How is this effect done?

[Adam Haro]

I've noticed a real cool effect in a couple movies lately, My Bloody Valentine and Chronicles of Narnia 2 for example. The camera appears to be zooming in on the subject but the background looks like its zooming out. Thats the best I can describe it. Hope it makes sense.
Anyone know how its done?

[Adam Haro]

I think I found it, would this be a dolly zoom?

[Jacques E. Bouchard]

Originally Posted by Adam Haro

I think I found it, would this be a dolly zoom?

Yep. Old effect. The camera appears to remain stationary in relation to the subject while the background gets "longer". It's a great demonstration of how your zoom affects perspective.


J.

[Matthew Overstreet]

Yeah, I think this effect was pioneered by Hitchcock, and made famous by Spielberg in Jaws.

[Robert Martens]

Also called a "zolly", from what I've heard, if you like cute names for this stuff.

If I may take a moment, though:

Originally Posted by Jacques E. Bouchard

It's a great demonstration of how your zoom affects perspective.

I would say it's actually a rather poor example, since zoom does not, in fact, affect perspective. Dolly zooms don't make such a clear distinction between changes of focal length and camera position, thereby blurring the line and making the whole thing harder to understand.

You need some sort of a lens to get any perspective at all, naturally, but beyond that the geometry of the image is determined by camera-to-subject distance. Focal length magnifies or shrinks the picture, changing its size, but doesn't alter the perspective.

The best way I've been able to come up with to demonstrate the difference is the fact that most of us, long before we had any interest in filmmaking, could distinguish between a zoom and a dolly move when we watched movies and TV shows. Maybe you couldn't figure out exactly why, but something told your brain they weren't the same thing. The giveaway is the lack of perspective shift in the zoom. Try it out. Zoom in on something, and note the lack of any parallax while you do so. Assuming the camera stays still, the objects in the scene don't slide in relation to one another. The whole image enlarges or shrinks, but the change is uniform. Ratios of distance remain the same.

Everyone says that to compress the geometry of a shot, you need to use a long lens, but while this is appropriate shorthand in real world shooting situations, it flip-flops the truth. The belief is that long focal lengths compress the view, but the subjects increase in size and you need to move the camera backward to compensate. The opposite is what's actually happening, however; compression is accomplished by moving the camera farther from the subject, but doing so shrinks them relative to the bounds of the frame, and longer lenses are needed to magnify them as necessary to create the composition you're after.

Ron Brinkmann has an excellent explanation in his book, "The Art and Science of Digital Compositing", if you'd like a better description of the effect, and comes complete with photo examples.

[Jacques E. Bouchard]

Originally Posted by Robert Martens

Zoom in on something, and note the lack of any parallax while you do so. Assuming the camera stays still, the objects in the scene don't slide in relation to one another. The whole image enlarges or shrinks, but the change is uniform. Ratios of distance remain the same.

That's not what I meant when I said that the zoom (not zooming in) changes the perspective.

What I mean is that in two shots composited the same by moving the camera in relation to the subject (with a short and a long focal length), the long focal length (zoomed in) will compress your background.


J.

[Robert Martens]

Originally Posted by Jacques E. Bouchard

... the zoom (not zooming in) changes the perspective.

Forgive me for being confused, but what is the difference between "the zoom" and "zooming in"? I suggested zooming in on a subject only as a way of demonstrating the optical principle I'm trying to describe; since a zoom is just a lens with a variable focal length, you would be able to observe any supposed compression taking place as you changed it. Since no such effect is visible, I thought it a good demonstration of my point, which is simply that focal length cannot affect perspective.

In two shots composed the same way, one with a short focal length and the other long, the longer shot will exhibit a compressed perspective only because the camera's been moved away from the subject, not because of the chosen focal length. In the case of the famed dolly zoom, the "zoom" half of the process only serves to maintain framing and subject size, where the "dolly" is what actually compresses (or stretches, depending on the direction you're moving) the scene.

[Aric Mannion]

Adam, you can either pull forward while zooming out, or back up while zooming in. A good example is in Martin Scorces' Goodfellas, during a conversation in the coffee shop.

[Shaun Roemich]

Originally Posted by Robert Martens

In two shots composed the same way, one with a short focal length and the other long, the longer shot will exhibit a compressed perspective only because the camera's been moved away from the subject, not because of the chosen focal length. In the case of the famed dolly zoom, the "zoom" half of the process only serves to maintain framing and subject size, where the "dolly" is what actually compresses (or stretches, depending on the direction you're moving) the scene.

Incorrect. A long focal length lens will magnify the entire shot: foreground, midground and background and will optically compress the shot.

[Robert Martens]

Shaun, I must respectfully disagree; as you say, longer lenses will enlarge the entire shot, foreground, midground, and background, but I'm trying to make a clear distinction between mere magnification and genuine compression. You could describe the cropped image that results from magnification as claustrophobic, closed in, and perhaps "compressed", but I'm not referring to a mood, feeling, or atmosphere, I'm talking about the definite, observable, measurable geometric distortion that is the result of a change in perspective. Since perspective is a function of distance, not focal length, simply using long lenses will not itself provide said perspective shift.

As I described above, the difference between a zoom in and a dolly in is the complete lack of change in perspective during a zoom. If alterations in focal length do not change an image's perspective, how can they possibly be held responsible for an effect that is by its very definition a perspective change?

[Shaun Roemich]

Sorry, again, wrong. If you have a person in the foreground and a clock in the background:
With a wide angle lens position (short focal length), frame the person to fill one half of the screen with the clock over his shoulder. Normal size person, tiny clock.

With an extreme telephoto lens position (long focal length), move the camera WAY back and get the exact same image of the person filling one half of the frame with the clock again over his shoulder, what do you get? BIG clock.

The effect the OP was describing is performed thusly:
- The camera in a wide angle lens position starts close to the subject on a dolly.
- The camera is moved (dollied) backward while the camera operator zooms in to maintain the same size of subject (this takes a certain amount of skill)
- The effect of this is that the subject appears to stay the same size and space while the background "comes forward" to meet him.

You can reverse the above steps to have the background "move away" from the subject.

The best photographic explanation I've seen of this is in Herbert Zettl's book "Television Production Handbook": in the book, Mr. Zettl shows a Greek Columned building along the left side of frame in a wide angle and then telephoto lens position. In the wide angle lens position, you can see "between" the columns. In the telephoto lens position, the distance between the columns becomes "compressed" so that the ability to see between them is completely obstructed.

Lens compression is a result of focal length of lens. The "magic" in the OP's noted effect is in fact lens focal length COUPLED with a well timed dolly. Or vice versa.

[Robert Martens]

Originally Posted by Shaun Roemich

Sorry, again, wrong. If you have a person in the foreground and a clock in the background:
With a wide angle lens position (short focal length), frame the person to fill one half of the screen with the clock over his shoulder. Normal size person, tiny clock.

With an extreme telephoto lens position (long focal length), move the camera WAY back and get the exact same image of the person filling one half of the frame with the clock again over his shoulder, what do you get? BIG clock.

The effect the OP was describing is performed thusly:
- The camera in a wide angle lens position starts close to the subject on a dolly.
- The camera is moved (dollied) backward while the camera operator zooms in to maintain the same size of subject (this takes a certain amount of skill)
- The effect of this is that the subject appears to stay the same size and space while the background "comes forward" to meet him.

You can reverse the above steps to have the background "move away" from the subject.

The best photographic explanation I've seen of this is in Herbert Zettl's book "Television Production Handbook": in the book, Mr. Zettl shows a Greek Columned building along the left side of frame in a wide angle and then telephoto lens position. In the wide angle lens position, you can see "between" the columns. In the telephoto lens position, the distance between the columns becomes "compressed" so that the ability to see between them is completely obstructed.

Lens compression is a result of focal length of lens. The "magic" in the OP's noted effect is in fact lens focal length COUPLED with a well timed dolly. Or vice versa.

I've added bold tags to your post, around the important statements. In every single example of this optical compression effect, what do you do? You use a long focal length, yes. What else do you always, always, always do? You move the camera farther away. You cannot achieve this effect unless you move the camera, and that is the distinction Mr. Brinkmann makes in his book. The focal length magnifies, the distance increase compresses.

I dismissed it as frankly stupid the first time I read it, since I'd always had it drilled into my head that if you want that squashed geometry you have to use long lenses, and after all what does some visual effects guy know about optics? They're not trained in that field, and I didn't give his comments any heed until I read the book a second time. The logic is all there, as are the photo examples.

I remain convinced that it's the camera-to-subject distance that produces the effect, based on his arguments and evidence, but I'm apparently not as good a writer as he and can't make my points as clearly. I stick by my posts so far, but if my insistence is bothersome I apologize, and I'll drop the subject.

[Jacques E. Bouchard]

Originally Posted by Robert Martens

I've added bold tags to your post, around the important statements. In every single example of this optical compression effect, what do you do? You use a long focal length, yes. What else do you always, always, always do? You move the camera farther away. You cannot achieve this effect unless you move the camera, and that is the distinction Mr. Brinkmann makes in his book. The focal length magnifies, the distance increase compresses.

So you're saying that, in Shaun's photo offered as example, you could see the space between the columns compressed even when zoomed out? That's interesting. I'd like to see an eample of that.


J.

[Robert Martens]

Yes, Jacques, that's exactly what I'm saying, and I only now realize I neglected to offer that exercise in my explanations. That's one of the photo sets presented in the Digital Compositing book, as a matter of fact.

With the camera locked down, in one location, shoot both a wide angle and telephoto shot of a subject, something that will adequately exhibit compression (the columns mentioned would be perfect). When you compare these two in a computer, the initial impression may be that the telephoto shot is the only one that's compressed, but that's not the case. Crop the wideangle shot to match the telephoto one, blow it up to the same size, and the perspective will match.

I have no samples on hand, but if you give me a few minutes I'll grab a few shots and upload them. Hold tight, I'll be right back.

UPDATE

Here you go! Please excuse the dirt, I haven't vacuumed this room recently. Give me a break, I'm a guy.

I arranged three bottles of rubbing alcohol I had handy in a staggered line in front of a cardboard box, with a bit of space between each of them. Setting my camera on the floor, I took one photo at full wide, and another at full telephoto (for curious parties, on my Cybershot T500 the 35mm equivalent angles come out to 33 and 165 millimeters, respectively). The camera did not move.

Comparing the originals may make the telephoto shot seem compressed at first, but look at the third file I uploaded; this is the first photo, the wide angle, simply cropped in Paint Shop Pro to match the framing of the telephoto picture as closely as I could, then resized to match the other two images' dimensions. Though the resolution differs, you can see how even though the focal length changed, the perspective did not, and the perceived distance between the objects is the same in both photos.

[Alan Certeza]

So, I'm a student at the Art Institute of Seattle and I tried the zolly and shot and realize that it's not that simple as it's describe.

manually zooming and tilting while trying to pace to the dolly is very hard, or for me as I'm still learning about videography.

I just thought I should share my experience with this shoot.

:)