My latest flow motion video Panta Rhei Wrocław just won december Best Timelapse award on Hollywood International Independent Documentary Awards! If you haven’t seen it, check out the video below.
I’ve decided to extract the blog section from TL video site and create the new one dedicated for the blog content – Beyond The Time. What does it mean for the viewers? More tutorials, more posts, more content 🙂 You can like the new fanpage HERE or join Beyond The Time newsletter HERE. Don’t worry, I don’t like spam either. You’ll recieve emails only with fresh tutorials 😉
So see you on the Beyond The Time website!
This tutorial has been moved to my new website. Please follow that link: http://beyondthetime.net/hyperlapse-dolly-zoom-tutorial/
We’re going level up than shooting a ‘simple’ hyperlapse. In this article I’d like to show you how to accomplish vertigo effect with hyperlapse technique. Using frame by frame shooting you can create really dramatic effects, which is super difficult in standard dolly zoom filming. This kind of shot can diversify your timelapse video. If you’re not familiar with hyperlapse, check out my last tutorial “How to shoot hyperlapse” before reading this article.
This technique has various names: dolly zoom, vertigo effect, push-pull, zolly etc. It was firstly shown in Alfred Hitchcock’s movie “Vertigo” in 1958. Later this effect was used by many directors. You can find it among others in Steven Spielberg’s “Jaws” or Martin Scorsese’s “Goodfellas”.
The dolly zoom effect is when the foreground stays the same size while the background is growing or shrinking (depending on the camera movement direction).
To accomplish this effect you’ve got to use the zoom lens. Basically what you have to do is moving forward (getting closer to your object) and zooming out the lens in the same time or moving backward and zooming in. Simple, right? 😉
FilmmakersIQ made a nice video about dolly zoom. If after this short introduction you’re still not sure you understand what’sthe vertigo effect, check out their video below.
In my opinion, it’s not easy to find a really good location. It’s easier when we’re talking about medium quality spots, but you’re not here to learn how to make average hyperlapses 😉
For this kind of shot you need to find an object, that won’t move or resize during your shot. Basically, it’s not necessary, but let’s discuss only this case as the easiest way to get how vertigo effect works. However, you should remember, that long corridors are also good for that, as the staircase in the ‘Vertigo’ movie.
If you want to use as wide range of focal lengths as possible and get the most dramatic effect you can get, the ideal location is where the object is surrounded by a lot of space in front and behind it. It has to be a straight road or a sidewalk; any curves, especially on the side you are shooting (in front of the object), could make it harder to make. Space in front of the subject lets you move far away with your long focal length and still see the object. Some space behind your theme maximizes the effect because far oriented background gets really close when you use a telephoto lens and extremely far with the wide angle. When the background is close to your theme, this change wouldn’t be that huge. Look at the gif below, it was taken with 14mm (28mm for 35mm equivalent) at the night (darker pictures, background far away), and on the brighter side (background close to the object) focal length was 140mm (280mm for 35 eq.). Look how dramatic is the background change between them, while the main theme is the same size (that’s basically how the vertigo effect works ;)).
I find my locations by walking around with my lens set to the longest focal length. I found out that it’s easier to get close to the object than move far away and still have a good view. There is always something that covers up your theme.Then it’s good to check how your theme looks like with the shortest focal length you have.
TIP: to find a good location use your longest focal length.
I also like to be on the same height that my object is. It minimizes perspective changes on that object and maximizes the effect on the background because you don’t have to tilt your camera.
To be clear, it doesn’t mean that when you’re shooting 6m/20’ tall monument you’ll get a crappy result. These tips show you how to get the strongest effect possible, if you stick to the steps in the shooting section, for sure you’ll get the dolly zoom effect 🙂
Of course if you can’t find a perfect location, you can create it yourself. Find a long, straight road, sidewalk or just some big square and put there an object. It could be anything, even your friend, which can stand still for an hour 😉 As you see in the gif, I used a teddy bear on a chair in the park next to my house.
Another important thing is a line on your path. In zolly shot, if there isn’t any line to follow on the ground and you can’t create the line by yourself I suggest avoiding this location.
It’s extremely hard to create a perfect shot like that without any line to follow, because when you’re using long focal length every centimeter (or inch ;)) of sideway movement will cause noticeable background shift.
In a simple hyperlapse, I sometimes shoot without obvious path on the ground, but in vertigo I wouldn’t try that. If I have to, I am as precise as possible. In the gif below you can see the effect accomplished on very uneven sidewalk, you can see the corridor sideways movement in the distance.
This location also shows you, that you don’t have to search for an object in the center of the frame. Drawing on the “Vertigo” movie, every corridor also gives great results, but it’s harder to shoot. In this article I’ll skip this case, becase if you’ll learn the basics of the hyper-dolly-zoom, I think you’ll know what to do to get this kind of shot 😉 If you want another article, about only the corridor version of this effect, let mi know in the comments.
The goal is to get as wide range of focal lengths as possible. Of course, prime lenses won’t work here (however, later I’ll mention what you can do with prime lenses). You’ve got to have a zoom lens. But you don’t have to use only one lens! When you reach your final focal length with one lens, you can just simply change for a different lens, set the same focal length & shoot again (for example use 18-55mm and 55-200mm lenses).
More about lenses in vertigo effects you can find in my tutorial on Timelapsenetwork.com.
As always, set your camera to manual mode. In the hyperlapse dolly zoom there is one exception: autofocus may be useful here. I believe a lot of lenses, especially for photography, lose focus while changing the focal length.
The main object that should be in focus stays in the same place in a frame during a shot, so it’s easy to choose autofocus points to match this object. Just watch out for anything that can cover up this object, because it can ruin the focus in some frames.
The second topic I want to mention here is the aperture. I would use an example here. I’ve got 14-140mm F3.5-5.6 Panasonic lens. As you probably know, these numbers mean that at 14mm I can set maximum aperture F3.5, and at 140mm it’s F5.6. The same situation is with my set of three lenses for Canon I mentioned earlier. I set the higher minimum value, that’s also 5.6.
These are very common values for that long lenses. Best thing to do here is to fix aperture at 5.6 or higher. When you’ll be changing focal length, aperture won’t change automatically for the smallest value available, which would cause really annoying exposure changes. Set your lens to the shortest focal length available and set the aperture to 5.6. Then you’ll be sure it won’t change (if you’re at M mode of course, but you should be if you’re taking seriously my words 😉 ).
If your lenses are all 2.8 (24-70 2.8, 70-200 2.8 etc) you don’t have to worry about that you lucky man. Or lady 😉
There are a few ways for placing the fixed object in the frame. As you can read in my HYPERLAPSE TUTORIAL you can use grid inside your camera. For example, I used the rule of thirds grid in the shot from Cracow (the ‘shield’ with K). Vertically I want to aim the object to the center of the frame, it was quite easy. Horizontally I positioned it using two horizontal lines in that grid. You can see this example in the picture below.
You can also use focus box, any cropmarks (inside Magic Lantern) or put a transparent tape on your LCD and draw lines with marker. I didn’t try the last method, I read that a few people use it 😉 Personally, I could use it on my 550D, especially that I’ve got there a screen protector foil, but I wouldn’t try that on the touchscreen in my GH4.
If you’ve got a Canon camera, you probably heard about Magic Lantern. As a timelapser, you can use a lot of great features in this software. In this case, you should check out the ghost image function. In the live view mode, it displays semi-transparent picture of your choice.
You can choose the first picture of your shot and position your camera to match your object placement. It’s quite an accurate, but slow method. It took some time to get perfect match when another image covers up your live view. Make sure you’ve got unchecked ‘auto update’ in menu using this feature. To be more precise, you always want to reference to the first picture, not to the last taken picture.
I suggest using a tripod in this technique. It’s really hard to place your subject in the correct place when you’re shooting handheld, especially with long focal length. It’s possible, but why make your shooting harder? 😉
Finally, use your intervalometer or just take pictures manually. I like to release the shutter manually if there aren’t any slow moving objects in the frame, like clouds for example. Different intervals could make the clouds movement jerky, however, the clouds probably will be visible only in the wide angle frames. So I’ll leave this decision to you 🙂
If you’re shooting a small object in the location with a lot of people, intervalometer could also ruin the shot. Especially when you’re far away from the object, on telephoto focal length, your theme could be covered by people, cars etc. This can screw up the stabilization, but also the effect will be less visible.
In my Cracow hyper-dolly-zoom shot, while shooting with long focal length, I’ve got to wait for a few minutes to get the clear view for one picture, because of the walking tourist. The whole few seconds shot took me 1,5 hours.
Now let’s discuss how to shoot a push-pull effect using a hyperlapse. In the previous paragraphs, I gave you a lot of information so this part wouldn’t be as long as you expect. If you found your perfect location, composed your shot, it’s time to shoot. Just take the first picture, move your tripod a little bit (I hope you’re using a tripod), change your focal length: zoom out when moving forward or zoom in when moving backward, and take another picture. Repeat until you run out of your zoom range.
It’s not necessary to rotate the zoom ring every time you move forward. When I’m using 55-200mm lens, at longer focal lengths it’s almost impossible to see any change in the picture when I move 20cm/8”. Even if you’re zooming only every 3-5 shots you make, Stabilizer will handle it 🙂 I always try to zoom as often as possible to get the best result.
Another tip in this paragraph would be to change the distance that you move your tripod depending on the focal length. Long focal length requires longer distances and short focal length (wide angle) requires shorter distances.
You can of course move equal distances every time. But the same distance, for example 10cm/4” on 200mm wouldn’t be noticeable at all, while on 14mm this movement would be quite big. I always start with longer distances using telephoto, then shorten them while getting close to wide angle.
If you know hyperlapse workflow, you know what to do here. Stabilize your footage in software of your choice, for example using After Effects Warp Stabilizer. Additional step could be adding digital zoom, if you shot your footage this way. I described that process in the prime lens section.
If you were using more than one lens, check out the transition between them. Sometimes it’s necessary to delete some frames. Unfortunately, 50mm doesn’t look the same on every lens. There is some tolerance in the optics, that lets producers call 50mm something that actually could be a little different focal length. Of course, 50mm is only an example, it could happen with any focal length.
If you want to read more about hyperlapse post production, check out my previous Hyperlapse Tutorial, where I described it in details.
Dolly zoom with a prime lens?
There is a technique to accomplish this effect with a prime lens (or zoom with fixed focal length), but it has limitations. Grab your camera and record (or make hyperlapse of course) the footage as you are getting closer to your object, like a classic dolly shot. You can also use your slider. Then open your footage into your editing program and increase the size in the part, in which you’re further from your subject. If you get the same size of your object through the whole clip, keyframe the size and you get pretty nice push-pull effect as a result. Of course you’re losing some resolution, but maybe for someone with 50MPix Canon 5Ds it’ll be the best solution 😉 You can also try that with 4K video footage.
How to improve the effect
I’m not a fan of the prime lens solution from the previous paragraph, but I have a reason to mention that. You simply can combine these two techniques to get even better result. It’s super easy and gives you a lot of flexibility in post-production. If you’re starting at the long end of your lens, find the point where you want to start zooming out your lens and move a few steps backwards. Then start your shot by simple forward move, as it would be classic hyperlapse. But don’t forget to start turning zoom ring in due course 😉
When you don’t have to worry about zooming, you can make really small distances between pictures and make it bigger and bigger, so it’s good method for a ramp start. A few images at the end of your footage will give you opportunity to start your clip smooth in post by just resizing them. We’re talking about hyperlapse, so you’re not losing much quality, because the image is much bigger than your final video, even if you’re exporting as 4K.
It’s also possible for stopping at wide angle, but you should decrease the size of the clip in post. Only way to do this is to leave increased size of the whole clip, than decrease it at the end. It also allows you to maximize the vertigo effect by combining digital post production zoom with your lens zoom 🙂
We got to the end of this tutorial. I hope you feel motivated to try this technique 🙂 As always, I’d love to see your work. Let me know in the comments here or on my fanpage if you’ve tried dolly zoom. Also, please like my page on Facebook, it’s great when I see that you appreciate my work, that motivates me to write more tutorials 🙂 Thanks to all guys from fb groups that helped me correct the last article, you’re the best! 🙂
This tutorials has been moved to my new blog website, please follow that link: http://beyondthetime.net/how-to-shoot-hyperlapse-tutorial/
In this article I’ll explain and give you some tips on hyperlapse shooting. You don’t have to buy expensive gear or motion control equipment to accomplish great effects with hyperlapse. So this article could be really interesting for low budget filmmakers. Patient low budget filmmakers.
Tips from the author: you can always pause the gifs just by clicking/taping on them.
I assume, that you’re familiar with timelapse photography. I consider hyperlapse as a special case of timelapse, where we add specific type of movement.
There are a few ways of adding movement to your sequence. You can make a movable timelapse with a slider (even without motion control system). But what when you want longer movement than your for example 1,5m/5′ slider? Here comes the hyperlapse (you can’t just keep buying longer sliders 😉 ).
In the hyperlapse we add hyper-movement to the sequence. In different words – we move for much longer distance. The movement is accomplished without any tracks, just by moving your camera between shots. In this technique you replace gear like sliders etc with post production stabilization.
So is every movable timelapse a hyperlapse? You obviously can’t call ‘hyper’ something that moves only 60cm/2′ forward. That’s great technique, where you really can make ‘every frame a painting’, because you literally compose every frame separately.
Basically, you need a camera. That’s your minimum equipment. In addition you need some patience. Maybe lots of patience. Like I mentioned earlier, you compose every frame separately, that takes a lot of time comparing to the simple video clip. I love hyperlapses, because you don’t need fancy & expensive gear to make great looking shots.
What else can be useful?
My favourite gear for the hyperlapse is a tripod with any kind of a video head. It doesn’t have to be smooth, it’s just better to adjust pan and tilt separately. I think you can work with a ball head, but it would be a little nightmare, unless you’ve got perfectly even ground.
Some people also use monopods, I probably will also try that when I buy one, for example for bike travel purposes where I’ve got to choose between handheld and monopod shooting due to limited luggage.
The advantages of using a tripod for me:
More steady shots with tripod doesn’t mean that you can’t shot perfect steady hyperlapse handheld. It’s just difficult 😉 More about smoothness later in this text.
I also use an intervalometer, provided with Magic Lantern. Differences between shooting with and without intervalometer in the next section.
Like in a classic timelapse, I highly recommend to set your camera to manual mode. Not only exposure, also white balance and focus. There are a few ways to deal with changing white balance during shot (especially shoot RAW), but when your focus jumps between objects it could ruin a few hours of your work.
Of course you can use autofocus for your shots, which sometimes I also do, but it should be done purposely. When you are constantly changing your distance to the object you’re shooting, it’s comfortable to use autofocus. If you want to avoid mistakes, you can choose only one autofocus point, because you know what you want to be in focus, camera electronics doesn’t have to choose for you 😉
If you’re a Canon guy, you may be interested in Magic Lantern firmware. I’m going to write an article, about how ML helps me accomplish some tricky shots. Don’t miss it, follow me on Facebook .
Fixed interval or shooting with your hands
Intervalometer is not necessary to shoot hyperlapse. You can release the shutter manually. But is it better? It depends on your shot. For example, when there are clouds in your frame, you probably want them to move smooth, not jump because of different intervals.
When you’re shooting only cars or people, they probably will get on 1-2 frames, because of the long interval between shots (when you make one photo of some car on the road, the car will probably leave your frame by the time you’ll be moving tripod). So if you can’t shoot their smooth movement anyway, you can shoot with your hands. Also, with the long exposure or telephoto shots I always use intervalometer or 2 sec self-timer. Nobody wants to touch camera during exposure 😉
How long my interval should be?
Making a hyperlapse can get you use longer intervals than you would probably choose for timelapse. You’ve got to have time to move your camera (and maybe tripod/monopod), then adjust position of your camera to match the last shot as closely as possible.
I use intervals between 5-12 seconds. It depends on the length of the movement, amount of bumps on the road, sometimes distance from your reference point in the hyperlapse (more about this point in the next section) and also your experience. In most of my hyperlapses I was shooting about 8-10 seconds intervals. It’s a good time for clouds movement (when they are moving with an average speed, not fast as hell 😉 ).
RAW or JPG?
That’s eternal struggle for some people on the internet. For me it’s not. I shot my first hyperlapses as JPGs. It’s very nice to have 4 times smaller files. But when I dig into RAW files I was impressed how much mistakes I can fix in post production 😉 Also, it provides better overall quality of your films.
When I’m shooting bigger project, where I know I’ll spend a few weeks or months to finish it, I always shoot RAW. When I spend so much time on something I want it to be as good as possible. Sometimes I shoot shorter timelapses in JPG, especially on better cameras (there is quite big difference in quality between 550D with kit lens and Sony A6000 with Carl Zeiss 24 f/1.8). RAW format is forgiving many mistakes, like incorrect white balance or exposure.
There are a few types of movement here. You can move straight (example in the gif below), parallel tracking an object or around the object. I would say, that the second type is the easiest. When you’re moving straight it’s hard to find one point to aim the camera, and when you’re shooting around the object it’s hard to keep fix distance from the object (unless there are some perfectly circle lines on the ground).
Most of my hyperlapses looks like this: I set my camera on a tripod, frame my shot finding a fixed point, plan the path and shoot. I think the best location is where you can find some lines on the ground. For example connection of the paving tiles, curb or any other available line. You can even draw the line by yourself using chalk if you’ve got enough time.
If it’s possible, I try to place two legs of my tripod on the line. It’s simple geometry, when a line passes through two points, third point will be on line parallel to the first one. That means third point will have fixed position too. Following line with only one leg can provides perspective changes which are hard to stabilize. After first shot I move a little following my path, reframe to find my fixed point and shoot again (or let intervalometer do that for me).
Reference point – how to find it?
First of all, you don’t need a fixed point to shoot a hyperlapse. There are conditions, where you can’t find any point, for example long shot (some road maybe) when you’re moving straight. It’s hard to shoot perfectly stable shot like this. If it’s possible, find a fixed point.
Fixed point is a point in the frame, where you aim your camera making each photo. It could be anything, a lamp, a window, a sign or a weird stain on a building. You need to be sure you won’t mislead this point with something similar nearby. When you choose this point, you’ve got to choose how you aim your camera.
When I shoot with a DSLR using optical viewfinder I like to aim one of the autofocus points to hit the reference point. It’s not the best solution, because you’re limited by amount of that points. On the other hand, in a DSLR the battery life is much longer using this method and strong sun wouldn’t disturb you composing your shot. In liveview mode you can use focus box, some kind of grid (for example rule of thirds, which even your phone should have) or Magic Lantern cropmarks.
I’ve also heard that some people just take a transparent tape, place it on the screen and make a dot with a marker. If you haven’t got any other ideas, you can try that 🙂
In my GH4 I found out that I can set my own ‘grid’, consisting of 2 lines and I think this is a great solution for a hyperlapse.
As you can see in the gif above, the reference point doesn’t have to be in the centre of the frame. You can choose whatever point you want. However, it’s easiest for stabilizer if the point is around middle of the frame.
Take a look at the upper right corner of that gif. There is some joggling in this shot. When the reference point is placed at the side of the frame, you should be incredibly accurate, because every single miss at the reference point causes the big change of perspective on the other side of the frame. In this shot, it’s caused by long distance from fixed point to the end of this building & the angle of camera in relation to the building.
How long should I move between shots?
Basically it’s up to you. I’m always trying to plan how long I want each sequence to be & how fast the end result should be. Then I can calculate how fast I’ve got to move. With smaller movement you should get more stable result. To let you imagine it better: first GIF in this tutorial was shot moving about 15cm / 0,5′ between each frame.
Many times, when I don’t have any lines or points on the ground for reference, I move my tripod for length of my foot, sometimes one, sometimes even a few steps. The second method is definitely less accurate, but it provides fast movement in the finished movie.
You don’t have to move equall distance every time. You can get it shorter or longer. That’s creating speed ramping effect. I suggest, instead of increasing distance from 10cm/4″ to 1m/3′ immediately, doing short increments that are actually ramping, for example such distances:
You can get best results with small changes between each frame (but avoid repeating distances during ramping).
Final tip in this section: always think about your stabilization. For example, you should avoid any lens flares, cause after stabilization, they would jump through the screen like crazy.
That’s where you make your footage look fabulous. The main thing we do here is adding digital stabilization.
If you’re shooting RAW, first you should process your photos in your software of choice. Now, when you import your footage to editing software, it’ll probably look like crap. Don’t worry, most of the times it will be shaky as hell. But even if you did a great job shooting and your footage is kind of stable, you should digitally stabilize it. More stable input footage gives better end result.
Our work flow will be: import footage to sequence that’s set to resolution matching input pictures, for example 4608×3072 for GH4 16Mpix 3:2 photographs. You can slightly down scale the image before stabilizing if your computer isn’t fast enough to stabilize in reasonable amount of time, but always keep image bigger than your output movie resolution.
If you try to stabilize 1080p footage and export as 1080p film, you’ll loose some resolution and quality, because stabilizer increases size to do it’s job (you can use ‘synthesize edges’ option in warp stabilizer, but when your input footage is even 12Mpix picture, it’s pointless).
I recommend using Adobe Warp Stabilizer, available in After Effects and Premiere CC. The only difference between these two programs that I use is ‘Show track points’ checkbox, which I’ll explain later.
Now, there are a few other options to work with inside stabilizer. Firstly, it’s not always better to use ‘Detailed analisis’ (Advanced tab). Sometimes it helps, sometimes it makes footage look awfull. I always start with unchecked detailed analisis, and if the result is not satisfying then I check it. It could save you some time, cause if you pay attention to my tips in movement section, simple analisis should be enough.
In most hyperlapses you should stabilize position, rotation & perspective (sometimes also scale, but in most cases I’ve got checked ‘Preserve scale’), so I suggest to stay with default method: Subspace warp.
Smoothness option also depends on your footage. Stabilizing takes less time than analisis so you can try different values. With hyperlapse you can go further than with movie clip, due to lack of motion blur (you can perfectly stabilize your movie clip but you can’t get rid of motion blur recorded on it – that’s not always bad, Film Riot did a video about creative use of warp stabilizer with motion blur, you can watch it here). I have never go with smoothness above 100%. I don’t think it’ll be ever necessary for me.
In the picture above you can see a clip with track points showed. It allows you to see points, which program found to stabilize & delete some of them. I use that option, when manipulating with other values doesn’t give me satisfying result. It’s a powerfull option, because After effects doesn’t always choose the same points we want to stabilize.
If there is something that moves in the frame (a car, a train, clouds etc.), it shouldn’t be chosen to be stabilized, because it can ruin all your stabilization. In the picture above you can see some points that could be deleted. First of all, three points on the stars. As we all know, stars are moving, so it could mislead stabilizer (unless you’re doing shot that tracks stars). Second, points on trees. It was a little windy out there and they are swaying between frames.
To be clear – it’s not necessary to delete this point, esspecially when the stabilization result is good. Algorythm always finds lots of points, but it ignores some of them. It’s a problem only when it ‘thinks’ it should stabilize wrong points.
If you read all of this tutorial you deserve congratulations! I hope you learn something 🙂 I’d love to see your work, you can send me links on my facebook page. Also, I really appreciate every comment. That’s my first tutorial on this page, so your feedback is highly desirable.
So now go out and shoot! Experience is gold 😉
If you want to learn more, check out my tutorial about flow motion video on Time Lapse Network.
Welcome to freshly launched TL video website. We’re passionate about timelapse, hyperlapse & stop motion films. Soon you’ll find here some tutorials about this techniques & more! Stay tuned 🙂