The picture is sharpening for how higher frame rates will manifest in cinemas. But it’s not the prettiest scenario for exhibitors. As often happens when new technologies are introduced, early adopters will be rewarded with early opportunity, and the possibility of early obsolescence.
It’s long been known that motion images shown at 60fps are visually more compelling. Doug Trumbull found higher frame rates so compelling he patented the range of 50-72 fps as a means to create a “heightened physiological response in observers.” Doug’s format, Showscan, was focused on the presentation of 60fps content. John Rupkalvis, a well-known and longtime stereographer, has been known to say that the improvement in visual quality from 24fps to 30fps isn’t repeated again until increasing the frame rate to 60fps.
Given these observations, one might wonder why Peter Jackson is shooting Hobbit at 48fps? Peter Jackson is a director that likes to push technology to the edge, and has long been a proponent of 3-D. It is likely that Hobbit would have been shot at 60fps per eye if the technology were there to play it back in cinemas. But the bottleneck that exists today is that media block design decisions restrict compression bit rates, in practical terms, to no more than 450Mb/s. This is the data rate that the compressed image file in the DCP requires. Peter Jackson’s technical crew determined that 48fps per eye 3-D material compressed at a combined rate not to exceed 450Mb/s is acceptable for the release of their movie, but that 60fps per eye at 450Mb/s is not. Jim Cameron is holding out that 60fps per eye will be possible for the release of his next Avatar. Exhibitors need to note: even though your projector company may tell you that their product can play 60 fps per eye, it’s the media block that will limit the content that you can play.
If this were film, the likelihood of lots of movie releases at high frame rates would be small. High film duplication costs would limit this format to all but the most promising of releases. That was a major limiting factor of the Showscan format 20 years ago. But things have changed. High quality, relatively low cost digital cameras capable of capturing images at cinematographic quality are now available. Doug Trumbull recently confirmed that today’s technology allows anyone to shoot at frame rates much higher than 60fps per eye. If today’s audience perceives the value in higher frame rates that Showscan viewers experienced, then higher ticket prices should follow, which will also cause the production of more higher frame rate content. The cost of this to the exhibitor would be a more capable in-projector media block, more storage, and more time to load the movie files.
The determination of parameters for the more-capable higher frame rate media block is a role of the SMPTE Higher Frame Rate Study Group. Co-chaired by Michael Karagosian, a major task of the group is to determine the compression bit rate needed for high quality 60fps per eye exhibition. The community has responded well. Doug Trumbull is volunteering high frame rate test material that he will shoot. Pace-Cameron has opened up its studio for the shooting of custom high frame rate test material by cinematographer and HFR Study Group co-chair David Stump. Tests will be run later this year with compression bit rates of up to 1Gb/s to determine the rate at which the industry should standardize.
The messy part is that cinemas with more capable in-projector media blocks will be few, driving the cost of upgrade higher than may be expected. It’s important to note that the improved features can NOT be accomplished with software upgrades. Higher bit rate compression decoding requires special hardware. The result will be three tiers of cinema in terms of frame rates:
- “DCI” auditoriums capable of 24fps and stereo 24fps (48fps),
- Level 2 cinemas capable of up to stereo 48fps (96fps), and
- Level 3 cinemas capable of up to stereo 60fps (120fps).
Needless to say, just as this poses challenges on exhibitors, multi-frame rate distribution will be a major challenge for distribution. This is an area we’ll explore in a future article.