Home Tutorials Should you use ND filters? Motion Blur in Timelapse

Should you use ND filters? Motion Blur in Timelapse

32 min read
5
0
967

Motion blur is a very important topic when it comes to distinguishing an amateur timelapse from the more professional one. In this article, you’ll learn how to make your timelapse shots smoother and more pleasant to watch by controlling the amount of motion blur and how to choose the right ND filter for your camera and the type of work that you do.

If you’re not sure what motion blur is, where it comes from and what is the 180 degree rule, before continuing check out my explaining article HERE.

Applying 180° rule to timelapse photography

The 180° rule (if you’re not familiar with that, once again, check this post first: POST) is most of the time explaining for 24, 50 or more fps video. But how about timelapse? When you’re shooting, for example, an interval of 2 seconds, you get 0,5 frames per second. Of course, that’s ridiculous calculation (there are a few situations when you can say you capture only a part of a frame, for example, interlaced video or HDR timelapse 😉 but it’s for sure not important here).

So, let’s say you want to stick to the 180° rule in your timelapse shot. What shutter speed should you choose? The simplified method is to take your frame interval, in our example it’s 2s and divide it by 2 – the shutter speed would be 1s.

However, cameras count intervals as shown in the picture below (for most built-in intervalometers, external devices just release the shutter, so shutter speed won’t affect the interval). It’s different than in a video mode, cause in video changing shutter speed won’t affect the interval. In timelapse, if you set your interval in camera for 2 seconds, and exposure time for 1s, you’ll get overall interval time of 3s, because the camera starts counting interval after it takes a picture, not after the previous interval finishes. With shorter exposure time, the simplified method is good enough, but the longer the exposure time, the less accurate that calculation is.

real-interval-time

So, that means the 1 second exposure time won’t give us a real 180° shutter angle. But the facts is, 1” exposure with 2 s interval may give you already pleasant image 😉 If you want to use the real 180° rule for 1” exposure, you should shorten the interval to 1”. Now you get 1 second exposure and 2 seconds (1” exposure + 1” interval) overall interval time.

Choosing the right shutter angle

I assume that you’re familiar with the shutter angle concept. If not, you probably missed my Cinematic motion blur article, which I mentioned twice already 😉

It’s easier to talk about shutter angle, not shutter speed, because timelapsers use all sort of frame intervals, sometimes it’s 1 second and sometimes even 5 minutes. I would say, that timelapse photography loves long exposure. As you know, you can use shutter angle up to 360°. For video shutter angles above 180° are very rarely used, but for timelapse it can give you nice ultra smooth movement (it’s possible to get the 360° shutter angle in low fps video mode, not in photo mode).

Of course, I won’t tell you that you should use 360°, 180° or 90°. That’s something that you should choose on your own, according to your experience. I could only suggest – most of the times the 180° shutter angle would look good.

Interested in Timelapse, Hyperlapse or Stop-Motion? Feel free to join the Facebook Group, where you can find BTS content of my tutorials: https://www.facebook.com/groups/timelapse.hyperlapse/

There is never one solution for timelapse. It always depends on what you’re shooting. Let’s take this example:

Your subject is a bridge construction site. The workers use a technique to slide the previously prepared bridge onto the supports (doesn’t really matter what it is, the point is that’s a very slowly but constantly moving object). Let’s say it will take 4 hours and you’re shooting with an interval of 24 seconds to get 10 seconds shot. If you want to use the 180° rule here you should choose the shutter speed around 12 seconds. The bridge moves constantly for all the time, so 12seconds of exposure will slightly blur it (of course it depends on the distance that the bridge covers in this time). Finally, you’ll get natural, filmic blur on the bridge.

On the other hand, the workers are constantly walking around and checking, if everything is okay. They could be even invisible in 12 seconds shutter speed, because they move a lot faster than the bridge. So, if the bridge is your main theme, 12s would be good, but if you want to show the work of the people around the bridge, you should set the much faster shutter speed, for example, 1-2s. Everything depends on what final effect you’re looking for, there isn’t one good set of settings.

P1260624-2

ND filters

I said earlier, that timelapse loves longer exposures. So, why there are so many timelapses with very jerky movement, for sure shot on very high shutter speed? The answer is simple – because those timelapsers are not able to set long enough shutter speed. That’s a struggle for timelapsers: you’ve got a bright sunny day and you want to nicely blur the moving people, but the light meter in your camera says, that you should use an exposure time of 1/500 even with the aperture closed all the way down. Overexposing your images isn’t a solution. Here comes the ND filter.

I really like the Film Riot’s explanation of NDs – it’s like sunglasses for your camera. It’s a piece of glass that (in most cases) you mount in front of your lens. It shouldn’t change anything in the image – the colors, contrast etc should stay the same as without the filter. It only makes your image darker for the specified amount of stops.

There are generally two reasons for doing that: in the video to open the aperture (to get shallow depth of field) while maintaining the shutter angle at 180° and in photography to extend the exposure time. There could be more reasons for doing that, for example to compensate the minimum 3200 ISO for S-log in a7s, but it’s not important for timelapse.


For youtube version click HERE.

 

As you can see, in the gif above I used two different shutter speeds for a quick timelapse. The other parameters stay the same, because I used an ND filter – Hoya PROND1000. These two shots were rendered with a JPG files straight from the camera. The main thing of course is the motion blur. As you can see, cars on the right shot are just “blinking” in the frame, while the left shot has lots of motion blur, which creates a feeling of very fast moving vehicles. The second thing to consider is flickering. In the gif there are JPGs straight from the camera, without any anti flickering filters. As you can see (if you need better quality, check the youtube version HERE ) the right shot has visible flickering problem (the easiest way to notice that is to look at the sky), while it is far less visible on the left one. There are also differences in the sky color. Well, I said the NDs shouldn’t affect colors on your pictures, but they do.

nd-tabela

There are basically two different naming in the NDs. Let’s take as an example the filter, that provides 3 f-stops of light loss. Some companies, like Hoya, would name it ND8 (that’s because it provides 8 times less light than without filter), and some would name it ND0.9 (that’s from the parameter called optical density of the filter). Of course, you’ll get the same effect using ND0.9 and ND8. You can also find the naming od NDx8, which is the same thing as ND8. It can be a little confusing at the beginning, but many producers or even sellers gives you the information about f-stop light loss and that’s the parameter the most interesting for me. If you can’t find that, use the simple table I put above that combines different naming and parameters.

The last column in the table might be interesting for you. I assumed that with a constant ISO and aperture (no matter what value), you get a shutter speed time of 1/500s to get the correct exposed image without ND. In this column you can find how applying the specific type of ND filter affects the exposure time. Of course, that’s simple calculation (1 stop light loss gives you 2 times less light, so the exposure extends twice), but I know that not everyone understands it correctly at the beginning of their photography adventure.

ND faders – wonderful versatile solution or a piece of crap?

If you’re looking for ND, you probably saw that there are also variable ND filters available on the market. The cheapest start from around 20$ and provide the range of about 2-8 stops. So the calculation is simple –  you get equivalent of a few filters cheaper than when buying it seperately – looks like the best solution available. But it’s not. It’s a very handy tool, but if you decide to buy it, you’ve got to know the disadvantages of variable ND’s.

Variable ND is two polarizing filters – a circular and linear. When you change the position of one of them, the images get brighter or darker, as you can see in the gif above. The principle is simple – more glass you place in front of your lens, less quality you get. And this kind of ND is already 2 pieces of glass.

The second disadvantage is the color tint. Cheaper ND filter = more tint (sometimes it’s green, sometimes it’s magenta). In fixed NDs the tint also exists, but it’s much easier and cheaper to buy one, that doesn’t have any (or almost doesn’t have).

The last one is limitations of usage. ND faders don’t darken the image evenly in some part of the range. With the longer focal lengths, it can be invisible, the problem is with the wide and ultra wide angle lenses.There are a lot of examples of the black cross appearing in the image when using ND fader on a wide angle lens. And let’s say it –  I love my UWA 10-20 mm lens for timelapse (on APS-C sensor). So that’s the most important disadvantage for me.

What are the NDs great for? For video. If you’ve got enough time and filters to get the exposure parameters that you want it’s good to use the fixed NDs, but it’s great to have a good quality ND fader.

For sure the advantage of using ND fader is that you have continuously changing density of the filter, not only by 1 f-stop steps or even ⅓ stop.

That’s definitely something that you should consider thinking about the type of work you do and the budget you have. I’ll write a full review of my NISI ND Fader in different post, so don’t forget to sign up for my newsletter 🙂

Screw mount NDs – which size to choose

IMG_6678-2

I’ve got a couple of lenses. Of course, I won’t buy a different NDs set for every lens. The solution is simple – you buy the filters that fit to the biggest filter thread you have and an extra cheap adapter. For me, the biggest thread is 82mm in the Sigma 10-20mm.

Of course, the bigger filters are more expensive, but it’s still cheaper to buy one filter and a few adapters to fit every lens I’ve got. Very small lens looks ridiculous with a filter that big. But if something looks stupid, but it works, it ain’t stupid 😉

gh4-fader-1

Of course, there is a probability that you won’t find an adapter that fits both to your filter and your lens. In this situation, you just buy more ring adapters 🙂 You can screw one adapter to another, wich I use a lot. I couldn’t find that big adapter like 58mm-82mm etc., but it wasn’t a problem to buy a 72mm-82mm and 58mm-72mm. Especially, my Sigma 18-35mm, which I use a lot, has a 72mm filter thread.

IMG_6686-2

ND filters shootout!

I own a few ND filters, including fader, so I decided to test them out for you (and for myself, too ;)). The results are pretty interesting for me. I’ve shot 6 pictures of my wife’s cute teddy bear, one without the filter and 5 with different filters or filter setups. I used also a polarizing filter, because it also causes light reduction by about 1 stop. Of course everything in the camera was manual. The first 6 images were taken with Panasonic GH4 with 0,71x Speedbooster and Sigma 18-35mm at 35mm f5.6 and ISO200. I was changing the shutter speed to maintain the same exposure.

Of course, that’s not fully scientific test, I change the exposure using zebra stripes and judging the results with my eyes and the histogram.

nd-1-1
No filter 1/30s f5.6 ISO 200
nd-1-2
Marumi Light Control ND8 1/4s f5.6 ISO 200
nd-1-3
Hama circular polarizer 1/10s f5.6 ISO 200

 

nd-1-4
Hoya ProND1000 30s f5.6 ISO 200

 

nd-1-5
NISI variable ND set to minimum 1/8s f5.6 ISO 200

 

nd-1-6
NISI variable ND set to maximum 1,6s f5.6 ISO 200

There are a few interesting facts about this test. First one is the sharpness. Every filter causes a slightly less sharp image, but I don’t think it would be noticeable by a viewer. Every one of them has pretty similar sharpness reduction except for NISI fader set to maximum, which is a little bit worse than the other filter or even NISI set to minimum.

hist-nd-1

The most important thing should be the color shifting. So here’s my conclusion: with Marumi and Hama CPL I don’t see any noticeable color shifting. The image with Hoya ND1000 is cooler than the initial picture (it is more bluish). The most important for me here is the fader, and unfortunately, it has the worst color shift, as expected. As you can see in the histograms print screens below (from RAW files in the Lightroom), NISI filter has a yellow-green tint in the image. I was hoping for better performance with this filter, especially it wasn’t cheap. I can only add, that the light reduction specified by the producer isn’t accurate. I’ll write more about this filter in a review on my blog, subscribe to my newsletter to be notified, when it’ll be ready.

hist-nd-4 hist-nd-5

The last thing I want to talk about is the quality of the cross-effect on ND fader. Faders aren’t good for ultra wide angle lenses. I use a lot my 10-20mm lens on an APS-C camera (and GH4 + speedbooster). Even I paid about 130$ for my fader, it has a cross effect, when I set it to the maximum.

I need to tell you one more thing, a very strong vignette in the corners of the image is caused by the speedbooster. Micro 4/3 sensor with 0,71x speedbooster gives in photo mode crop about 1.42, and the lens is designed for the APS-C sensor (crop 1.5-1.6x).

10mm, no filter
10mm, no filter
10mm, NISI fader set to maximum
10mm, NISI fader set to maximum

As you can see above, the image isn’t darkened evenly. there is a cross between corners of the image. If it’s not visible for you, check out the shot of the white background at -2EV below. Of course, the background was lit quite evenly. This situation is only for the maximum density of the filter, around the middle of the range and for minimum the cross isn’t visible. As you can see in the pictures above, for longer focal lengths it’s also invisible.

10mm, NISI fader set to maximum
10mm, NISI fader set to maximum

Of course, the round ND filters aren’t the only ones on the market. There are systems like Cokin, that include filter holder and square filters. It’s a good quality system. I don’t have any of these, so I can’t test it and compare to my filters. I think, especially for the beginners, the round filters are easiest, because you have only one piece of glass. The advantage of the square filters system are that you buy one set of filters, one holder and different adaptor rings (in round filters you can also accomplish that) and the more important – the ability to move graduated filters up and down (graduated filters are good for landscape photography, when you want to darken the sky, but leave the ground bright – that’s the topic for different article).

Source: Cokin filters
Source: Cokin filters

Summary

So, for the conclusion: if you want the professional look for your timelapses, you should get some ND filters 🙂 Try to avoid extra short shutter speeds, unless it’s justified in some way. And the most important as always – go out and shoot! You won’t be able to decide, what shutter angle is best for your scene just by reading tutorials 😉

But I hope my tutorial will help you and you come back here to learn some more cool stuff! 🙂 If you like that, I recommend signing up for my newsletter – now you’ll receive free storyboard template for this!

 

 

5 Comments

  1. Jim Jacobson

    15 September 2016 at 16:12

    Excellent article. Had figured most of this out but had little knowledge of the where and why of things. Thank you.

    Reply

    • Tomasz

      15 September 2016 at 17:14

      Thanks 🙂 yep, that’s kind of basic topic for timelapse 😉 I hope in the future write both basics and advanced tutorials. If you want to check out something more advanced check out Hyperlapse dolly zoom tutorial, that’s not that easy 🙂 Thanks again for the feedback!

      Reply

  2. […] In the picture below you can check out the chart I made during the shot 🙂 I can’t draw, but I wanted to make a shot with my whiteboard, so I decided to just write something, that’s always useful for timelapsers. It’s a simplified chart from my Should you use ND filters? Motion Blur in Timelapse tutorial. […]

    Reply

  3. […] desired shutter speed in hard conditions (especially when it’s too bright). Check out the post: Shutter speed in timelapse – ND filters. If you dont want to miss new tutorials, click the little red bell in the bottom left corner. You […]

    Reply

  4. […] It’s so important that I wrote specific, in-depth tutorial on this subject – Should you use ND filters? Motion Blur in Timelapse. The other important thing here is that long exposure = better low light 😉 You don’t have to […]

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *