just what i read.

 

[Barry_Green]

Re: To 85 or not to 85 ????

It is Michael Phillips' assertion that CCD's are balanced for tungsten light.

I have no knowledge that could confirm or deny that. When I first heard him say it I thought it sounded odd, but after a bit of explanation it made more sense. For example, you can shoot tungsten-balanced film in daylight, with no filter, and color correct in post -- but the results will not be identical to what you would have achieved had you shot daylight-balanced film in the first place. The blue portion of the tungsten film will be overexposed. The film is made to process light in certain proportions, and that's why it's better to use an optical filter rather than relying on post color correction.

This issue came up about a year ago when the DVX was first released, because there was an issue of "purple fringe" around some objects, especially high-contrast objects. It was Michael's opinion that the fringe came from overexposed sections of improperly-balanced lighting conditions, especially using CineGamma. He said that the CCD's were balanced for tungsten, and if you shot outdoors using the daylight preset and CineGamma (with no blowout rolloff), the blue section was in jeopardy of overexposing and causing the pink fringe. He further said that if you used the tungsten preset and an optical 85 filter, you could eliminate the fringe.

A bold assertion, and fascinating, and it sounds plausible. I never tested it, but I probably should -- I do have a glass 72mm 85 filter around here somewhere.

Funny thing, you don't hear much about the "purple fringe" anymore, whereas once it was a raging controversy. Tizzy's posts contained some purple fringe which is what made me think of it again.

MichaelP, I hope I've accurately relayed your opinions and assertions -- if not, please correct anything I got wrong. And if you'd like to chime in here and clarify things, it would be most appreciated!

 

[Barry_S]

Re: To 85 or not to 85 ????

Well, it should be easy enough to test, but I can't say it sounds plausible to me.  The explanation doesn't even seem consistent--the CCDs are balanced for tungsten, but blues are being overexposed?  Wouldn't it be the other way around?

What is true is that CCDs have a spectral sensitivity curve that is most sensitive in the red and yellow portion of the visible spectrum.

http://www.retrievertech.com/Product/ccd_spectral_sensitivity.htm

So CCDs have a higher sensitivity to tungsten light--true, but an 85 filter *subtracts* blue wavelengths of light--it doesn't add light in the red-yellow spectrum.  So--you balance your camera for tungsten--effectively telling your camera that for this given set of voltages you want the R, G, and B output to be equal.  Then you go outside and pop on your 85 filter and what happens?

The 85 filter cuts 2/3 of a stop of light hitting the CCDs and guess what?  You're cutting out mostly blue wavelengths--the area of the spectrum where your CCDs are least sensitive.  Does it matter?  Probably not as there's a lot of full spectrum light outside.  The 85 ends up acting like a neutral density filter.

 

[Barry_Green]

Re: To 85 or not to 85 ????

[quote author=Barry_S link=board=news;num=1082602943;start=0#4 date=04/23/04 at 12:31:29]the CCDs are balanced for tungsten, but blues are being overexposed?  Wouldn't it be the other way around?
[/quote]
Well, no. You're telling the CCD's to expect tungsten light, and then you feed it daylight, which is stronger in the blue spectrum, thus overloading the tungsten-expecting CCD's. That doesn't sound inconsistent...

Keep in mind, I have no real idea what I'm talking about here, this is something I gathered from some of Michael Phillips' earlier posts, and in the posted sample photos it did seem to have some effect on curing the "pink fringe".

Then you go outside and pop on your 85 filter and what happens?

Well, the plan is the 85 filter converts the daylight into tungsten light, matching what the camera is (at that point) calibrated for.

The 85 ends up acting like a neutral density filter.

Sorry, I don't follow you here. We're calibrating the camera to accept tungsten light. We then take the camera outside into daylight. If we expose without the 85 filter, we're going to get blue footage. If you put the 85 filter on, it corrects the color to be what the camera's expecting, resulting in proper color balance. That's not acting like a neutral density filter... it's correcting the color to what the camera has been set to (remember, we're using the 3200k preset under 5600k light).

I think the gist of Michael's argument is that CCD's are inherently calibrated to be at their most accurate under tungsten lighting, and getting them to read proper colors under daylight can be accomplished in two ways: through voltage balancing (white balance correction) or optically (through using an 85 filter). It was his contention that the optical path is more "correct" because there's only so much that the voltage balancing can do, and if the blue portion of the CCD is being overexposed, there's only so much correcting that can be done for that, whereas by optical filtering you reduce the amount of blue light entering the CCD, thus preventing the overexposure situation in the first place.

With that said, I've never heard of anyone ever using an external #85 filter on a video camera. But what about the rotating filter dial on a BetaSP camera: that's an optical filter wheel, isn't it? And if so, they're using glass 85 filters rather than voltage balancing to correct for daylight.

 

[Barry_S]

Re: To 85 or not to 85 ????

Ok, I think we're reaching the point of diminishing returns here    This would be better debated over several pints of local ale or lager--culminating in a tipsy experiment.

Why does the 85 end up acting as a ND filter?  Because you end up with the same color balance for the final image, but a lower voltage being delivered to the CCDs.

The optical balancing vs voltage balancing argument just doesn't make any sense to me.  A filter can only *remove* light of certain wavelengths.  This is no advantage to this over lowering CCD output.

I'm not saying this wacky procedure hurts--just that it isn't going to gain anything.  Although one clear disadvantage is that if you balance to exactly 3400K and use an 85 filter to warm the daylight you could end up with a poor white balance.  An 85 filter will drop 5500K light to 3400K.  But what if you start with 8000K or 10,000K light outside?  The color temperature of light outdoors will vary depending on the time of day, cloud cover, etc.  So your indoor/85 converted balance won't take changing conditions outside into consideration.

Now that's my last word, but you're welcome to have a final go--as long as you buy a round.

 

[MichaelP]

Re: To 85 or not to 85 ????

Barry - You got my reasoning correct. I got the native tungsten CCD from very reliable sources in the video industry. I will go back and get it agian. But my tests have show that the pink fringe is eliminated when shooting with the 85 filter on areas where there are blowouts. This can be seen in a RGB color corrector where the blue channel will have twice as much value in the same area when compared to the R and G channels. By using the 85, the blue channel is cut is in half at time of exposure to have the same values in the RGB channels thus being even (and this white) and not pinkish.

I havev alos found in some shots on a production I was editing that the DP did not balance for outdoors. So one take was very blue. The following take was properly balanced. In the shot, there was a blue sky. I was able to extract the blue in color correction using a gray garbage can in the shot matched to RGB 128 to get proper color balance. What I disovered was that the sky in the incorrect shot remaing a nice deep blue after the correction is post whereas the properly balanced shot in take 2 had a blown out sky and was more white than anything. I have been meaning to do more of these tests to see what the effect is over different shooting situations but haven't had the time. Between a full time job during the day and editing two features in my spare time.... shooting has been minimal of late...

Michael