Using ND filters to reduce dynamic range?

[ArchRival]

Possible to use ND filters to reduce dynamic range to avoid gross over and underexposure?

Reasoning goes like this:

If we assume sensors have a linear response to exposure (which they nearly do, or try to do), we can substitute exposure for intensity.

ND filters function to scale the intensity, hence dynamic range, of a scene. Which means dynamic range can be compressed. Also, by varying the exposure time, the histrogram can be shifted to a favourable position.

So scenes with a dynamic range exceeding sensor capability can be solved simply by sticking a ND filter on the lens.

True?

 

[night86mare]

sorry, i am losing you because i am not a techie.

how does linear response mean --> nd filter = scale intensity and thus compress DR?

from personal experience with ND filter, exposure DR remains roughly the same whether using 10 stop nd filter on it or not.

 

[ortega]

a full ND will only cut the amount of light going through your lens

a grad ND on the other hand, if used correctly will do the trick

 

[diediealsomustdive]

The simple answer is no - your DR will remain the same and the ND filter cuts by say 2 stops throughout.

The so-call industry practice is to use graduated ND, usually 2 to 3 stops, typically use to cut the sky down by 2-3 stops and thus allowing better rendetion of the foreground that is of less illumination.

Much as I dislike KRW's stuff he sometimes gives easy to read and relatively good stuff. See his write-up on grad ND.

 

[Gengh]

ND filters reduce all incoming light by a given percentage, and does not discriminate against the actual intensity or colour.

Supposing that your highlights and shadows are 10 stops apart (dynamic range is 10 stops). After the ND filter, highlights drop by 3 stops, midtones drop by 3 stops, shadows also drop by 3 stops.

Result: highlights and shadows are still 10 stops apart, no change in dynamic range.

Perhaps you're thinking of the absolute (linear) intensity values? The absolute intensity range is indeed compressed, but unfortunately what counts in photography is the dynamic range in the logarithmic scale - the stops.

 

[giantcanopy]

Too difficult for me to assimilate. though an interesting read, but it will not impact my ( traditional ) way of using ND / GND

ryan

 

[ArchRival]

Thanks for the replies guys.

I guess i got what i wanted. 2 images of a lamp, first without filter, second with filter. Don't have ND filter at hand, so used black polymer filter instead. Tends to give an orange tinge, but it's supposed to be similar to ND filter. The histrogram in first is clustered at the ends while the second image is more clustered in the middle. The filter is not of good optical quality, so second image don't look too good.

 

[uzume002]

you can use spot meter to do the same thing. don't even need a ND filter

 

[sabee]

you can use spot meter to do the same thing. don't even need a ND filter

Lost you there, what has spot metering got to do with ND filters?

 

[ionicbeam]

Thanks for the replies guys.

I guess i got what i wanted. 2 images of a lamp, first without filter, second with filter. Don't have ND filter at hand, so used black polymer filter instead. Tends to give an orange tinge, but it's supposed to be similar to ND filter. The histrogram in first is clustered at the ends while the second image is more clustered in the middle. The filter is not of good optical quality, so second image don't look too good.

It seems like the light has diffused across the polymer hence the light in the dark regions. Seems like the light is coming from the diffusion than from the dark areas. How about performing the same test in backlit subjects in daylight?

 

[uzume002]

first pic, prob. use auto metering.

AV mode, adjust 0 to +1 ev, point the spot meter at the brighest part,lock expose , reframe and prob. you will get around the same histrogram as the 2nd pic.

nothing to do with ND actually

 

[HLZQ4]

Hi Archrival, looking at your example, it seems that your neutral density filter is not very neutral beacause if you look at the histogram, some colours of light are missing, which might explain your orange hue. A neutral density filter is supposed to reduce intensity across all wavelengths (ideally though). So by sticking the ND in your case, you scaled down the intensity, and in the process moved the histogram from the right to the middle. But at the same time i think you also scaled and removed the clipped shadows. As to whether a ND can help to reduce the dynamic range, I might not think so. A GND might help in this case.

 

[night86mare]

Thanks for the replies guys.

I guess i got what i wanted. 2 images of a lamp, first without filter, second with filter. Don't have ND filter at hand, so used black polymer filter instead. Tends to give an orange tinge, but it's supposed to be similar to ND filter. The histrogram in first is clustered at the ends while the second image is more clustered in the middle. The filter is not of good optical quality, so second image don't look too good.

not like it makes any difference here, really.

there is no shadow detail to speak of. you should do the test with something in the shadow.

the second just looks like a hazed out version of the first, with a yellow filter added, and contrast dropped.

add that along to the fact that the histogram isn't going to be very helpful due to thoroughly different color casts.......

 

[sabee]

I think you need to analyze what the histogram is showing you. In the first shot, there are only 2 "parts" to the histogram, representing the lamp which is entirely white at the far right of your histogram, and everything else which is almost completely black on the far left.

When you slapped on the polymer, it diffused the light across the polymer itself such that there are no more dark areas. It did not compress your dynamic range in any sense but instead added tones that took up the middle part of your histogram instead. If you had any detail in the dark areas, they would still be outside of the histogram and not visible.

 

[ArchRival]

http://www.cokin.fr/filtres.html?=#153.
See the ND4 results.

 

[night86mare]

http://www.cokin.fr/filtres.html?=#153.
See the ND4 results.

before you go off on a rampage, note the tones

the nd4 results are darker, because the nd4 choked the exposure, not increased DR.

 

[eyes]

Technically ND filters will not reduce the dynamic range but derived or perceptive decrease in the dynamic range of the scene can be attained through these filters.

 

[Gengh]

http://www.cokin.fr/filtres.html?=#153.
See the ND4 results.

The Cokin pics simply demonstrate the light cutting ability of the filters, you can achieve the same before/after effect by underexposing the scene by 3 stops.

And in your black polymer experiment, other than the fact that the polymer is diffusing a lot of light (lowering contrast and potentially dynamic range) and not neutral at all, I also have to point out that the horizontal scale of the histogram is linear, representing pixel values from 0 to 255. If you don't manage to get the same effective exposure in both shots, the histograms are going to look very different. This means that if you want to carry out this experiment properly, you need to at least use spot metering on a calibration object like a grey card for both shots.

 

[ArchRival]

before you go off on a rampage....

wat tokking you?

the nd4 results are darker, because the nd4 choked the exposure, not increased DR.

Please explain "choked the exposure".

 

[ArchRival]

http://www.cokin.fr/filtres.html?=#153.
See the ND4 results.

left is w/o filter, right is w filter. The histrogram with filter is more shifted towards the middle. Still get a peak at on the left but the peak on the right is gone.