auxilary data re D2X noise

[Iliah]

Overall noise is at minimal if camera is set to 5300K;
Noise in green channel is amplified between 2500K and 3200K due to white balance scheme.
At 2800K blue noise is amplified 1.5 stops.

 

[Steve S]

Huh? :-/

Does this mean we should always set our cameras to 5300k?

 

[Iliah]

Dear Steve,

I mean that decision to use 80A colour compensation filter can be based on that.

 

[cwilt]

Are you saying that 5300K is the native color temp of the sensor?

I know nothing about filters but would like to know what this has to do with an 80A filter?

 

[Iliah]

For D2X at 5300 all three colour channels are least affected with colour balancing coefficients. 80A filter is conversion filter that raises colour temperature of incanescent light close to 5300..5500K.

 

[cwilt]

Damn your quick!

I just went looking for info on 80A. It adjusts WB of tungsten to that of daylight film. The conclusion I get from this, is that if you use a filter to adjust WB you will have less noise than if you change the WB coeffiencts of the red or blue channels.

Right or wrong?

 

[Iliah]

The conclusion I get from this, is that if you use a filter to adjust WB you will have less noise than if you change the WB coeffiencts of the red or blue channels.

Of course. Under incandescent light blue channel is to be multiplied 2.6..4 times (3200..2500K), that effectively is 1..2 stops underexposure. As we normally shoot pretty high ISO under incandescent, that makes for a fair amount of noise. 640 ISO underexposed 2 stops and corrected in postprocessing - how is that for the noise?

 

[JeffKohn]

How much impact does 80A filter have on exposure (ie how much light loss)?

I'll have to think about whether one of these would be worth the purchase. I usually use flash indoors...

 

[Chris101]

You lose 2 stops. This was a very popular filter back in film days. It's blue.

 

[cwilt]

2 stops of under exposure noise or 2 stops slower shutter. Thats a tough trade off to decide. For a static shot with a tripod it would not matter, but on a moving subject in poor light 2 stops can be the difference between a stop action shot and a blurry one.

I picked up an 80A and an FLD filter to experiment.

I think the first test will be to look at coeffiecients with and without filter. These shots will include a whibal card to sample WB from.

 

[Creative Edge]

the 80a filter is a 2x filter which equals
one stop, not 2...

 

[Iliah]

the 80a filter is a 2x filter which equals
one stop, not 2...

and below 3200 red channel is scaled down - up to 0.75 at 2500.

Filter factor depends on the brand, but for top brands 80A has factor of about 2.2, very slightly more then 1 stop.

For usual incandescent bulbs below 150W perfect neutrality can be achieved with KB20 filter which is rated to decrease light by 1 2/3 stops on film, but has much wider transmission in greens and blues then 80A. For digital it seems that "non-orthodox" filters like KB series work better due to narrower absorption - we do not want the green channel to be affected that much as it is often with classic film filters. I like KB 12, which is close to 80B and preserves the atmosphere of in-doors; and of course I do not apply exact white balance. "Blue" filters factor determined for the film is usually lower for digital in any case.

In studio using halogen lamps full CTB light filters are easier then lens filters. Light filters also allow for more creativity. BTW, there are also special sleeves for fluorescent lamps.

 

[cwilt]

What am I doing wrong?

WB set to 5300K using an 80A filter with standard incandescent lighting. The blue channel is under exposed a lot! I thought at first the D2x was using its colormeter so I tried the exact settings on my D70 and got the same image.

Enlighten me, please!

Subscribe to see EXIF info for this image (if available)

PS. I like VR. This was a 1/2 second exposure.

 

[Chris101]

An 80A filter is designed to convert light from a tungsten photo-flood bulb (3400K) to daylight (5500K). A normal incandescent bulb is 500-1000K less than a photo-flood. So your photo still looks warm because the filter has converted the incandescent light to only 4500-5000K instead of all the way to daylight's color temperature.

Here is an experiment I just did. A gray card was placed against a white canvas and photographed it with and without an 80A filter. The shutter was set to 1 second and the aperture was set by the camera. The Exposure Compensation was adjusted until until the gray card fell in the center of the histogram, and the white was at, but not climbing, the right wall. This EC was used for both exposures.

Here is the shot with no filter, WB set to incandescent, and the color temperture was calculated to be 2500K by ACR's eyedropper:

Subscribe to see EXIF info for this image (if available)

Now the 80A filter was put over the lens, the WB set to daylight and then the color temperture adjusted to 4850K with ACR's dropper. A quarter of a stop of exposure (-0.25) was taken from this exposure with ACR to bring the gray card to the same shade as the above image:

Subscribe to see EXIF info for this image (if available)

First note that the temperture of the filtered light is 4850K and not 5300K. This is consistant with the initial light source being 2500K and not 2900K. The tint portion of the white balance was set to zero for both pix so as not to complicate the math. This causes a slight color shift between the two shots.

Second look at the difference in exposure. The difference is (7.1-0.25)/3.5 = 1.96 f-stops. Michael, the published filter factor of 2X is for white light. Incandescent light has very little blue light in it however, so this blue filter attenuates yellowish incandescent light more than it does the even mix of colors in white light. This means the effective filter factor for an 80A filter is closer to 4X (actually 3.8X) than it is to 2X.

I used to shoot slide film with the 80A filter. Using a 2X factor always gave me underexposed film with photos made under regular light bulbs.

 

[Creative Edge]

Chris...
you are right.. my mistake...
real life and what the makers publish sometimes do not coincide...
thanks...

 

[Iliah]

Dear Chris,

Thank you for your examples and explanation.

To compare noise in shadows, gray card is exposed to about 60..70 RGB, and strongly OOF. To facilitate comparison, tripod helps.

IMHO exposure for 2500 and 3400 will be different, as filter factor indeed varies with light spectrum. And it also depends on the brand of the filter.

 

[cwilt]

If I do a custom WB in NC from the whibal card in the shot the blue coeffiecent is increased by what looks like 2 stops (if thats what the grid means) and the red decreased 1 stop. Isn't the purpose of the filter to help with blue channel under exposure which increases noise? I don't see the gain at the moment. It may be like not seeing the forest for the trees though.

 

[Iliah]

Dear Charles, 1 stop in white balancing a channel is 2 times.

 

[Catz]

for Iliah or anyone with experience with this filter

Ilah, How do you feel about the ExpoDisc? I was thinking of getting one and try it out. It sounds like a great filter for WB.

Which one would you recommend? The warming or the original or both?

Thanks for the info,

 

[Iliah]

Dear Melissa,

I beleive it is not better (to say the least) then the ambient sensor you have on your D2X. I tried ExpoDisk https://www.nikoncafe.com//forums/viewtopic.php?t=91