Actually, one person asked me why film isn’t shot at 30 fps making the film to video transfer easier. The answer to that was easy: Compatibility with earlier film and 30 fps wouldn’t do European video any good at 50 fps anyway. The conversation then veered into how 24 fps was settled on in the first place.
I had some shrewd guesses. Various rates were experimented with back at the dawn of cinema and about 16 fps was a compromise of enough images for the illusion of motion without spending too much for film stock. Sixteen frames of 35-mm film is exactly a foot, but that may be due to them setting the vertical frame size so that sixteen of them came out to a foot.
The question was why the increase to 24 fps when sound came in. My guess was that the higher speed was necessary to get enough film passing under the sound head to stretch out the higher frequencies. At one foot per second a 3,000 Hz tone’s wavelength would be only .004
Googling the subject came up with lots of sites talking about the what but not the why. Can anyone out there confirm my guess?
Well changing from 24 fps to 30 fps in film would theoretically make it easier to transfer film to video, as NTSC video runs at 30 fps (or 29.997…). However, as a large part of the world uses PAL, which runs at 25 fps (not 50, unless it’s progressive of course), such a change would be meaningless.
As it is now, Film -> NTSC video transfer is slightly more troublesome than Film -> PAL video transfers, seeing as they have to do the 2:3 pulldown to transfer 24 frames smoothly into 30. However, when transferring to PAL, one simply speeds the film up to 25 fps; a change which is completely unnoticable (unless you comparing running times for NTSC and PAL DVDs of course )
Depends on what the f stands for. Interlaced PAL is 50 fields per second and NTSC is 60 fields per second (or 59.94, but let’s not get into drop frame time code).
I’ve also heard of converting PAL -> NTSC by just speeding things up also (I’ve never seen such a transfer, but some people don’t like the 4% increase in sound pitch). And from NTSC -> PAL I think it’s a 3:2 pulldown just like film.
To me, it’s completely unnoticable. Of course, I would imagine that, depending on the original sound pitch, the conversion process would also include a lowering of the pitch in the PAL version.
I don’t know about video transfers but the 24 frames per second is the slowest you can do it and have it work. Film stock was fairly expensive in the early days. Well, it still is. The average ‘print’ of a film today costs about $5,000.
Changing film today to be 30 frames per second would be a nightmare. The thousands of movie theatres would have to get new projectors. You can not just adjust the speed a bit. This is one of the problems with converting to digital movies. Every theatre in America would need new projectors.
Not to mention that all other countries use 24 FPS and either they would have to change, and all their theatres, or US theatres that wanted to show movies for other lands would have to have dual systems. And of course US movies that want to be show abroad well, you get the picture.
So, we are stuck with 24 frames PS because of tradition or momentum.
Some old theatres are equipped for 30fps—the process was ToddAO. If you EVER get a chance to see a film in ToddAO, do so. The higher frame rate delivers more information to the brain, and with a shorter time in the gate, more light can be used to project with, equaling a better picture.
The Showscan movie rides are 70mm at 48fps–a huge increace in information for the brain, with a corresponding increase in sharpness and clarity.
Origial Showscan setup was at 60fps—it looked like 3D without the glasses, and was so sharp they needed makeup in the actors ears.
One point is that 16 fps was hardly a universal standard in the silent era. Indeed arguments about the appropriate speed are a perennial dispute amongst both silent film fans and scholars.
One of the standard discussions of the issue is this article by Kevin Brownlow from Sight&Sound in 1980. On the question of why the standardisation to 24 fps with the advent of sound, he writes:
In other words, there was probably no speed-up at all.
There is a threshold below which the persistence of vision does not come together, and a flicker is too apparent. (Roughly) 16fps is a fair but not ideal threshold. As for Kevin Brownlow’s explanation, I agree. At the time it no doubt seemed to be the tail wagging the dog, to let the fledgeling sound people dictate anything having to do with camera. As a projectionist in film school, I had at my fingertips a variable shutter rate projector. Tasty item to have. We’d screen old silent movies, and I’d strike a good balance between the flicker and the motion of actors. You’d be surprised how good an old silent looks, once you take the time to try to find just the right projection rate. Of course, human movement was the best key to use in this situation.
The Douglas Trumbull’s Showscan system was indeed a higher frames per second film/project system. It also used 70mm negative with a 70mm release print.
I would vigorously argue, however, what The Vorlon said about having to put makeup in the actor’s ears. Urban Legend, I say. While you are indeed increasing the amount of information delivered to the eyes/brain by increasing A) size of negative, and B) Frames per second show to eye/brain, you are in NO WAY increasing the raw resolution of the film negative. That is to say, you are taking bigger pictures and more of them, but there is nothing inherently SHARPER about any of those individual images. If you pan at the wrong speed, you get strobing. If there is sufficient speed of movement, blur is seen on a given frame. ( The blur is resolved in the brain, which compensates for the blurred images by filling in the missing slightly blurred information, letting you see a moving picture that appears sharp. )
A more reasonable analogy in terms of sharpness is the jump from NTSC television to High Definition television. In this case, the literal resolution is being raised from 575 lines of resolution to roughly twice that. ( The standards are still being debated, but 1050 or 1170 is a close shot ). In that case, I can see details in a High Def shoot that are absent from capture in an standard NTSC shoot.
I’m sorry, but the whole 70mm/ 48fps ( or, 70fps) argument doesn’t hold. In order to increase the clarity of the images, you have to alter the grain structure of the film itself. This happened in the 1980’s with the advent of T-Grain motion picture and still film, which altered the physical grain structure of the silver halide crystals on the surface of the film from spheres ( s.i.c. ) to t-shaped flat objects. Clarity took a huge leap upwards, but that had nothing to do with Showscan.
Sidenote: Showscan cameras and projectors were made by Cinema Products Corporation, the company that built the Steadicam system until a few years ago. Cutting edge guys, those folks at C.P.
One other tidbit. ShowScan, Super Panavision 70, Arri 70 and such make use of so-called 70mm film. The film runs vertically through the gate. VistaVision and Imax cameras make use of 70mm negative, but run the film HORIZONTALL through the gate. If you think about it for a moment, you will see how IMAX gets away with a 10 story high image. The negative is 70mm tall and much, much more than that wide. Per exposed frame of film. That’s a lotta information for your brain. Still and all, you don’t have to make-up someone’s ears…
Have you ever seen 60fps Showscan? Not the 48fps they have today, but the original 60fps, interlocked 35mm magnetic sound version, normaly shown on a deeply curved screen with a Century JJ set to ToddAO speeds and a double-pinned intermitant?
Quite impressive.
The makeup-in-ears came from Trumbull, as I recall from a interview he did on the interesting problems they came across while pioneering the new format.
I saw it in the early 80’s at one of the Showbiz pizza parlors that was set up with the theatres, located outside of Washington, DC. My friend and I had to buy the whole house out so they would run it for us…
I worked at Lincoln Square in NYC. There is an IMAX there that is 3-D. They get tht 3-D from TWO prints running. So they have two IMAX projectors running. They have some very fancy headgear they you wear and the lens over each eye electronically turn on and off. (the opacity of them that is) This is done in synch with the two projectors. So the left eye projector is showing a frame the left eye on your visor is what you seee through Then as that projector closes it’s shutter your left eye on the headset ‘winks’ as your right lens ‘opens’ and you see the image from the right projector.
And, amazingly enough, standard AC electricity in Europe (where they use PAL) is 50 Hz, while standard AC electricity in America (where they use NTSC) is 60 Hz.
Thanks for the education, guys, especially the link to the Kevin Brownlow article, bonzer. Once again, something you “know” to be true proves to be not quite so simple after all.
Cartooniverse, is that T-grain film stock pretty universal now (more than one manufacturer)? I had noticed a difference in image quality between films in the seventies and more recent ones but thought vaguely it had something to do with lighting or perhaps print production improvements.
Eastman Kodak developed it first, but design patents run out eventually. For all I know, the T-Grain design is used by Fuji and Agfa as well. ( the two other motion picture stocks favored commercially ). I can’t find a definitive answer as to whether or not the design patent on the T-Grain formula is expired. Harrumph. I’ll call Rochester tomorrow.
Cartooniverse, I have always been under the impression that 48fps does lead to a great increase in perceived sharpness of the projected print. I am not a professional cinematographer, though I do work alongside them, so edumacate me on any misconceptions I may be holding. Most of my understanding of the advantages of shooting/ projecting at 48fps is the result of a drunken conversation with Denny Clairmont, so I could be talking out of my @ss.
If a film is shot at 24fps with a 180 shutter the effective shutter speed is 1/48th. Shooting at 48fps with the same shutter leads to a 1/96th shutter speed, which should decrease motion blur therby increasing the sharpness of the image. Also, Film grain is random in distribution, unlike the pixels of a digital image. As it was explained to me, this means that while with a digital image-capture the same “points” are always displayed, with film it’s like capturing random pixels with each frame. Then when projecting, the effective number of points (“pixels”) is increased by the factor of the number of frames per second. Obviously there will be considerable overlap in what information the grain is capturing, but at 48fps you would certainly be increasing the perceived sharpness of the projected image. Finally, the same brain mechanism which allows persistence of vision also allows the brain to perceive increased sharpness in an image projected multiple times (with random film grain patterns) where a simple blow- up of one frame would be blurry. I am fairly sure you were talking about this in your post, so maybe you were talking about actual sharpness as opposed to perceived sharpness.
To sum up, with a 48fps image capture you decrease motion blur, thereby increasing sharpness. When projecting, persistence of vision combined with the random grain pattern of film being projected at double the normal rate does indeed lead to a considerable increase in perceived sharpness.
How the loss of one stop of light in the capture process affects the speed of the film selected (and thereby the density of the grain structure) is another matter.
(On a total side note, DAMN I’m gonna miss film if digital takes over!)
Cartooniverse: You are right that the raw resolution of the film wasn’t changed, but the perception of increased clarity in Showscan was due to the fact that at 60ps you couldn’t perceive individual grains in the images, as you sometimes can at 24 fps. [I note on preview that **coven** has provided more detail than I wanted to go into on this point. I believe he is essentially correct in his post.]
If you haven’t seen it, you might not believe it, but looking at a Showscan image was like looking through a window at reality. (I think Vorlon will back me up on this.) In fact, in his first Showscan film, New Magic, Trumbull made use of that fact to pull off a really amazing trick on the audience. Showscan was truly one of the greatest cinematic experiences ever. Better, in certain respects, than IMAX.
Also, if you are saying that you can get strobing at 60 fps, you may be right theoretically, but practically speaking the high frame rate virtually eliminated it. One reason is that, unlike in 24 fps projection, which projects each frame twice, the Showscan projector only showed each image once. This was another of the great things about Showscan: fast-moving POV shots were very realistic because there was none of that jittering that makes quick pans or fast motion look so bad in conventional films.
Trumbull said that 72 fps and even higher frame rates were noticeably better than 60 fps, but he settled on 60 as more practical, since the U.S. power grid operates at 60 Hz and there are therefore more 60 Hz motors and other electrical components available. Sixty fps and the wide aspect ratio also made Showscan a perfect match for HDTV.
Finally, as a cinematographer, you should know that camera negative stock is 65mm, not 70mm. Seventy is used in release prints to allow space for the sound track that camera negative film doesn’t need. (But perhaps you knew than and were just keeping it simple for the civilians.)
Vorlon: I, too, first saw Showscan at the Showbiz Pizza Parlor in Fairfax, VA. It’s too bad they didn’t have a better business plan to spread the format around. But are you really sure about their rides running at 48? I don’t think I had heard about that. The original Tour of the Universe in Toronto was 60fps, although it is long gone now.
Zebra: The Loews Lincoln Square IMAX theater has one dual-strip 3D projector, not two projectors. Imax Corporation does make two-projector 3D systems, but the bigger installations use the dual-rotor GT system, which feeds both 70mm strips through a single projector (with two lamphouses of up to 15 kW each, water-cooled!).
BTW, In response to Showscan, in the early 1990s Imax Corp. developed its own high-speed system, called IMAX HD, which was the standard IMAX frame running at 48 fps. It was amazing, but only one film was made in that format (for the Canadian pavilion at the 1992 world’s fair in Seville, Spain). Even though it was an incredible experience, there were some strobing effects still visible that Showscan did not have. (Of course, the IMAX frame is 3 times larger than the Showscan frame, which exacerbates the problem.) If you’d like to see a comparison of the size of the IMAX frame with other giant-screen and conventional film formats, see here.
Oh yeah, Trumbull’s patent on the Showscan system, which covered motion pictures from 49-72 fps, has recently expired, so if you want to go out and make your own Showscan film and theater, you can. If you do, please let me know.
(If I’ve come across as too much of a know-it-all in this post, sorry, but IMAX and other giant-screen film formats are my business.)
)