Film-Tech Forum ARCHIVE: Dolby SR-D or SDDS at 16 fps?

This is topic Dolby SR-D or SDDS at 16 fps? in forum Film Handlers' Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=1;t=004470

Posted by Joe Beres (Member # 678) on 01-22-2003, 12:02 PM:

Please forgive my ignorance on the subject of digital soundtracks, but I'm curious about something.

Let's say an archive restores an early silent film, maintaining the fill-frame 1.33 aspect ration, meant to be run at 16 fps. (Keeping in mind that it is frowned upon in the archive/restoration world to print up to 24 fps) Obviously, the analog soundtrack area is unavailable, but conceiveably an SR-D track (or SDDS), of a score for example, could be applied outside of the image. My question is would either of the digital tracks work at 16fps if recorded and encoded for that speed, or do they only work at or around 24 fps? I know that the lack of analog back-up could be an issue, and that most archives would never add a digital soundtrack to a print of a true silent film. However, i am curious about the logistics of the digital systems and whether this potentially could work or not.

Posted by Joe Redifer (Member # 3) on 01-22-2003, 12:28 PM:

I would think so since they'd only have to worry about mono or stereo at best (not 5.1 or 8 channel). But the cost of designing new software and/or hardware to make it play would kind of make it not worth it, IMO.

Why not just press "MUTE" on your processor and have true analog silence?

Posted by Paul Linfesty (Member # 214) on 01-22-2003, 12:33 PM:

Actually, DTS timecode sounds like a good bet here. They already have a system that can be used for 16mm film.

Posted by Ian Price (Member # 14) on 01-22-2003, 12:43 PM:

But I'm betting that the DTS sync track is in the visible area of Full-Frame, Silent Speed film.

Posted by Lionel Fouillen (Member # 1475) on 01-22-2003, 12:45 PM:

From the little I know about movie digital sound, DTS would fit better than Dolby or Sony in my opinion. I wonder whether on-film digital sound could ever be possible at 16fps considering the reduced film length, not to mention the "perforation jump" of SR-D.

This is an interesting topic. Any expert out there to think about it and share conclusions?

Posted by Daryl C. W. O'Shea (Member # 1303) on 01-22-2003, 02:08 PM:

Depending on how the variable bitrate decoding is handled in the Dolby processor 16fps may or may not be possible. Since the easiest way to do variable bitrates (without building a ton more custom hardware) is with software, I'll assume that this is how it is done. That said, there's no reason why a firm-ware update couldn't make 16fps digital sound happen.

Last time I tried something like this, I just had the sound totally drop-out anywhere under probably 20fps.

AFAIK, the DTS track does fall on the Academy frame area.

Since, AFAIK, the SDDS system is also variable bitrate, it should be able to do the same as I said about a Dolby processor.

Posted by John Pytlak (Member # 331) on 01-22-2003, 02:53 PM:

The position of the (DTS) data track on 35mm prints is specified by SMPTE Recommended Practice RP115 as having a centerline that is 0.297 inches (7.54mm) from the reference edge of the print --- which would be in the projectable area of a full frame silent print.

Posted by Christopher Seo (Member # 7) on 01-22-2003, 03:04 PM:

The range of tolerated speeds for a DA20 was given as 21.4 fps - 25.7 fps. I can't recall if this was an inherent limitation of the hardware of the Cat. 670 video preamp/ADC card.

Posted by Bobby Henderson (Member # 840) on 01-22-2003, 06:17 PM:

All three current digital audio systems would have to be modified into order to work with full aperture 35mm projection at 16fps.

With DTS time code needing to be relocated to a different position, wouldn't that require a different kind of setup for recording the timecode? On normal prints, the time code is white against black. Wouldn't the film prints need some kind of dark background on the outer edge to hold the DTS timecode? Or would the DTS reader pickup the Morse Code looking data if it were printed in reverse?

Dolby Digital and Sony SDDS would not track correctly since so many bits have to be coming in per second off the film for it to work. SRD and SDDS data is compressed and stored in packeted form. It is not linear like LPCM.

My guess is SDDS just wouldn't work at all under 16fps. The 2.2 million bits per second of data both full quality and half quality tracks contain would need a smaller data spot to crunch that data down into 75% of the normal space.

Dolby Digital would have a greater chance to work without modification. The theatrical system normally runs at 320kb/s with 550kb/s of data total for 24 frames. For 5.1, you could drop the bitrate down to 256kb/s (but it probably would not sound all that great). The best bet --if the theatrical systems can process it-- would be encoding a 192kb/s 2.0 mono/stereo track and using the remaining data space available for error correction.

Posted by Karen Hultgren (Member # 395) on 01-22-2003, 08:00 PM:

DTS 16mm formatted film does not have any soundtrack printed on the film itself. The DTS timecode is printed where the mono track usually resides. So, when film is run, the optics picks up the timecode, feeds it to the DTS timecode preamp, Model 376, that conditions the timecode for processing by the DTS digital player. The 6-track digital soundtrack is contained on CD-ROM discs. The timecode tells the player when to play the discs and what to play off the discs.

The process is normally used on new projects by film students and is a low cost solution for presenting 16mm films in 6-track digital sound.

The speed you are referring to (16fps) is what we refer to as "special venue". We'd have to do some testing on that to be sure all would playback correctly. If you are truely interested in persuing this, please send me an email and I'll get you touch with our special venue guy.

Karen at DTS
khultgren@dtsonline.com

Posted by Scott Norwood (Member # 30) on 01-22-2003, 09:23 PM:

Why do we need digital sound with silents? Part of the fun is having live music performed as the film runs, and a real orchestra will sound better than any electronic reproduction system. Many silents are intended to be run with variable-speed projection, anyway, and not at a constant 16 or 18fps. [Smile]

Posted by Joe Beres (Member # 678) on 01-22-2003, 11:31 PM:

It was just a point of curiosity for me. I didn't know what boundaries that technology itself has. It's kind of an unrealistic idea, but an interesting one, I think. I've been lucky to work in places that bring in musician(s) to play along with a silent film, but I'll probably never see some of the big budget silents performed with their original orchestral score. It would be interesting to see the image presented properly, but also hear what people would have heard at the film's premiere. I do agree though, there is nothing like the experience of seeing a silent film accompanied by a great musician or group.

Posted by John Hawkinson (Member # 1135) on 01-23-2003, 12:09 AM:

You could probably also free-run an external music source, since the alignment is not that precise.

How much smarts are in SRD readers, anyhow -- can you get raw access to the CCDs? If so, you could do 16fps 35mm with a Cat700 (+-1) and a DTS-6D with a little converter widget inbetween, just like DTS-16. Of course, at that rate, you might as well just use a shaft encoder.

--jhawk

Posted by William Hooper (Member # 71) on 01-23-2003, 02:06 AM:

Like Scott said, silent films were shot & are projected at different speeds. For the most part, only the earliets were 16-16 fps

For the love of Mike, please don't run them at 16 or 18 fps. It was even expected at the time that in projection speed would be slightly faster than taking speed, & the slight speed increase was part of the "art" of movies.

Anybody who's ever sat through Safety Last, Thief of Baghdad, etc. projected by someone who didn't do their homework well & insisted that silent movies are 16fps knows how much it makes you want to just start slapping the guilty. Those averaged around 24 fps.

Something earlier like Birth of a Nation has much that was taken about 16 fps & was meant to be projected at about 18 fps.
David Pierce's page at

http://www.cinemaweb.com/silentfilm/bookshelf/#March1998

has info on this, & includes period instructions from different films for particular projection speeds. The projection speeds are feet per minute, so you must divide by 3.75 to get frames per second.

And if you look at the later SMPE RP's for silents, you'll find an interval of from +/- 2 feet per minute to +/- 6 feet per minute! Which makes sense, since in the days of hand-cranked cameras & projection a couple of fps margin of error would be inevitable.

When in doubt, run it at 24 fps. Unless it's real old stuff like BOAN or Lumiere's. Historically, the projection speed was what looked best to the projectionist, barring outside influences like the manager telling him to crank them through as fast as possible to get more shows in per day.

Posted by Michael Schaffer (Member # 1204) on 01-23-2003, 05:29 AM:

Theoretically, it should be possible to use SDDS at 16fps as the data blocks are adjacent to each other. While for Dolby Digital vertical tracking is provided by the perforations and the edge of the data blocks, with SDDS vertical sync is provided by patterns in the data blocks. However, the optical image of the blocks would have to be compressed by 1/3 so that the data is read at the rate specified for the system. That would require film stock with a very high resolution and quality. Also the pixels would be rectangular rather than square. But I think it could work as there is a zigzag pattern on the side of the track which is called the zipper track which regulates the electronic "shutter" of the CCDs. That means the zipper track tells the reader when to read a new line of pixels.

BTW, neither Dolby nor SDDS use variable bitrates. It wouldn`t make sense to use variable bitrates if the playback speed is defined by the medium from which the data are read. As both systems use data compression, the goal is to use the limited data rate allowed by the playback medium as well as possible. Dolby uses a technique called Parametrical Bit Allocation. It shifts the amount of data dynamically between the channels depending on which channels need more information at any given moment. But that is not a variable bit rate. The information rate read from the print is constant.

Posted by Leo Enticknap (Member # 534) on 01-23-2003, 05:53 AM:

Not only does the intended projection speed differ between silent films, but some purists believe that the speed should be varied between an individual shot or scene. When Kevin Brownlow presents roadshow screenings of his Napoléon restoration he frequently makes ongoing speed adjustments. A digital system which not only encoded, say, two-channel sound outside the full-gate frame area (i.e. Dolby or SDDS), but also contained speed data which was then used by the processor to send signals to a variac could potentially automate the whole process.

This would potentially make silent films presented reasonably well available to a much wider audience than is able to see them at present. The leading silent film pianists, e.g. Neil Brand or Phil Carli, are amazing to hear but seriously expensive to hire. There are also not that many of them (in a recent radio interview, Brand estimated that there were probably around 50 people who made their living out of accompanying silent films in the world). But as Karen points out, we are talking special venues with an inherently limited audience, so whether it would be economic to develop such a system I really don't know.

As for running soundtracks 'wild', this was done at the NFT when they ran a Buster Keaton season in the early '90s. The sound came from a VHS hi-fi track. A monitor displaying the video was placed next to one displaying a CCTV image of the cinema screen in the box, and an operator continually adjusted the projector speed to match the video. It worked very well, but was obviously a bit labour intensive.

[ 01-23-2003, 11:14 AM: Message edited by: Leo Enticknap ]

Posted by Joe Beres (Member # 678) on 01-23-2003, 09:29 AM:

William Hooper said:

quote:
For the love of Mike, please don't run them at 16 or 18 fps. It was even expected at the time that in projection speed would be slightly faster than taking speed, & the slight speed increase was part of the "art" of movies.

This brings up an excellent point. The academics I have worked with in the past have always insisted upon a particular frame rate, determined by the period, and region, in which the film was created, i.e. a film from the US made in 1910 would be shown at 16fps, period. More often than not though, as William stated, the films were projected faster than they were shot. I showed Fritz Lang's Die Nibelungen a couple years ago, and a pianist came in from Germany. (He reworked the original orchestral score for piano, it was absolutely amazing) He had actually worked with Murnau and Lang in the past and talked with each of them about projection speeds of their silent films, and neither took any issue with their films having been projected faster than they were shot. Rather than try to mimic the speed in which a film was shot, shouldn't we try to mimic the way the film was initially presented? That would make much more sense to me, both as a projectionist and as an audience member. Despite what many academics think, to me, it would seem much more "historically accurate" to present a film at the frame rate it would have initially been shown at.

As I mentioned in an earlier post, it is wonderful to see a silent film accompanied live, but as Leo said, it is fairly rare (outside of Pordenone, Italy, perhaps) to be able to see a film with some of the best accompaniment. After seeing Die Nibelungen, I had a much better sense of what the possibilities could be. I don't mean to get down on the musicians I have worked with, but the best are pretty tough to beat. Sometimes, it would be great to hear the scores, improvised or otherwise, that those individuals and orchestras have played.

(Leo, do you know if that radio interview is available on the 'net anywhere?)

Posted by Stephen Furley (Member # 1277) on 01-23-2003, 10:14 AM:

Leo wrote:

quote:

As for running soundtracks 'wild', this was done at the NFT when they ran a Buster Keaton season in the early '90s. The sound came from a VHS hi-fi track. A monitor displaying the video was placed next to one displaying a CCTV image of the cinema screen in the box, and an operator continually adjusted the projector speed to match the video. It worked very well, but was obviously a bit labour intensive.

Leo, I have twice seen a very good print of 'The General' at the NFT, do you know if this was presented in this way?

This was a remarkable print, the quality was quite outstanding.

Posted by Leo Enticknap (Member # 534) on 01-23-2003, 11:00 AM:

Joe: Neil Brand was interviewed on BBC Radio Four's Today programme as part of the NFT's 50th anniversary celebrations towards the end of last year. You could try going to the programme's website and seeing if you can search under 'Neil Brand' for the item.

Stephen: if you heard the Carl Davis soundtrack and it was around 10 years ago, then probably yes. If they were doing that now I imagine that they'd use a DVD, but the method of synchronisation is probably the same.

Posted by Joe Beres (Member # 678) on 01-23-2003, 11:18 AM:

Thanks, Leo. Here's a link to the page with the Neil Brand Interview, in case anyone else is interested:
Neil Brand on Radio 4's Today

Posted by Michael Schaffer (Member # 1204) on 01-23-2003, 02:21 PM:

quote:
He had actually worked with Murnau and Lang in the past and talked with each of them about projection speeds of their silent films

How old is this guy?

Posted by Joe Beres (Member # 678) on 01-23-2003, 02:28 PM:

Michael, now that you say that, I realize that my memory must be failing me. Lang was definitely mentioned, but Murnau would have been an impossibility, as he died in 1931. The pianist was probably in his mid-sixties, and I cannot remember his name for the life of me.

Posted by Michael Schaffer (Member # 1204) on 01-23-2003, 08:00 PM:

The whole idea could maybe work better the other way around. Say if you had a multichannel mpeg2 audio encoded playback source, then you could use one or two channels to drive the projector at variable speeds. There could be analog steering signals which get encoded as audio and then used to regulate the projector speed.

Posted by William Hooper (Member # 71) on 01-24-2003, 01:47 AM:

Scott Norwood mentioned that he will be working a sci-fi festival at which they'll be running the Kino 35mm print of the recent Metropolis restoration. That film runs at 24fps, is *recommended* to run at 24fps. Additionally, the Kino print apparently has an orchestral accompaniment on the soundtrack for theaters not using live accompaniment. So apparently it's not full-frame (I can't imagine Kino sending out the print with no optical track when it would be most likely to play many older theaters).

Maybe Scott can check to see if it's 1:1.37 (meaning it's lost information at the top or bottom) or windowboxed.

Joe:
University folks run on power differential, in that they must be unquestioned in academia to be influential with their peers or students. They can never, after making a blanket statement, back down. They make a lot up, & a lot is based on an observation of only 1 instance.

So you'd need to arm yourself with more authoritative sources than academics who just dabble in silent movies.
Print the stuff from Brownlow on David Pierce's site.
Also, Brownlow's book "The Parades' Gone By", which is a survey of the silent film era, mentions how when sound movies came in, everyone noted that things slowed down. It was not due to just restricted camera movement & tortured elocution for sound, but because the films in the silent era were projected faster than they were taken. That's noted in Brownlow's book.
Also print the stuff from SMPE from David Pierce's site. You will note that its RP for projection speed is FASTER than the RP for taking speed.

In the silent era, it was expected by the filmmakers & by the audience that the projection speed would be faster than the taking speed. By what degree was up to the operator, working from notes furnished by the filmmaker when available. Otherwise, they ran it by eye: zippy chases & comedy, but not unbelievably fast. Guidance for operators as to speed when running films in the silent era can be found in the book Motion Picture Handbook for Managers and Operators by F. H. Richardson. It sometimes shows up on eBay.

You're the projectionist, you know your craft, you do the actual work to find out how something works, the folks who dabble have no stake & don't work in a world where something either works or it doesn't & thus they can just argue about whether something is right.

By the way, another source for this silent/sound speed confusion seems to be silent/sound speed switches on 16mm projectors. Those switches apply only to 16mm films, not "silent movies". "Silent" meant amateur format, home movie cameras, they had no sound & ran at 18 fps. But you could buy 16mm prints of theatrical movies, which are of course 24 fps so you'd have to switch the 16mm projector speed. Silent: 16mm home movies. Sound: 16mm print downs of 24 fps 35mm movies.

Posted by Leo Enticknap (Member # 534) on 01-24-2003, 04:42 AM:

quote:
University folks run on power differential, in that they must be unquestioned in academia to be influential with their peers or students. They can never, after making a blanket statement, back down. They make a lot up, & a lot is based on an observation of only 1 instance.

Some university folks are guilty of this, but I think it's a little bit unfair as a generalisation. Far from defending dubious blanket statements, most academics working in non-science subjects try to avoid making them at all, because they know that blanket statements are hostages to fortune. That having been said, a little knowledge can certainly turn out to be a dangerous thing when humanities academics start trying to write about film technology, and I guess we've all seen instances of this.

Posted by Scott Norwood (Member # 30) on 01-24-2003, 06:00 AM:

William--the new Metropolis prints are full-frame Academy with a Dolby SR track containing the musical score. We had live organ accompaniment, though.

Posted by Michael Schaffer (Member # 1204) on 01-24-2003, 11:15 AM:

Why does nobody respond to my posts? I tried to find some answers to the original topic! :weep:

Posted by Michael Schaffer (Member # 1204) on 01-25-2003, 12:49 PM:

It has been said here that early films were filmed at 16fps, but projected at different speeds. But, since the cameras were hand-cranked, how can you say they were filmed at this or that framerate? When were motor-driven cameras and projectors introduced? They must have had a defined framerate then.

Posted by David Graham Rose (Member # 1450) on 01-25-2003, 02:10 PM:

I believe the upper and lower limits of linear speed for Dolby SRD (35mm) are 21 f.p.s. and 27 f.p.s. respectively

Posted by William Hooper (Member # 71) on 01-27-2003, 01:20 AM:

Leo wrote:

quote:
That having been said, a little knowledge can certainly turn out to be a dangerous thing when humanities academics start trying to write about film technology, and I guess we've all seen instances of this.

Their writing doesn't bother me as much as the instances I've encountered where they've presented themselves as consultants or "free" unbidden advisors in theater projects, getting a spiff on their bio while screwing a theater up for years. After spending a pile on the physical plant, public non-profits can't admit it when they discovered they've gotten advice from people who had NO real clue but were quite assured that they did, nor can the facilities find money for another HUGE outlaw to repair the problems. I will admit that the statement was inaccurate as a blanket statement, but I will assert that it is completely safe as statistically representative.

We've all seen civilians giving weird, erroneous explanations to their friends of how film projection works. Let me tell ya, when a unversity prof gets in there to show folks around, it's an ASTOUNDING flow of spontaneously synthesized information. It's a performance that's fascinating, like people speaking in tongues. And the strangest thing is, that their egos have become so huge from being unchallenged in classrooms & needing to defend their statements at all costs academically, that they really do believe every thing they say...even though they are making it up on the spot!

Michael wrote:

quote:
It has been said here that early films were filmed at 16fps, but projected at different speeds. But, since the cameras were hand-cranked, how can you say they were filmed at this or that framerate? When were motor-driven cameras and projectors introduced? They must have had a defined framerate then.

The framerate was not locked, but dynamic. "Around 16 fps" is a baseline. Cameramen, even if cranking at what they believed to be 16fps for their baseline, would undercrank or overcrank if they felt it was appropriate for the scene. Film speed, before sound required that it be locked, was a film making tool to be used just as variously as exposure & composition.

I think some of the material on Brownlow's site mentions that cameras & projectors had tachometers available. Motorized cameras were available also, but cameramen did not like them because 1) they were bulkier 2) they restricted the tool of varying taking speeds.

The FIRST movie cameras were driven by motors; Edison's cameras built by W. K. L. Dickinson & used for all the Edison pictures. Edison was Mr. Electricity, & of course used motors. Those things were monsters. The Lumieres' small, mobile, handcranked cameras were the revolution. The nifty things also doubled as projectors.

Posted by Leo Enticknap (Member # 534) on 01-27-2003, 01:45 AM:

quote:
The FIRST movie cameras were driven by motors; Edison's cameras built by W. K. L. Dickinson & used for all the Edison pictures. Edison was Mr. Electricity, & of course used motors. Those things were monsters.

Another problem with the Dickson/Edison technology was that it needed a shooting and reproduction speed of 48fps, because the Kinetoscope (i.e. as near as Dickson got to a projector) moved film continuously past a disc shutter, and without an intermittent movement it needed to be that fast to appear as a moving image to the viewer.

Michael: Two possible reasons why 16fps became established as a de facto shooting speed for studio use in the late 00s and teens: firstly it is the minimum needed to project a flicker-free image which appears to be continuously moving (most projectors had three-blade shutters in those days), and secondly, 16fps equalled two revolutions of the cranking handle per second on the Bell and Howell model 2709 studio camera, which was introduced in 1912 and was the most widely used camera mechanism in the American film industry during the teens and 20s. Suggested reading. I suspect William would approve of chapters 2-4, which are a comprehensive demolition job on the way film studies is taught in universities.

Posted by William Hooper (Member # 71) on 01-28-2003, 12:59 AM:

Ramsaye's _Million & One Nights_ states that the 1845 Lumiere camera design was for 16fps as taking speed, & gearing for 2 turns per second because it lent itself to being easily maintained for 16fps. Most other successful cameras evolved from its design (or just lifted features outright).

Posted by John Pytlak (Member # 331) on 01-28-2003, 08:04 AM:

Did you mean the Lumiere camera of 1895 (not 1845)?

Posted by Leo Enticknap (Member # 534) on 01-28-2003, 04:14 PM:

Maybe it was Dickinson who invented the 1845 camera? [Big Grin]

Posted by William Hooper (Member # 71) on 01-29-2003, 01:25 AM:

I thought it just got mentally typo-tangled with the link to the 1840s St. Charles playbill, but when I pulled out the book I found it's a funnier manifestation of recall/typo tangle mush than just that!

From the chapter "Lumiere's Sixteen-A-Second" :

"They worked through the winter of 1894-5 and into the spring. They were not satisfied with the exposure speed of the Edison pictures, which required forty-eight frames of the film a second. They tried reducing the speed to ten images a second and found that too slow. Finally sixteen images a second became their adopted rate. The Lumiere camera was evolved with a mechanism which gave an exposure of eight frames per turn of the crank, requiring two turns per second for normal operation.

"This two turns of the crank per second was also the subject of some experiment aimed at discovering the speed which the operator could most readily maintain with minimum variation. Unsteady cranking would have meant jumpy pictures. The Edison Kinetograph camera avoided fluctuation by using a battery driven electric motor, which made the machine expensive, complicated and immobile. The Lumiere camera was cheap, small simple and light. The Edison Kinetoscope made its first trips from the "Black Maria" on a one horse spring wagon. Louis Lumiere carried his camera in one hand. This obscure little fact was to have considerable bearing on the first few years of screen business.

Two important notes:
1) Ultimately everything you vow you'll never do, like type "it's" for "its" or "Dickinson" for "Dickson", you will some day do anyway.

2) No matter how great an example Jeffry Johnson will constantly provide, pileups will continue to occur from mushy recall & failure to actually haul the book down & quote precisely from it.

Posted by Leo Enticknap (Member # 534) on 01-30-2003, 01:57 AM:

Point taken: obviously the convention of two revolutions=16fps was established by Lumière and then adopted by others. Barry Salt attributes this to Hell & Bowell (see - even I can make typos, too! [Smile]

), which suggests that he picked it up from a secondary source and hadn't seen a Cinématogràphe.

I must re-read the Dickson SMPE Journal piece reprinted in Fielding's Technological History of Motion Pictures and Television - I bet Edison tries to take credit for that, too, even though his films did run at 48fps at the time!