Do I need ND filters?

[skunky1]

Hi

Im a hobby flyer doing mainly videos but some photos.

I know that in bright sunlight the footage looks overexposed but I can dial that down in the settings.

So do I really need filters? Which ones..there are so many. Also there is the variable dial type.

Please advise

 

[Cisco62]

Yes, you definitely need ND filters if you shoot video. Best aperture on Hasselblad lens is between 2.8-4.0 and shutter speed should be set at 1/60 for a fluid render of the video at 30 fps. So you need at least a 8-16 ND if you shoot at sunset or in cloudy weather and a 16-32 ND in bright sun.
I have Polar pro 4/8/16 ND-Polarizing set and I use them both for video and photos.

 

[Meta4]

You don't need ND filters for stills shot from a drone.

I know that in bright sunlight the footage looks overexposed but I can dial that down in the settings.

Proper exposure settings are the way to deal with that.
ND filters aren't for exposure control.

So do I really need filters? Which ones..there are so many. Also there is the variable dial type.

ND filters are a forum favourite discussion topic.
Lots will try to convince you that you cannot shoot video without them.
They'll tell you about something called a 180° rule and use the word cinematic and motion blur.
But the fact is that it's a matter of personal choice.

Here are a couple of videos that explain what it's all about to give you more of an idea than a forum post can.
Watch them and do a little testing to see what you like and what works for you.

 

[Cisco62]

ND filters aren't for exposure control.

So, what do you use them for?

 

[BadWolf1]

You don't need ND filters for stills shot from a drone.
Proper exposure settings are the way to deal with that.
ND filters aren't for exposure control.

So, what do you use them for?

Well they are and they aren't - If the meter in the camera wants to over expose a given scene by 1 or 2 stops it will still do that with an ND filter.

The job of an ND filter is to use a longer exposure for the same aperture. Video looks a bit odd if you shoot 1/1000th 25 or 30 times a second. The eye can tell it is seeing a succession of sharp images, with big gaps. If nothing is moving quickly across the frame it won't matter, otherwise a little motion blur it puts the illusion of motion back, hence the rule of thumb of shutter time is half the frame rate.

 

[Starz]

If you go that route, 60.00 shipped (USA). Mods remove if inappropriate.

 

[Not A Speck Of Cereal]

They still are not for exposure control. You can reduce your exposure by normal photographic means (higher shutter, smaller aperture, lower ISO).

I know that in bright sunlight the footage looks overexposed but I can dial that down in the settings.

Yes. To correct the exposure in this scenario, you already have the controls you need.

They ARE for frame-rate control (which is not an exposure modifier). One could say "well, I'm controlling exposure to control frame rate". Why not just say "I'm controlling frame rate", because in the end, the exposure should be the same.

When you put on an ND filter, you are reducing light and so (for proper exposure) compensate with an exposure control (such as shutter speed) to give you the SAME exposure result.

That is, the histogram in a video editing app (showing luminance values of your video) should not be different between frame-rates (of the same scene, same light, photographed properly).

Chris

 

[SkyeHigh]

You don't need ND filters for stills shot from a drone.
Proper exposure settings are the way to deal with that.
ND filters aren't for exposure control.


ND filters are a forum favourite discussion topic.
Lots will try to convince you that you cannot shoot video without them.
They'll tell you about something called a 180° rule and use the word cinematic and motion blur.
But the fact is that it's a matter of personal choice.

Here are a couple of videos that explain what it's all about to give you more of an idea than a forum post can.
Watch them and do a little testing to see what you like and what works for you.

I have to disagree on the photograph side Meta4. In Australia it is extremely bright and to get an exposure at a wide aperture for the camera’s best lens definition often requires an ND4 and sometimes an ND8 in the desert regions. I know you can shoot at very high shutter speeds but they’re not optimum for CMOS and CCD sensors.

Totally agree for video use and necessary for our slower 25FPS.

 

[Meta4]

I have to disagree on the photograph side Meta4. In Australia it is extremely bright and to get an exposure at a wide aperture for the camera’s best lens definition often requires an ND4 and sometimes an ND8 in the desert regions.

I happen to have quite a lot of experience photographing in Australia in bright conditions.
Your concerns are exaggerated and unnecessary.

 

[Repaid1]

Well they are and they aren't - If the meter in the camera wants to over expose a given scene by 1 or 2 stops it will still do that with an ND filter.

That is a bit odd of a statement, as one should know not to use auto with a ND filter attached?

Quite simply for the OP the mere fact that you had to ask if you need ND's means you don't as your happy with your output at present. If however you would like to experiment with what a ND can do, then by all means do so. But at this point it is safe to say that you do not "Need" them however you might "want" them.

 

[BadWolf1]

Me > If the meter in the camera wants to over expose a given scene by 1 or 2 stops it will still do that with an ND filter.

That is a bit odd of a statement, as one should know not to use auto with a ND filter attached?

Quite simply for the OP the mere fact that you had to ask if you need ND's means you don't as your happy with your output at present. If however you would like to experiment with what a ND can do, then by all means do so. But at this point it is safe to say that you do not "Need" them however you might "want" them.

Through the lens metering just thinks everything has got a bit darker with the ND attached. If auto works without the ND, it works with it. If it is two stops off without, it will will two stops off with it - ND doesn't correct metering errors, but gives longer exposure for an aperture.

Need vs want ? Don't put up with staccato video for want of filer.

 

[SkyeHigh]

I happen to have quite a lot of experience photographing in Australia in bright conditions.
Your concerns are exaggerated and unnecessary.

Well we have to disagree again. I have been in the film, stills and TV industry most of my working life. And I live in Australia so I happen to experience the extreme conditions and did so for the whole of last year driving around the entire continent photographing.

We clearly have different opinions and experience.

 

[jlatas]

Most people use an ND filter to get their shutter speed to be about 2x their frame rate so that their video has a more realistic and cinematic motion blur. For example, a video with motion,recorded at 1/1000 second, doesnt have natural movement in it. Especially if there are people in the video.

 

[Meta4]

And I live in Australia so I happen to experience the extreme conditions.

You're not the only person that lives in Australia, but are the only person I've ever heard suggesting that higher shutter speeds cause some sort of issue.
I've not heard of that before and can't find anything referring to it now.
Can you point me toward some source to confirm this new revelation or give any details to explain it?

 

[franknitty69]

you don't 'need' ND filters, but they are useful. Mainly for reducing shutter speed without altering your aperture or iso. They can also reduce glare or increase color saturation/contrast. They also have polarizers which will help to polarize the image (reflections).

I keep a 16/PL on almost all of the time during sunny/partly cloudy flights. For clear skies at high noon, i'll go with the 32. Polar Pro are my favorite ND filters.

 

[Meta4]

They can also reduce glare or increase color saturation/contrast.

The neutral in Neutral Density is because the filter does nothing at all the the look or quality of the image.
It doesn't affect glare or saturation.
It just cuts light getting to the sensor.
Your ND16 prevents 94% of the light getting through but doesn't change the look of the image.

 

[Cisco62]

Well they are and they aren't - If the meter in the camera wants to over expose a given scene by 1 or 2 stops it will still do that with an ND filter.

The job of an ND filter is to use a longer exposure for the same aperture. Video looks a bit odd if you shoot 1/1000th 25 or 30 times a second. The eye can tell it is seeing a succession of sharp images, with big gaps. If nothing is moving quickly across the frame it won't matter, otherwise a little motion blur it puts the illusion of motion back, hence the rule of thumb of shutter time is half the frame rate.

It will overexpose if you use auto exposure mode not in manual.
I totally agree with your second sentence. Nd filters are used in photography to increase shutter speed to get motion blur when you can't (or don't want to) close the aperture.

 

[SkyeHigh]

You're not the only person that lives in Australia, but are the only person I've ever heard suggesting that higher shutter speeds cause some sort of issue.
I've not heard of that before and can't find anything referring to it now.
Can you point me toward some source to confirm this new revelation or give any details to explain it?

No new revelation, far from it. I thought you might have been familiar. Every CCD and active-pixel sensor (AKA CMOS sensor) manufactured have a number of noise related issues, the most obvious being random noise and fixed pattern noise that shows largely with long exposures in low light situations. Some of this can be mitigated with dark-frame imaging where a second image is taken with the camera shutter closed and the common noise in each imaged can be removed. However that isn’t possible in sensors with electronic shutters that are constantly exposed to light. There are also saturation issues with imaging cells that are constantly exposed that can lead to reduced contrast and detail. There’s more than one reason for having a physical shutter in a DSLR.

In the case of very short exposures an additional form of noise called shot noise is apparent. It is a quantum effect caused by random distribution of photons over very short periods. The longer the exposure the more the photons are distributed evenly across the sensor. Canon and Sony as the leading developers use several techniques to reduce both short and long exposure generated noise in their in-camera image processing engines.

There isn’t a lot of publicly available reference material but searching for ‘shot noise’ brings up quite a few academic papers detailing the cause and effect along with the other forms of inherent noise in Silicon-based sensors.

I was an honoured guest spending six very privileged weeks at Ota-ku, in Tokyo (Canon Global HQ) kindly arranged by my cousin after signing an NDA. She’s behind some of the work that went into the sensors developed for the new R-Series cameras. Her husband is an optical engineer working with the L-Series range of lenses. Their eldest daughter is now at university studying for her PhD in semiconductor design. Seems technology runs in the family.

 

[BadWolf1]

In the case of very short exposures an additional form of noise called shot noise is apparent. It is a quantum effect caused by random distribution of photons over very short periods. The longer the exposure the more the photons are distributed evenly across the sensor.

That explanation of shot noise is - let's say - "novel". It would mean that exposure from studio flash units which deliver their output in 1/10,000th or less would be noisy. Which generally, they aren't.

Lets say a 1 second exposure causes a pixel to count 10,000 arriving photons. One might assume every 1/1000th you'd get exactly 10. Except it doesn't work gaps between conversions form a Poisson distribution - the photons aren't perfectly equally spaced and not everyone gets counted. In some thousandths you might get 8,9,11 or 12 but the average will be 10. If you shoot 1/5000 the average will be 2 but you'll get some 1s and 3s
The issue is not the time, but the number being counted The 10,000 in 1 sec might vary by +/- 100 (square root of 10K) but that's 1%. 100 might vary +/- 10 or 10%; 10 might vary +/- 3 so 30%. The way to get low shot noise is not to count small numbers of pixels.

If you have a sensor with a deep "well" (i.e. big pixels) you might count 10,000 photons (+/- 100) for light grey and divide by 50 to scale the the sensor numbers to the range in a JPG. So it's 200 +/-2 on a scale of 256.
With a shallow well (small pixels) it might only 400 (+/-20) for the same grey, this time it's only divided by 2 give 200 +/- 10 on a scale of 256.
And of course when you turn the ISO up instead of counting 10,000 or 400 it might be 2500 (+/- 50) or 100 (+/-10) which scaled to 200, becomes 200 +/- 4 and 200 +/- 20

Small well depth, and/or ISO give high shot noise, because they're dealing with few photons.

In bright light at ISO 100 if you want to use f/2 it might 1/8000th. But there's no difference in noise between f/2 @ 1/8000 and f/16 @ 1/125. (or f/2 + ND @ 1/125)