How to film and edit dance to convey spatial continuity
Why do dance-centric films and television vary so widely in their ability to engage viewers? What principles can we follow to better convey the energy, beauty and athleticism of live dance, even though audiences will view the film in conditions that are highly detached from the live experience?
This topic merits a long book, rather than a blog post, and I’m purposefully ignoring many aspects of dance on screen, including the arts of choreographing for the screen, creating or choosing a set that works well for filming dance, staging dancers within that set, costuming dance for film, and varying techniques of lighting dance for film. For today, my emphasis will stay on two aspects: the camera and editing.
I will add to this post over time, so if you have any questions or suggestions, please leave a comment below.
1. Lens Choice
I have read a lot of discussion in the academic community about the supposed difficulty of translating a three-dimensional art form (dance) into a two-dimensional space (the screen). While those challenges would in theory apply to any performance art displayed on the screen, I suspect that this discussion stems from a generally-perceived flattening of 3D space and loss of location reckoning when audiences see dance on screen.
“Matched” (presented by The LXD, directed by Charles Oliver, DP Alice Brooks), was originally filmed and presented in 3D. Is its impact lessened in 2D?
Some recent efforts in 3D dance filmmaking by directors Jon Chu (“Step Up 3D”), Charles Oliver (“Matched,” “Santa Claus is Coming to Town”) and Ryan Landels (“Game On”), as well as their collaborating Directors of Photography, Ken Seng and Alice Brooks, have shown that 3D can work excellently with dance. While 3D can enhance certain aspects of dance, I don’t believe that shooting in 3D solves all of the stage-to-screen “flattening” problems in and of itself. 2D and 3D screens are only a facade that covers the core issue of why dance can appear flat on screen. The main problem (I believe) lies in the fact that dance is often filmed with telephoto lenses rather than wide-angle lenses. That may seem like a bizarrely-specific and totally-myopic point to focus on, but let me explain the reasoning behind it.
Telephoto lenses allow the camera operator (and the audience) to see up-close, intimate and fascinating details that a live audience wouldn’t easily experience. The expression, passion, and beauty of dance can be conveyed in a very intimate way with a long lens. Despite that level of intimacy, however, we regularly see two main problems when dance is filmed on a long lens:
- The audience only sees close-up details and isn’t able to see the larger picture (whether it’s the whole body of the dancer or the collective body of dancers in formation). As a result, the audience feels disoriented and abandons their mental efforts to perceive the movements as sequential and connected. (This is partially a conditioned response created by music video-style editing that sequences brief unrelated movements in an effort to create an atmosphere of excitement.)
- Telephoto lenses compress distance on the Z-axis (between the camera and the horizon). They optically manipulate the image so it appears that both far and near subjects are closer together than they are.
The telephoto lens creates the net effect of killing all Z-axis movement dynamics in the dancer. When a ballerina beautifully reaches her hand towards the camera, that motion is largely lost and compressed into Z-space (even though some hint of that may be conveyed through defocus due to a shallow depth of field, or through 3D parallax). If that same movement is filmed with a normal lens (35mm) positioned closer to the dancer, the Z-axis movement is not lost. If a wide-angle lens is used, the movement is exaggerated.
Note in this clip of a dancer filmed with a 10mm lens placed inches from his feet at their nearest arc, how the leg lines are powerfully conveyed:
In my ideal world, dance would largely be filmed with wide (or normal) lenses, and telephoto lenses would be used sparingly to bring the audience in for moments involving facial expressions, details of small movements, and other specialty shots.
Watch any Gene Kelly or Fred Astaire film and you will see that wide shots with long, long takes were the norm. This technique works in those films. While audience sensibilities are clearly different now (nearly 80 years later), the over-emphasis on close-ups during dance sequences creates an image that both disorients the audience and conveys movement that feels flat.
2. Camera Movement
I am convinced that there is no single perfect way to rig a camera for dance. Carlos Saura has done amazing, passionate, gripping work with Flamenco dance using static cameras.
Contemporary TV producers (on programs like “So You Think You Can Dance,” “America’s Best Dance Crew” and “Dancing With The Stars”) make heavy use of crane and steadicam movement when filming dance, and the effect is very engaging.
In my mind, the most important principle when filming dance is to allow the dance itself to be the emphasis. We might ask ourselves, WWFAD? (What Would Fred Astaire Do?)
“Either the camera will dance, or I will.”
(Fred Astaire in Winge, John. “How Astaire Works.” Film and Theatre Today, January 1950, pp.7-9.)
That Fred Astaire quote is loaded with meaning, personality, context and history, but I think that it in essence it conveys a simple principle: the emphasis in dance film should be on dance.
Astaire was very particular about the way directors worked with him. He was making films during the same era that Busby Berkeley was rigging spinning cameras to the ceiling and filming kaleidoscopic dance sequences that dizzied audiences with pure visual spectacle. My best guess is that Astaire (and others) were taking reactionary stances to Berkeley’s work, collectively choosing to emphasize spectacle created by dance over spectacle created by cameras, art departments and wardrobe.
When Alice Brooks back-lit dancers and filmed them largely with a steadicam in “The LXD,” the look was far from a Fred & Ginger dance sequence, but they were similar in that the emphasis always stayed on the dance itself. Brooks‘ decisions about camera movement never feel like they were created to add false energy to a scene; her decisions seem to be based around the primary goal of presenting dance itself, with the secondary goal of presenting it in a beautiful and engaging way. I believe that she follows the contemporary manifestation of Astaire’s mandate: she never attempts to place her talent with the camera above the talent of the dancers on screen.
Start watching at about 8:53 to see a good example of Brooks’ love of strong back-lighting, and how well it plays on screen:
The most important principle to remember is that editing dance is still editing. Some editors mistakenly conclude that dance sequences don’t follow the same principles of good storytelling that other sequences do, and their work suffers as a result. You must master the fundamentals if you hope to make something great. If you want to edit well, first learn and master the rules before you attempt to break them.
Visual continuity is still very important in dance film, otherwise the movement disorients the viewer. Based on conversations with other editors, it seems that some are unable to follow the storyline contained in the movement of a dance piece; they erroneously conclude that because they only see the dance as a series of unrelated movements, that’s how audience members will see it. The editor’s lack of understanding of dance will pass on to the audience, and vice-versa. An editor must comprehend and clearly portray the movement if she expects audience members to do the same. If the edit is made without a sense of conveying the storytelling inherent in the dance movement, the audience will mentally struggle to assemble a narrative based on visual fragments, and will often feel dissatisfied.
For example, eye lines are important to maintain in dialogue scenes, and they are important to maintain in dance. If one edit creates a jump between the dancer looking to frame right at the end of a shot to a shot where he is looking to frame left, it will be jarring to the audience, even if only on a subconscious level. When the line needs to be crossed, movements that spin are a convenient time to place an edit point, since the eye line is effectively reset.
I also encourage dance editors to resist the urge to cut too frequently. The most common complaint I hear from people when they come out of dance movies (as well as many contemporary Michael Bay-style action movies) is that the dance sequences were disorienting and hard to follow. Fast cutting can help to mask poor dance technique, and make boring movement look more exciting, but it does so by creating a slight of hand that tricks the audience, leaving them confused and ultimately less satisfied than they otherwise would have been. The ideal situation is to have excellent dancers who are well-rehearsed (or free to improvise), and let them present their art.
An editor may feel that a 15-second wide shot of two dancers is far too long of a shot for audiences to be engaged in, but practical experience shows that long, uninterrupted shots can also create a hypnotic effect as the audience fully loses itself in the story of the dancers’ movement.
Skip to about 2:20 in the season 3 finale of The LXD and take note of how many long takes are included:
In other words, if you’re tempted to “add” to a scene with clever editing or “exciting” cutting, stay out of the way and let the dancers dance.
More To Come
In the future, I hope to discuss more dance film-related topics, including:
- Choreographing for the screen
- Creating or choosing a set that works well for filming dance
- Staging dancers within a film set
- Costuming dance for film
- Lighting dance for film
Please feel free to leave a comment if you agree, disagree or would like to discuss something that I neglected to mention.