Are ND filters required at low light levels?

G

guyburns

Active member
I'm looking into the accessories I'll need when I eventually get a GH2. One of the accessories is ND filters, which I thought were needed under bright conditions to reduce the light so that the shutter speed and/or aperture can be kept at the appropriate setting.

Formatt seem to have a good reputation, but this bit from their website I found confusing:

Having undertaken detailed research with DOPs and other users of High Definition (“HD”) equipment, it became apparent that HD cameras require a more optically pure grey neutral density filter to avoid some of the colouration issues which are inherent in their use, particularly at low light levels.​

Do they mean "high light levels", or is there something I don't understand?
 
it sounds like they mean to correct oversaturation of colors with a grey ND filter to flatten the image
 
Are they referring to the infrared bleeding issue?

"Color shift is a critical factor in digital photography. This color shift is directly affected by the infrared spectrum. Specifically, in high definition photography, the sensitivity of imaging sensors to IR has become a challenge. This is particularly evident when using standard neutral density filters that absorb visible light but pass infrared.

Unlike many photographic films, the CCD or CMOS of a digital camera is inherently susceptible to infrared contamination, even when the manufacturer attempts to reduce this problem by adding an infrared absorbing filter on top of the image sensor. Because these filters do not have a sharp cutoff at the infrared wavelengths there has to be a compromise in order not to exclude visible red wavelengths. This results is some infrared wavelengths being allowed to pass.

While minor infrared contamination does not normally cause much of a problem for visible images, when using conventional neutral density filters especially denser grades, this issue is multiplied by the filter factor.

Tiffen has now addressed this problem with the introduction of its exclusive line of IR/ND filters. This is particularly important as it now allows for the use of the denser grades of Neutral Density filters needed for these light sensitive cameras while, at the same time, maintaining complete color balance."
 
I assume this info on the B&H site is nonsense – a B+W ND filter that increases light transmission. Or am I missing something?

A B+W 62mm #103 Solid Neutral Density (ND) 0.9 Filter (Single Coated) represents an innovative design in lens coating. Single coating is applied to one layer on each glass surface. This improves light transmission from about 92% (in uncoated glass) to about 98%. Single coating also has certain anti-reflective properties that help prevent against ghosting and internal flaring.​

How is a novice supposed to learn about ND filters with confusing statements like that?
 
You're reading too much of the in-depth design information and not enough about what ND filters actually do, though you do seem to have a basic understanding of what ND filters do.

To clarify, ND filters are designed to reduce the amount of light entering the camera without affecting the color spectrum, hence the "neutral" in "neutral density" filtering. ND filters, generally, are not useful in low-light situations since their sole purpose is to reduce exposure. There are exceptions to every rule, but I have never been shooting in a low-light setting where I've needed ND.

The reason for getting ND is to allow more control over aperture and shutter in brighter light. By reducing the amount of light coming into the camera, one can use a wider aperture or slower shutter speed... or both. For example, let's say the shot requires the depth of field attained at f/4 in bright daylight. Well, that could be achieved by increasing the shutter speed, but a high shutter speed has an entirely (and not always desirable) different look. To be able to shoot f/4 at 1/60 shutter in bright daylight, it will be necessary to reduce the amount of light coming into the camera using some other method. Enter the ND filter.

Further, if two parts of the frame are drastically different in light levels, it can be difficult to expose for both. Bring one up to the proper level and the other is overexposed. Bring the other down to proper exposure and the one is too dark. By using a grad ND (aka graduated ND), which has ND filtering on half the glass, the hotter section is knocked back a bit so that both areas can be exposed properly without losing too much to over/underexposure on either. Grad ND can be found hard or soft edge. Hard edge is good for knocking down specific structures, like buildings, that are overexposed but have a defined border. The soft (or feathered) edge grad ND is useful for knocking down images that may not have a defined edge between lighter and darker areas, or where a hard edge is going to be noticeable, such as landscapes. Use a feathered ND to bring the sky down and expose for the ground.

Formatt, Schneider (B+W), and Tiffen are the ones to look at. Don't worry about their explanations of coatings. Just be sure you don't cheap out on ND filters, and also be sure to get a decent selection. ND 0.3, 0.6, 0.9, and 1.2 make for a nice starter kit. You can layer if you need more. Get a grad 0.6 and 0.9 as well. Soft edged grad is a bit more versatile.
 
Last edited:
guyburns said:
I assume this info on the B&H site is nonsense – a B+W ND filter that increases light transmission. Or am I missing something?
A B+W 62mm #103 Solid Neutral Density (ND) 0.9 Filter (Single Coated) represents an innovative design in lens coating. Single coating is applied to one layer on each glass surface. This improves light transmission from about 92% (in uncoated glass) to about 98%. Single coating also has certain anti-reflective properties that help prevent against ghosting and internal flaring.​
How is a novice supposed to learn about ND filters with confusing statements like that?
Click to expand...


my guess is that it means the following:
* with standard coatings, 8% of the light is reflected, 92% goes into the filter; out of that, on a .9 ND filter, an additional 79.5% is absorbed by the filter and transformed into heat, 12.5% goes to the sensor
* with their improved coatings, that changes to 2%, 92%, 85.5%, 12.5%
the advantage of this coating is therefore reduced reflections, which for example can create ghosted highlights when shooting some candles in low light, as in this video:
http://www.dvxuser.com/V6/showthrea...namorphic-lens-look-without-Anamorphic-lenses.

as c2v said, you're overthinking this...

* ND faders = very convenient but bad image quality
* resin filters for cokin = convenient but bad image quality
* screw filters = not very convenient but good image quality (already at $25 each for 67mm, or $50 each for 77mm)

you may want to check my ND filter tests:
http://www.similaar.com/foto/lenstests/lenstestsn.html
 
Thanks for the responses. Cokin had already been crossed off my list courtesy of a comment by Barry Green. I had a look at Formatt – and you'll get a kick of my reasoning here – I didn't like the grammatical mistakes on their website and in their PDFs, so they're crossed off the list. I came across a few comments about Tiffen on dpreview; the comments, as well as their cheap prices, put me off a bit. So it's down to B+W and Hoya.

Norman – you're trying to trick me. For standard coatings, I finally worked out the maths: 8% reflected, 12.5% to the sensor (3 ND = 3 stops = 1/8), leaving 79.5% absorbed in the filter. What threw me were the next 4 numbers – then I realized they should have been: 2%, 98%, 85.5%, 12.5%.

Your testing was informative. I've added it to my references, along with this one: http://www.dvxuser.com/V6/showthrea...my-sharpness&p=1966610&viewfull=1#post1966610

I'm not sure I caught the "ghosted highlights" in that candle video. Did you mean the extra images of the candles that are above the real candles? I'm intrigued how they formed – they're a mirror image of the real candles (pause at 0.57). How did they get there?
 
guyburns said:
Norman – you're trying to trick me. For standard coatings, I finally worked out the maths: 8% reflected, 12.5% to the sensor (3 ND = 3 stops = 1/8), leaving 79.5% absorbed in the filter. What threw me were the next 4 numbers – then I realized they should have been: 2%, 98%, 85.5%, 12.5%.
Click to expand...

you're totally right, it's a copy-paste-then-forget-to-change-all-numbers typo

guyburns said:
I'm not sure I caught the "ghosted highlights" in that candle video. Did you mean the extra images of the candles that are above the real candles? I'm intrigued how they formed – they're a mirror image of the real candles (pause at 0.57). How did they get there?
Click to expand...

yes, those extra images
IIRC, on the thread they explain it's light being reflected by the lens' front element, then ont he back of the filter, then go through the lens and into the sensor; a better coating on the lens element, and/or on the back of the filter, would prevent that
 
You must log in or register to reply here.
Back
Top