Page 1 of 1

Projection Speeds for Silent Films

Posted: Mon Oct 25, 2010 7:20 pm
by SFBOB
Up to what year(s) were projectors hand cranked for silent films? During this period weren't projection speeds throughout a given film controlled by the projectionist?

Posted: Mon Oct 25, 2010 9:20 pm
by Gagman 66
:o In the late Teen's, certainly by 1920 most places had motor-driven projectors. Ironically enough, Motor-driven camera's didn't start popping up until 1925. Although, allot of film-makers seemed to ignore them. They just were not so versatile as the Hand-cranked ones were.

Posted: Tue Oct 26, 2010 9:29 pm
by sepiatone
In Europe Abel Gance may have had a motor run camera(?a Pathe) on J'Accuse and certainly had one by "Napoleon". In the US cameraman Alfred Gilks can be seen in one Paramount still operating a motorised Bell & Howell for "Old Ironsides"(1926).

Posted: Wed Oct 27, 2010 5:40 am
by drednm
'Tis a mystery to me.

My LOC copy of Getting Mary Married says it was trasnferred from 16MM at 19 FPS. Running time in 68 minutes.

When I converted the file to MPEG via AVS, the only choice was 25 FPS, but the run time is still 68 minutes.

I'm no tekkie but does this make sense?

Posted: Wed Oct 27, 2010 9:07 am
by Rodney
drednm wrote:'Tis a mystery to me.

My LOC copy of Getting Mary Married says it was trasnferred from 16MM at 19 FPS. Running time in 68 minutes.

When I converted the file to MPEG via AVS, the only choice was 25 FPS, but the run time is still 68 minutes.

I'm no tekkie but does this make sense?
Sure. Previous to sound films, all projector speeds were variable. Without sound, it didn't matter (though your orchestra leader would get pissed off if the screen cues come at a different rate from show to show). Just as the music was different at every theater, the projection speed could be different too. This was more pronounced before the mid-1920s, but there's plenty of evidence for variability in film speed up to the introduction of sound, including some theaters where the practice was to adjust the speed at different points during the picture for best effect (as in the notes for the NYC premiere score for the 1927 Wings). Where other theaters use a constant speed (the published cue sheet for Wings doesn't indicate any speed changes). There are reports of unscrupulous exhibitors running the most popular films (Pickford, Chaplin, Fairbanks, Lloyd) extra fast so that they could squeeze more shows in per night and sell more tickets. This may have lead to a film speed race where cameras were cranked faster to compensate (though at least one important cameramen denied that their cranking speed changed at all).

Once sound came in, a standard speed was required so that projectionists could keep the film and discs in sync, or (once sound tracks were put on the film) to keep the sound from being distorted. 24 fps was picked, since it was reasonably common by that time.

Video also has defined frame rates, but they aren't the same as film -- NTSC (used in the U.S. and elsewhere) runs at 30 fps (for odd technical reasons it's actually 29.97 fps, but now I'm geeking out). The PAL format used in Europe and elsewhere runs at 25 fps. That seems to be what you have.

When early silent films are shown at 24 fps, they're too fast (this is not controversial even among the extremes of opinion on this group). So, how do you screen these films? At high-end revival theaters, the projectors are still variable speed, so no problem; but many theaters only have sound-speed projectors. Experiments were made in "stretch printing," where every other frame was doubled. If you do the math, this slows the film from 24 fps to 16 fps (in every 24 frames you get 8 doubled frames and 8 non-doubled frames = 16 film frames). This is too slow for most silent films and also causes a noticeable stutter in the movement. Chaplin did this for some of his sound re-releases of early films since he thought 24 fps was too fast; but abandoned it because the solution was worse than the problem. Stretch-printing is pretty much universally regarded now as a failure.

It is possible to do this "frame doubling" solution when transferring to video with greater sophistication to create frame speeds between 24 and 16. In fact, this has to be done with all sound films, since to go from 24 film frames to 30 video frames you need to double each fifth frame. I'm simplifying, but sometimes frames are doubled, other times they're "blended." Sometimes it's done really badly, and motion that should be linear pulses instead. But the NTSC video player is still showing you 30 video frames in every second (or, in your case, 25).

Usually on a DVD player there's a technique -- on mine it's a combination of pressing the pause button then a skip button -- to step through the video frame by frame. By doing this, you can see the doubled frames (this works really well on scenes with fast action, or transfers from dirty prints, since the specks distinguish the frames well), and you can work out how many actual frames there are in each second. For your transfer, going at 19 fps on a 25 fps video, I'd guess that there are six doubled frames in each batch of 25.

Now, if you were to transfer your 25 fps PAL video to 30 fps NTSC, you'd have to impose an algorithm that doubles frames. But your transfer already has doubled frames. Now the algorithm might double them further (it can't tell which are doubled already, it just sees 25 frames). Now you've likely got a stuttery mess. Serge Bromberg's attitude is that you should always make transfers from film at 25 fps (he's in France, a PAL country, so at 25 fps there are no doubled frames), then slow it down to the speed you're after from there. If you do your original transfer at a different speed, you can never adjust it gracefully.

Hope that helps.

Posted: Wed Oct 27, 2010 9:28 am
by drednm
Thank you Rodney.... As I said Getting Mary Married started out at 19 FPS but I see no difference between what LOC sent me and the 25 FPS transfer I made. Maybe the hand is quicker than the eye.

I had no idea of this difference between PAL and NTSC but it explains why the Will Hay films I got from Australia, which I transferred from PAL to NTSC, have a odd jerky motion especially during "fast-paced" scenes.

I think I had read that the sound re-issue of The Birth of a Nation had some weird speed problems because of the music track.

Wouldn't the FPS have been in the shooting notes from most silents? Or what is left to the projectionist to figure out?

Posted: Wed Oct 27, 2010 9:34 am
by boblipton
Projectionists were given wide latitude -- mostly because what could you do about it. You often got theater owners running things faster in order to get another show in.....

Although sound fixed projection speed to deal with the sound track, there was a lot of leeway in sound projection. I read a piece about how the sound was carefully composed for BONNIE AND CLYDE -- and one of the producers (Beatty?) was really annoyed at a British technician who had 'fixed' the 'problem' in projection.

Bob

Posted: Wed Oct 27, 2010 10:16 am
by Rodney
boblipton wrote:Projectionists were given wide latitude -- mostly because what could you do about it. You often got theater owners running things faster in order to get another show in.....

Although sound fixed projection speed to deal with the sound track, there was a lot of leeway in sound projection. I read a piece about how the sound was carefully composed for BONNIE AND CLYDE -- and one of the producers (Beatty?) was really annoyed at a British technician who had 'fixed' the 'problem' in projection.

Bob
I know some musicians with perfect pitch, and when projectors run at the wrong speed, it bugs them no end. Especially with classical standards where they know what pitch it's supposed to be at...

Posted: Wed Oct 27, 2010 10:26 am
by Rodney
drednm wrote:Thank you Rodney.... As I said Getting Mary Married started out at 19 FPS but I see no difference between what LOC sent me and the 25 FPS transfer I made. Maybe the hand is quicker than the eye.
Well, apparently I confused you with my lengthy answer. Sorry.

Despite the label, your source for Getting Mary Married is running at 25 video frames per second, because PAL video can't run at any other speed. But LOC's transfer already has frame doubling introduced, so that in those 25 video frames there are only 19 film frames. In each second you see 19 film frames, distributed across 25 video frames. Since "19" is the number that matters, LOC put "19 frames per second" on the label. If you were to run the film in a projector showing 19 fps, it would match the speed on both of your videos.

Of course, it's possible that the LOC disc is in NTSC (at 30 fps) and your computer converted it to 25 fps (perhaps to save space). If so, I'd do your transfer over at 30 fps to avoid weird timing artifacts. Keep it at whatever speed or video format they sent it to you in, if at all possible.

Posted: Wed Oct 27, 2010 10:51 am
by sc1957
Two posts here mention the idea that "theater owners might run films faster to fit in more showings." I have a hard time believing that this would really work -- that they'd really save enough time to drop another film into the mix. Maybe if all the films are an hour or less in length. Has running films faster to show more of them actually been documented, or is it just a "lore?"

PS: And how can you determine that films were run faster in order to fit more in, versus being run faster so the theater manager could go home early?

Posted: Wed Oct 27, 2010 11:09 am
by silentfilm
OK, here's everything you ever wanted to know about projection speeds, but were afraid to ask, courtesy of Oscar-winner Kevin Brownlow...

http://www.cinemaweb.com/silentfilm/boo ... 8_kb_2.htm

Posted: Wed Oct 27, 2010 11:18 am
by Rodney
sc1957 wrote:Two posts here mention the idea that "theater owners might run films faster to fit in more showings." I have a hard time believing that this would really work -- that they'd really save enough time to drop another film into the mix. Maybe if all the films are an hour or less in length. Has running films faster to show more of them actually been documented, or is it just a "lore?"

PS: And how can you determine that films were run faster in order to fit more in, versus being run faster so the theater manager could go home early?
It's not to add another film -- it's to have another audience turnover in the evening, which means a complete set of new ticket sales. It's a story I've come across in a couple of books -- one was a biography of Pickford (most of whose movies were around an hour in the early days), but I forget which at the moment. If I had the reference at hand I'd give it to you, but I don't. You can believe it or not, of course, as with all of this stuff.

But as for whether it's possible, I can do that just sitting here. Let's say you've got a program of films that lasts a particular time at 18 fps. Crank it at 22.5 fps, and you can fit five shows in the place of four. Crank it at 24 fps, and you can fit four shows in the place of three. Many theaters didn't advertise "start times," you just walked in when you arrived and left when you got to "this is where we came in." I can well imagine the manager telling the projectionist "We've got long lines, crank faster so the folks inside will leave sooner and we can sell their seats again!"

This wasn't true everywhere, of course, since projection practices (and integrity) varied. We're mostly talking about inexpensive nickelodeons, not the movie palaces where they made a pretense of art.

When the "Lord of the Rings" films came out I saw an article where local theater managers were complaining that they couldn't fill as many screenings per day as they did with other shows, because the films were too darn long. I'm sure they'd have sped up the projection speed if it weren't for that pesky sound track.

And sure, the manager might also want to go home early (and with dud films that no one came to see, he probably did), but I bet he'd keep the place open as long as there are lines outside of people with money in their pockets wanting to see "Our Mary."

Posted: Wed Oct 27, 2010 12:46 pm
by Chris Snowden
Rodney wrote: It's a story I've come across in a couple of books -- one was a biography of Pickford (most of whose movies were around an hour in the early days), but I forget which at the moment. If I had the reference at hand I'd give it to you, but I don't. You can believe it or not, of course, as with all of this stuff.

But as for whether it's possible, I can do that just sitting here. Let's say you've got a program of films that lasts a particular time at 18 fps. Crank it at 22.5 fps, and you can fit five shows in the place of four. Crank it at 24 fps, and you can fit four shows in the place of three. Many theaters didn't advertise "start times," you just walked in when you arrived and left when you got to "this is where we came in." I can well imagine the manager telling the projectionist "We've got long lines, crank faster so the folks inside will leave sooner and we can sell their seats again!"

This wasn't true everywhere, of course, since projection practices (and integrity) varied. We're mostly talking about inexpensive nickelodeons, not the movie palaces where they made a pretense of art.
I doubt that this sort of thing happened very often, except maybe in rural areas where Mr. Exhibitor had no competition. Because if there was another theater in town, running films at a reasonable speed, all his customers would gravitate to that theater instead, and Mr. Exhibitor would be out of business.

In watching silent films, there are few things more aggravating than intertitles that disappear before you've had a chance to read them.

If an exhibitor really wanted to squeeze in an extra show, I believe he'd more likely just drop a short subject, thus shortening the overall program. If his feature that day was more expensive to rent than usual, he'd be looking to save a few bucks on short subjects anyway.

Posted: Wed Oct 27, 2010 1:04 pm
by Rodney
Chris Snowden wrote:
Rodney wrote: It's a story I've come across in a couple of books -- one was a biography of Pickford (most of whose movies were around an hour in the early days), but I forget which at the moment. If I had the reference at hand I'd give it to you, but I don't. You can believe it or not, of course, as with all of this stuff.

But as for whether it's possible, I can do that just sitting here. Let's say you've got a program of films that lasts a particular time at 18 fps. Crank it at 22.5 fps, and you can fit five shows in the place of four. Crank it at 24 fps, and you can fit four shows in the place of three. Many theaters didn't advertise "start times," you just walked in when you arrived and left when you got to "this is where we came in." I can well imagine the manager telling the projectionist "We've got long lines, crank faster so the folks inside will leave sooner and we can sell their seats again!"

This wasn't true everywhere, of course, since projection practices (and integrity) varied. We're mostly talking about inexpensive nickelodeons, not the movie palaces where they made a pretense of art.
I doubt that this sort of thing happened very often, except maybe in rural areas where Mr. Exhibitor had no competition. Because if there was another theater in town, running films at a reasonable speed, all his customers would gravitate to that theater instead, and Mr. Exhibitor would be out of business.

In watching silent films, there are few things more aggravating than intertitles that disappear before you've had a chance to read them.

If an exhibitor really wanted to squeeze in an extra show, I believe he'd more likely just drop a short subject, thus shortening the overall program. If his feature that day was more expensive to rent than usual, he'd be looking to save a few bucks on short subjects anyway.
That's all good reasoning. If I come across those quotes again, I'll post them here (and I understand that just because something's in a book doesn't mean it's true... it just means that somebody else said it too).

But as for people abandoning theaters that have bad practice -- I don't think that's always true except in an ideal world. There's the quote in Brownlow's article about the projectionist who had to run movies at different speeds at different times of day, and we know that matinees for kids were run without music, and reports of theater organs sadly out of tune -- but why tune them when that requires shutting down the theater and hiring someone.

Posted: Sun Oct 31, 2010 6:04 am
by josemas
sc1957 wrote:Two posts here mention the idea that "theater owners might run films faster to fit in more showings." I have a hard time believing that this would really work -- that they'd really save enough time to drop another film into the mix. Maybe if all the films are an hour or less in length. Has running films faster to show more of them actually been documented, or is it just a "lore?"

PS: And how can you determine that films were run faster in order to fit more in, versus being run faster so the theater manager could go home early?
Back in the late 1980s I worked with a gentleman who had worked as a projectionists for decades. He had started when he was a teenager during the silent era at a small town theater using a hand cranked projector.
I once asked him as to whether he was ever asked to crank faster by the theater owners. He told me "no" but then admitted that when ever he had a "hot date" on a Friday or Saturday night he would crank up the speed considerably in order that he could get out of there quicker. He told me that nobody ever complained when he ran these weekend films (usually westerns) faster.

Joe Moore