24fps film & sound-on-film

[JaredSMark]

Way back when, on top of the obvious reasons they settled on 24fps as the standard (cost of film), did they settle on the 24fps frame rate since it was pretty much the least speed they could run tape when recording sound-on-film to retain high frequencies? At least, is that one of the reasons? I also heard it may be due to the noise floor.

 

[j1clark@ucsd.edu]

Way back when, on top of the obvious reasons they settled on 24fps as the standard (cost of film), did they settle on the 24fps frame rate since it was pretty much the least speed they could run tape when recording sound-on-film to retain high frequencies? At least, is that one of the reasons? I also heard it may be due to the noise floor.

At the time that sound was introduced, there was no 'tape' recording. Well, in general. 'tape' recording was a post WWII development. At the time of the first 'talkies', the two methods that have been somewhat 'proven' were recording mechanically to disk, similar to recording phonograph records, and optical recording on film. 'recording on film' won out for a number of reasons.

Hence, 24 fps, was 'found' to be the lowest speed to provide the 'best quality audio recording', while recording the film image.

There may have been other reasons why 24... as opposed to 25... but it was sort of a cost/benefit type choice.

 

[Richard Crowley]

Sound was not recorded ON film (the same film or different stock) at the time that 24 FPS was established as the standard speed. As Mr Clark says, sound was recorded on transcription disks (phonograph records) which had their speed synchronized to the film speed.

24 FPS was selected for visual reasons, the slowest (=least expensive) speed that would work with the POV (persistance of vision) of the average viewer to produce the illusion of a moving picture. Of course technology has allowed us to raise the frame rate for electronic media and higher rates are used for premium production and exhibition.

24 frames per second was chosen because it was the slowest (and thus cheapest) speed which allowed for sufficient sound quality.
http://en.wikipedia.org/wiki/Film#Technology

The first half of this sentence is correct, but the last part is completely fanciful and bogus. Note that WIkipedia has flagged this article for multiple problems, including sources and atrribution.

This article has a better explanation of frame rates and appears more factual: http://en.wikipedia.org/wiki/Frame_rate

 

[Noiz2]

As mentioned above "tape" was first developed during WWII by the Germans. After the war a soldier shipped back two of the machines (prototypes?) and formed AMPEX. Sound was recorded to optical tracks on 35mm film. I believe this was a second "sound camera" that recorded the optical track and then this was optically printed back to the release print. Originally sound was on records that played back in sync to the film, kind of like SDDS does today. But sync was hit and miss and optical tracks printed on the film became standard.

I also remember something about 24 being the slowest speed with acceptable wow and flutter. I think that is correct since silent films were shot at around 19-20 FPS. Generally hand cranked so speed was a bit flexible. One of the reasons most people think of silents as jerky films is because most have only seen silents played back incorrectly at 24fps so everything becomes a bit herky jerky.

With sound you needed a stable speed and picture and sound needed to run at the same speed to stay in sync so film sped up to meet the needs of sound. Probably the last time picture changed to accommodate sound ;~).

 

[JaredSMark]

https://vimeo.com/3388256 - at 1:14 he mentions, "audio engineers determined the absolute minimum speed they needed to optically record synchronized sound was 24fps". So it didn't have much to do with sound quality, but more about the synchronization issues? I'm wondering exactly how this synchronization works and why 24p was the lowest they could use.

I was under the impression that recording 'faster' on tape gave you a better frequency response (kind of like digital sample rate), so that's why I guessed that when the article said "minimum speed that allowed for sufficient sound quality", they were talking about frequency response (quality).

 

[BRH]

24 FPS was selected for visual reasons, the slowest (=least expensive) speed that would work with the POV (persistance of vision) of the average viewer to produce the illusion of a moving picture.

I believe this is correct. 24FPS with a standard 2 bladed shutter would be 48 interruptions per second... just at the point where you eye/brain starts to notice the flicker. Has nothing to do with sound.

 

[j1clark@ucsd.edu]

I believe this is correct. 24FPS with a standard 2 bladed shutter would be 48 interruptions per second... just at the point where you eye/brain starts to notice the flicker. Has nothing to do with sound.

The silent films ranged from 14-24 fps. This was set in stone with the introduction of sound, since the sound recording/playback was 'better' than when 20-24, 'common enough' silent speeds, were used. The actual film transport rate was set at 18 in per second.

I'm pretty sure that if the 'studios' at the time could have gotten away with 16 fps 'sound', even with 6 bladed shutters... they would have done it... that would have been 66% less film to buy at any stage of the film making process.

There are two basic ways that sound was recorded optically. 1) varying the density of the optical track, 2) varying the width of the optical track. The variable density option was the first used. This consisted of a light controlled by the 'audio' signal, and a slit opening that exposed the negative film media to the light.

If an 'audio signal' varies faster than the transport moves the film, the 'slit' acts as an integrator, and as such is a 'low pass' filter. Therefore the lower the speed, the lower the cutoff frequency of this low pass filter. For 'reasonable' sound reproduction, one would want to have 5-10 KHz worth of 'bandwidth', and so, the 'fastest' film transport that would deliver that, would be the 'better' choice, than a slower speed.

At the time in the US 24 fps was the standard 'top end'. There were demonstrations of higher speed, but at that rate, given, cost of film, exposure time required for the visual image(*), exposure time for the 'audio' optical track, 24 seems to have been the 'best' solution.

I still don't know 'why 24' as opposed to 25... but it may be that the perception of motion starts about 12 fps, so 24 is 'double' that. But still some number of silent films where not shot at 24 fps. However blades in the projectors did chop the image to smooth it out.

*What seems to escape some people is that for a given ISO value, doubling the film transport speed, halves the exposure, and therefore requires double the light for the same resultant density... then in film film there's 'reciprocity' failure when this 'simple' law fails, and even more light is needed than a simple doubling...

In the case of 'optical audio recording' this would correspond to faster transport, would require a doubling of the light until reciprocity failure took over, and requiring even more light...

 

[Noiz2]

Well as you can see this is a little like asking what ADR stands for.

I'm not sure about the 48 frame thing. In science classes we were taught that it takes 1/4 of a second for the eye to register an image. Certainly they were getting smooth motion with 16 fps so?

The sound quality issue has to do with speed but also the way a projector works. There is a loop that goes through the plate and is projected. The film jumps frame to frame so the "audio head" is located elsewhere where the transport is smoothed down. This is where the speed comes into play. Even then sound wasn't at a frame rate but at a speed.

As Richard mentioned, originally the recordings were on platters. But once things standardized recording was on optical recorders. Sync is simple. All the machines move sprocket for sprocket together. If you have a start mark you are mechanically locked and always in sync. For release all standard films were released with an optical track that was part of the film so sync was never a playback issue.

Yes in theory faster speed in analog tape gets you better fidelity. But a lot also depends on things like bias and noise reduction. This is kind of moot though when your talking about optical tracks, especially early mono ones. No noise reduction and the Academy curve eliminates pretty much anything over 10K and below 100hz. So speed doesn't get you much.

 

[Rick R]

I believe this is correct. 24FPS with a standard 2 bladed shutter would be 48 interruptions per second... just at the point where you eye/brain starts to notice the flicker. Has nothing to do with sound.

Yes, that and because the US electricity is at 60Hz. In some other parts of the world (the UK for instance) is @ 50Hz. and the FPS is 25.
FYI, the confiscated WWII era German tape machines were know as "Magnetophone".

 

[Noiz2]

Yes, that and because the US electricity is at 60Hz. In some other parts of the world (the UK for instance) is @ 50Hz. and the FPS is 25.
FYI, the confiscated WWII era German tape machines were know as "Magnetophone".

Well kind of, but your mixing video and film. 35mm film is 24 everywhere PAL video is at 25fps (because of 50Hz power) and NTSC was 30 fps (for the same reason). The 30 got bastardized into 29.97 to accommodate color (and because some political maneuvers, we were going to have something very similar to PAL).

Film was never synced to the AC freq.

 

[Allan Black]

Yep, in Berlin at the Reichstag, before the latest rebuilding, I saw a reel to reel recorder in their museum. I managed to get around to the back of it and saw the name plate .. 'Magnetophonband 1936'.

In 1944 the Allies couldn't figure out how German orchestras and concerts could b/cast at midnight, it was from the first BASF tape.

As a new kid in the biz, one of the best recorders I ever worked with was the Ampex 350, mono, it was built like a tank. Years later as manager of a studio complex, I got the budget for new recorders. We opted for Ampex and bought 5 stereo Ampex 440s and 2x4 track 440 machines.

They were a disaster, the mono rigs would not rewind a 10" roll of tape. Oh boy I had a terrible time, Ampex was a world away in Redwood City Cal. and was no help. But I had some great tech guys and they rejigged them to spool properly.

On the other hand I worked with 24tk Ampex 124s and they were great, but the 440s signalled the end of Ampex as a builder of pro tape recorders. Allen Sides at Oceanway in Cal bought a heap of 124s and rebuilt the electronics.

Incidentally we had MCI consoles from the late Mac Emerman in Florida, great desks from a great guy. The Bee Gees and Bruce Jackson who started Apogee, thought so too, I worked with all of them, very sad only Robin left now.

The day Bruce left Sydney to join Elvis as audio guru, he came by Col Joyes studio to say bye, and I bought his 2 JBL 2011 speakers for $A400. He had them in the trunk of a car!

Cheers.

 

[David G. Smith]

Film was never synced to the AC freq.

Well, yes, you are correct that picture and sound frame rates were not based on the AC frequency, but... It is my understanding that the camera and optical recorder were both slaved to each other using AC powered electric motors, or more specifically, their AC motors were slaved from a third device driving both camera and optical recorder from the AC mains. Moving picture and sound recording with machines not connected together did not happen until the advent of crystal sync in the early 1960s. (EDIT: Crystal sync also used a 60 cycle frequency to drive the recorder and camera motors for sync at 24 fps).

The 24fps standard was finalized by the precursor to the SMPTE (The SMPE, the Society of Motion Picture Engineers), who also changed the aspect ratio of the film to 1.33:1 (4:3) so as to give space for the optical sound track. I would imagine that the frame rate was a choice to give the best picture AND sound quality as efficiently as possible with more than a smidge of arbitrary final choosing for the best compromise.

 

[Noiz2]

You know I'm not sure about production sound in the early, post disk, days. It's quite possible that it was an electrical sync signal.

Post was a sprocket sync. By the time I got to it that was electrically transmitted.

There was also a sync signal that was sent out of the camera (I want to say this was pilot tone but I may be wrong). I had a Bolex that had a connector on the (Electric) motor that sent a signal to a recorder so that they would run at the same speed. There were also dual system projectors (interlock) that were (electrical or mechanical?) that played back the audio "film" locked to picture.

In the big "machine rooms" (circa 1990's) you had lots of reels of magnetic 35mm film and those reels had start marks. You set up all the dubbers with the start marks in the frame and the film also at the start mark and you flipped a switch that locked the dubbers with the film and they all moved together. I'm sure it was electrical but it was also effectively a sprocket to sprocket sync.

What complicates things in the early days is that AC power was a new thing in the twenties. The first cities electrified were done (in the US) with DC so no sync freq. So I suspect that the electrical sync if used was self generated. It was in the later years because the camera might vary a bit (speed wise) and you wanted the sound to vary with it so the sync came from the camera.

The wikapedia and other sources report a much wider variance in seed than I am familiar with so I'm wondering if it varied a lot and then evolved to being around 19fps or if that just happens to be the silents that I worked on happened to be around that speed???

In some ways it's a bit of a moot point. The part that is really pertinent to the OP's question is that money was the biggest deciding factor. The frame rate was driven by Hollywood because at the time it was the biggest producer of films by a wide margin and Hollywoods decision would have been driven by the cheapest frame rate that is acceptable. It was ether sound of vision but it wasn't "the best" but the least we can get by with.

Video puts it all in flux but a constant in film has been compatibility. As long as we were releasing everything on 35mm the "big boys" set a standard and everyone else followed and for a hundred years or so any film could be played back on anybody's projector and it "just worked".

Now it's a bit of a mess and until a new "standard" evolves it's going to be a constant battle to find compatibility.

Your first quote is I think very on point to the discussion. We are spending way too much time worrying about technology and way too little worrying about content.

 

[j1clark@ucsd.edu]

Well, yes, you are correct that picture and sound frame rates were not based on the AC frequency, but... It is my understanding that the camera and optical recorder were both slaved to each other using AC powered electric motors, or more specifically, their AC motors were slaved from a third device driving both camera and optical recorder from the AC mains. Moving picture and sound recording with machines not connected together did not happen until the advent of crystal sync in the early 1960s. (EDIT: Crystal sync also used a 60 cycle frequency to drive the recorder and camera motors for sync at 24 fps).

The 24fps standard was finalized by the precursor to the SMPTE (The SMPE, the Society of Motion Picture Engineers), who also changed the aspect ratio of the film to 1.33:1 (4:3) so as to give space for the optical sound track. I would imagine that the frame rate was a choice to give the best picture AND sound quality as efficiently as possible with more than a smidge of arbitrary final choosing for the best compromise.

Ah... but California Edison which served much of So Cal... ran at 50 Hz... during the critical period of sound development. Furthermore, even though the electric clock was invented in the 10's, power company tolerances were not regularized until much later even for 60 Hz.

The 24 fps of film I think was a totally independent selection, unlike the TV frame/field frequences.

I don't think the Phase Locked Loop had been invented yet to even 'sync' up... that was something that developed in the 30's along with prewar TV receivers.

 

[David G. Smith]

Ah... but California Edison which served much of So Cal... ran at 50 Hz... during the critical period of sound development. Furthermore, even though the electric clock was invented in the 10's, power company tolerances were not regularized until much later even for 60 Hz.

The 24 fps of film I think was a totally independent selection, unlike the TV frame/field frequences.

I don't think the Phase Locked Loop had been invented yet to even 'sync' up... that was something that developed in the 30's along with prewar TV receivers.

Now I am starting to get confused. What time period are we talking about? The first commercial sync sound system was the Fox Vitaphone system, introduced in 1927 with "The Jazz Singer". This was a sound on disc system. That system was quickly abandoned for the superior Western Electric recording system that used optical sound on film. I am pretty sure that the Western Electric system slaved camera and sound recorder via the AC mains, but was not dependent on the AC frequency, as far as I know.

 

[Richard Crowley]

I am pretty sure that the Western Electric system slaved camera and sound recorder via the AC mains, but was not dependent on the AC frequency, as far as I know.

The cameras and sound recorders ("dubbers") were operated from a private AC mains supply on the studio property. And they were very much dependent on the AC frequency, else there would have been no point to installing the system. They used 3-phase power from a studio-lot alternator to operate all the speed-dependent equipment around the studio. I believe it was 60Hz and was not derived from the local utility power.

When the sound recorder(s) locked in to the mains frequency, the operator would proclaim "Lock!" or "Speed!" so that the director would know that the shot could proceed with sync audio.

The photo below shows three of these synch-driven "dubbers". The black machine behind the operator is likely an optical recorder, in that era, used to copy (dub) the final mix (which was on 35mm mag stock) to the optical sound track which was printed along with the picture for theatrical distribution.

 

[David G. Smith]

Ok Richard, just to clarify... So the mains power was used to run the camera and recorder at a constant speed, based in the 60 cycle frequency, but the frame rate of 24 fps was independent of that. The 60 cycle time clock (I don't know the correct nomenclature for this) could have as easily driven them at 20 fps, 30 fps, or what ever, correct?

 

[Richard Crowley]

A self-created, local, proprietary "mains" power grid was used to drive (and synchronize) all the cameras and sound recorders (and players and projectors, for that matter). This was NOT the utility mains power that neighbors use to operate their homes or businesses. It was three-phase (similar to the utility) to make it easier to get rotating loads (like motors) to lock). They could have chosen any frequency they liked. The frequency controlled the linear speed of the film based on the design of the equipment, and they were designed to produce 24FPS based on the sync power frequency. Yes, they could have designed the equipment to operate at any other speed by using different gears, etc.

 

[Noiz2]

Guys...

First you need to separate PRODUCTION recording and POST PRODUCTION recording (or actually "rerecording).

Richard you are correct those are dubbers. Other than most of them had switched to dual machines these are the same machines that were in everyday use up to the late 1990's. They are post production recorders. The machine in the back is probably a projector, notice the hole in the wall. The guy looks like he is setting up the last one so it looks like he is setting up for a mix. It looks like full coat so those are probably 3 to 6 tracks a peace, but hard to tell.

Mix stage projectors put out a pulse for each sprocket hole, the dubbers lock to that signal. It has nothing to do with line frequency. Line frequency didn't come in to play as a common sync signal till B&W TV's used it.

But the question originally was about production recording. Dubbers were not production recorders.

So...

The Jazz Singer was not the first sync sound picture released, it was the first feature length "talking" picture. Jolson ad-libbed a couple of lines and and they kept them in. SUNRISE was the first feature length sync sound picture and it was a Movietone release. Sync sound shorts date back to 1923 and "talkies" in 1924 (of Calvin Coolidge).

Since even I was getting confused I looked a few things up. So VitaPhone (1926-1931) was used on features and a large number of shorts. It recorded to wax discs and indeed sync was maintained by using synchronous motors with a common power source. MovieTone started out as single system sound. The sound was recorded on to an optical track on the same film as the picture. Later they went to a dual system with dedicated recorders, but the single system was in use for newsreels into the late 40's ~ 50's. I couldn't find a specific reference to how sync was maintained on dual system MovieTone but it's probably safe to assume it was similar to the VitaPhone one.

I did find this which jives with what I experienced later.
""Double-system" sound used independent cameras and sound recorders. The first sync sound standard used recorders and cameras both powered by AC (alternating current) motors - essentially clock motors. [SUP][1][/SUP]Later the 50Hz or 60Hz sine wave, called a "Pilotone," was recorded on a second parallel track of an audio recorder.
In double-system filming, speed variations of camera and recorder, as well as the elasticity of the magnetic recording tape, requires some positive means of keying the dialogue to its appropriate film frame.
The inclusion on the sound recorder of a second, parallel, sync or "Pilotone" track, is the most common method in use today.
If a location camera is driven by a DC motor, with some sort of governor control to hold it fairly accurate at 24 f.p.s., a sync pulse generator geared to the movement or motor shaft could be employed to provide the sync pulse output. A cable conducts the sync pulse from camera to sound recorder. The sync pulse is typically a sine wave of 50 or 60 Hz with an RMS amplitude of approximately 1 volt.[SUP][2]"


[/SUP]

VitaPhone used mechanical sync for playback. After 1931 the name was still used but it was a sound on film system then.

 

[David G. Smith]

I started out shooting double system sound with Super 8mm film. What I used was a system that used the pc flash contact of the Super 8 camera (Available on higher end Super 8 cameras, I was shooting with a Canon 1014 XLS) which gave a pulse per frame (Designed to fire a flash gun per frame during time lapse filming). I had a gadget that took the pulse and turned it into a sinewave signal that was recorded to one track of any stereo recorder (I used a Tascam Porta Studio 4 track cassette recorder). For post, I used a Super 8 Sound Super 8mm fullcoat recorder. What you did was run the audio track in to be recorded to the fullcoat and feed the sync track into a socket and this drove the recorder and allowed you to "Resolve" the sound so that it would record the audio in frame by frame sync to the original camera recording. You then took the fullcoat and synced it on an editing bench to the film. It was a major pain in the butt.