Creative Frame Rate Presentation November 26, 2015

One of the more fascinating qualities of moving image technology is that it is at once intuitively comprehensible and yet endlessly, fascinatingly complex. No presentation at 2014’s The Reel Thing demonstrated this quite like Jonathan Erland’s remarkable “The Speed of Cinema.” Both a rigorous technical and historiographic examination of moving image exhibition, Erland’s presentation further managed to achieve that rarest of feats in academic discourse – it was damn entertaining to boot! Pinging freely between past and present, technical minutiae and the fundamental cinematic tenants, Erland’s presentation was a wild (and often quite humorous) joyride, embracing key ideas of the past to point a way toward exhibition’s future.
At a time of such fundamental technological change as the current industry-wide switch to digital distribution, a fair amount of hand wringing is to be expected among adherents to the old guard. One popular point of contention among members of the archival community – indeed, the one around which Erland framed his entire presentation – is a seemingly problematic limitation built into the standards of almost all current digital projection (DCP) systems: a fixing of projectors’ display frame rate to a 24 frames-per-second (fps) standard. This standardization presents a problem for the display of archival silent footage, much of which was produced at a variety of frame rates before the advent of sound established a standardized speed of display. Ironically, while many of today’s most technologically innovative directors (Jackson, Cameron, et al) are similarly pushing to free projection from its longstanding 24 fps yoke, it was Erland’s assertion that the key to developing a more flexible standard of cinema display lies not in new technology itself, but rather in its repurposing of principles as old as the film medium.
If, as Godard has famously asserted, “cinema is truth 24 frames-per-second,” it nevertheless remains that truth is a slippery thing to pin down. While, indeed, very nearly all sound film since the medium’s development has run at 24 fps (a speed necessary to achieve accurate audio fidelity), it is nevertheless true that 24 flashes of image per second is both an unnecessarily high and woefully inadequate speed for the effective reproduction of the perception of motion. The key to this paradox lies in a frequently ignored distinction of cinema technology: that between the number of frames displayed per second (fps) and the number of cycles per second (hertz, or Hz) at which an image is displayed. Thus, while frame rates as low as 16 or even 12 fps have been shown to produce a satisfactory illusion of motion, studies have nevertheless shown that in order to prevent a distracting strobing in the image caused by too long of intervals between the light of display and dark of the image being changed, an interval of at least 48 Hz is necessary. Thus, while sound film may indeed run nearly always at 24 frames per second, the fact is also true that it runs at 48 hertz per second, with each frame being flashed twice in succession before moving on to the next.
Such systems of multiple-flashing have been the case since nearly the dawn of cinema, with frame rates as low as 16 and 12 being repeated as necessary in order to bring them closer to the threshold of acceptable motion reproduction. Modern projectors, stated Erland, are capable of inherently displaying anywhere between 20 and 120 Hz at present, and there is furthermore nothing to suggest that even these numbers are fixed limitations. As such, the problem of frame rate becomes one of merely processing and specification limitations, for while most projectors are currently capable of a much wider range of display rates, they are limited by their attached processing hardware, as well as the specifications of the DCP protocol itself, which standardizes a display rate of 24 fps.
To prove his point, Erland made truly effective use of the Academy’s cutting-edge digital projection technology, which has been equipped with a special processing card capable of variable-Hz projection. Unleashing a bevy of clips shot at both myriad frame rates and shutter speeds (the camera equivalent of projection display cycles), Erland revealed the startling array of possibilities offered by this new form of exhibition, not just for cutting-edge digital features like Jackson’s The Hobbit (itself shot at a frame rate of 48 fps), but for archival material of all lineages and speeds. Climaxing with a breath-taking, native-30 fps projection of the “Kansas City” number from Oklahoma! (one of few notable sound film exceptions to be shot at a non-24 fps frame rate), Erland’s multimedia essay proved a sterling example of a far-too-rare occurrence: an academic engagement with technology that pointed a way toward reviving the past, rather than simply eulogizing it.
– John Kostka




SMPTE Hollywood Frame-Rate Presentation

Last Monday, I attended a fascinating presentation at the Linwood Dunn Theater, a private,
>286-seat screening room in the Pickford Center for Motion Picture Study in
Hollywood, CA. Hosted by the Hollywood Section of SMPTE (Society of Motion Picture
and Television Engineers), the event was entitled "Digital Projection—From 12 to 120
fps." It was basically a pre-recorded presentation by Jonathan Erland, long-time cinema

industry technologist and founder of the new Pickfair Institute for Cinematic Studies,
named after the famous mansion occupied by silent-film stars Mary Pickford and
Douglas Fairbanks, who were instrumental in promoting motion-picture art and science
in the early days of Hollywood.
The presentation included clips at many different frame rates compiled into a single DCP
(digital-cinema package) played from a Qube XP-I server on a Barco DP4K-32B 4K
projector with a Qube Xi 4K IMB (integrated media block). Qube and Fotokem were
instrumental in developing a system that can present various frame rates sequentially
and even simultaneously!
Erland's focus started with restoration and preservation of very early movies and how to
display them using digital projectors, especially DLP-based digital-cinema projectors,
which can play files at frames rates from 20 to 120 frames per second (fps). However,
some of the earliest movies—experiments, really—were shot at many different frame
rates, from 12 to 1600 fps. (Frame rates lower than 12 fps posed a danger of igniting the
slow-moving nitrate film with the hot arc lamps in early projectors.) Not only that, the
cameras and projectors were hand-cranked, and the frame rate was not constant; in
fact, variable frame rates were sometimes used as a visual effect.
In order to preserve the look of these films, Erland adapted a technique used in those
early days that is still employed today—double and triple flashing, in which each frame
is shown two or three times with black frames inserted between them. (The black frame
between one image frame and the next was necessary to mask the film as it advanced in
the camera and projector; that black frame is called the pulldown.) For example, The
Dancing Pig (1897) was shot at 16 fps, so Erland played it at 96 Hz—each frame was
displayed as three images interleaved with three black fields. Of course, half the
projector's light output is lost in this process, but high-output projectors can mitigate
this problem.
I had always thought that the current standard frame rate of 24 fps was established
because it was the slowest rate that would yield smooth motion with no jerkiness or
flicker, but it turns out that lower frame rates can look surprisingly smooth, especially
with double and triple flashing, thanks to persistence of vision and other psychovisual
effects. Instead, Erland said that 24 fps was chosen as the standard movie frame rate
when "talkies" supplanted silent films because it was the slowest speed that could
support the audio bandwidth required to produce acceptable sound quality when the
soundtrack was printed on the film.
Now, of course, 24 fps is the universal standard frame rate for movies—in fact, it has
become somewhat of a straightjacket, limiting the creative options that were available
to early filmmakers. But with the advent of digital cinematography and projection, that
limitation no longer applies, and the creative freedom of the early days of cinema can
be reclaimed.
In an effort reminiscent of the experiments from the late 19th and early 20th century,
Erland and his team shot some experimental footage using a Vision Research Phantom
high-speed digital camera at 240 fps—two women in a fast-moving, choreographed
martial-arts stick fight and another woman doing Tai Chi. The camera shutter angle was
360°—that is, fully open with no blanking interval. The resulting footage was then
shown with a variety of manipulations to simulate different frame rates as well as
different camera and projector shutter angles.
Those manipulations included what Erland calls "block integration," in which several
consecutive frames are combined into one frame—for example, to simulate 24 fps from
the 240 fps master file, 10 frames are combined into one. To simulate a shutter angle of
180° at 24 fps, five frames are combined and the next five are discarded.
As you would expect, the sticks in the fight sequence looked less blurred as the frame
rate increased, but even at 120 fps, there was still some motion blur. Decreasing the
shutter angle reduced motion blur but increased flicker. Of course, there was virtually
no motion blur at almost any frame rate in the slow-moving Tai Chi sequence.
One interesting technique Erland developed is called "rolling integration," which is much
like block integration except that the set of frames being combined shifts one frame at a
time. Returning to the previous example, the first frame at 24 fps is the result of
combining frames 1-10 in the original 240 fps file; the second 24 fps frame includes
original frames 2-11, and third frame includes original frames 3-12, and so on. By
varying the number of original frames combined into the final frames—and using black
frames to maintain the desired final frame rate—you can vary the "exposure" of the
image independent of the frame rate.
For example, with a 360° shutter angle, the exposure time of each frame at 24 fps is
1/24 second or almost 42 milliseconds, but using rolling integration with a 240 fps
original file, you could increase that exposure time to 1/8 second (125 ms) or more. Of
course, that would result in even more motion blur, which might be just what the
director wants in a particular scene.
The point is, digital technology now allows frame rates to be used artistically. As
discussed above, frame rate can now be separated from exposure time. Also, different
rates can be used in the same movie and even in the same shot—according to Erland,
James Cameron will use a variable frame rate in the upcoming Avatar sequels. In a
contemporary example, The Artist was actually shot at 22 fps and shown at 24 fps—an
increase of 8%—which made the dancers look lighter on their feet.
Early cinematographers had a lot of control over the minds of the audience, but much of
that control disappeared when 24 fps was locked in. Audio is no longer printed on the
film itself, pulldown blanking intervals are no longer needed, and digital cameras and
projectors can capture and display many different frame rates, so modern moviemakers
have much more creative freedom than they've had since the first days of cinema. All
that remains is for them to use it.
- Scott Wilkerson