Timing Motor Ramp-Ups for Accurate Leader Start Frame

[Ryan Gallagher]

I contemplated re-testing the motor start timings for proper leader timing (Academy and SMPTE) over in this thread:
https://www.film-tech.com/vbb/forum/...0078#post50078

But as this is a general booth knowledge / best practice topic, thought I would share my findings in a new thread that is easier to find for anyone else looking to do the same or new to film projection.

Background:
We have two Century JJ's 35/70mm Units upgraded with VFD motor starts, 24fps, 30fps, and Variable speed controllers. For the purpose of this write up I'm focusing on 24fps 35mm. Booth practices as I started here were to frame on the "8"s for Academy Leaders, and honestly I'm not sure there was a consistent practice with SMPTE Universal leaders. We didn't have many change overs with Academy that revealed black on the next reel etc, but there were some, and our other operator has a tendency to leave an intentional "breath" between the c/o cue and mashing the buttons for this reason, which might relate to start frame timing issues.

Goals:
Learn our actual motor ramp up timing, time from target number to first frame, etc.
Learn if "8"s is in fact correct considering for Academy leaders in our booth.
Learn what would be ideal framing number for SMPTE Universal Leaders.

Prerequisite Knowledge:
https://en.wikipedia.org/wiki/Film_leader
https://www.sprocketschool.org/wiki/Cues

Excellent but optional read:
https://muse.jhu.edu/pub/23/article/657445/pdf

Academy & SMPTE Projection Leaders:
These are the ones with the numbers in a white frame with black between them, from 12 stopping at 3, the numbers are in feet (16 frames each).
If correct and nothing is missing, after the 3 foot mark, there will be 47 frames (nearly 2 seconds) of black before first frame.
Academy cue marks, if nothing is missing, should have 172 frames of picture between them (7 seconds + 4 frames, or 10ft + 12 frames).

SMPTE Universal Leaders:
Theseare the ones with the numbers in the sweeping radar or clock face, no black between them, ideally starting from 8 stopping at 2, the numbers are in seconds (24 frames each).
Like Academy, if correct and nothing is missing, after the 2 second mark, there will be 47 frames of black before the first frame.
Universal cue marks, if nothing is missing, should have 168 frames of picture between them (7 seconds, or 10ft + 8 frames).

Easy Method:

Easy Test, use a stopwatch on your phone and watch the leaders framed as you would normally with prior booth practices. For example if you frame Academy on 8, start the motor and the stopwatch together, but open the dowser to watch the second half of the leader on screen, attempt to stop the timer when you see the first frame of picture. This is best done with a reel other than R1 that has a bright picture for the 1st frame.

Warning: I'm not assuming any particular skill level here, so it goes without saying, DO NOT open the dowser and/or change over mechanism before the motor is up to speed!!

Your stopwatch time is how many seconds of motor time elapsed from motor start to first picture frame plus a little reaction time. You can compare this time to the known footage time from your start frame. 8 feet on Academy is 144 (8 x 16) frames before picture, 5.33 (144 / 24) seconds before picture. If your stopwatch time is not in the ballpark of the approx. 7 seconds between motor and c/o cues, you might want to consider a different "start" frame and repeat the test.

A More Intentional Method:

For my testing I made a short trailer reel with a full Academy leader spliced onto it, I added additional numbers every foot above 12 up to 16 with tape, and threw on some white leader for threading at those higher numbers if desired. I made the first frame of picture correspond to the ratings card, which is easy to spot and fills the frame. This is a now a nice bit of kit to keep around the booth as built, so you can repeat testing in a more rapid fire manner and test out different framing numbers.

To translate from Academy to SMPTE Universal, one simply has to take the deduced "best" framing number from the Academy testing, multiply by 16 for total frames, then divide by 24 to get the value in seconds, which correspond to the Universal leader numbers. If you are in between numbers pick the lower of the two to avoid showing black leader.

An even better test film might be, instead of rating card or picture at start frame, include a section of a Universal countdown with a known start number. Put some accurate cue marks on another short reel and you can test realistic changeovers, but the extra countdown where the picture would normally be will reveal how much footage you did not show in the changeover due to slightly miss-timed start-frame or reveal operator response times.

Two Projectors:

In a perfect world your motor ramp up times are matching, but don't assume they are, repeat testing with the other projector just to confirm you are within a workable range of sameness. If you find you are not, it would be better to service your VFDs/Motors than to create different threading practices per projector.

When things are not standard:

How to compensate for leaders missing footage, wrong or missing cues, or not enough picture after the c/o cue is a bit outside the scope of this write-up, but learning how it "should" work with correct leaders and cues is a foundation to learning the various method for being flexible and still having a good show when a print diverges from those expectations.

My findings for my booth:

Our prior practice of framing Academy 8 timed out at 6.95 seconds before picture on my stopwatch, which allowing for operator reaction times, is probably adequate/correct.

Framing on 7 might be a slight improvement because it would reduce our reliance on frames after the c/o, if you are too "Johnny on the Spot" right now you risk showing black on the next reel, but debatable.

So for our Universal leaders, we should frame at the equivalent seconds mark to 8 feet, (5.33s), and erroring on a lower whole number, so framing on 5 seconds seems best for Universal for our ramp up timing.

Corrections and other methods encouraged!

[Ryan Gallagher]

An afterthought. I suppose if you wanted even more rapidly repeatable with less rewind hassle... it would be possible to build your leader test strip as a suitably sized loop, might be a touch long and unwieldy unless you limit it to the countdown and nix the non-needed parts of the leader.

[Randy Stankey]

I don't know if there is a standard method but I can tell you how I might do it:

Get a roll of old film with the leader you want to use spliced on. Thread up just like normal, with your "start" frame in the gate. Start the projector just as if you were doing a real show. When your changeover point comes, take a Sharpie marker and touch it to the film (through the aperture) and make a mark on the film.

Do this three times. After the third time, take your film to the bench and see where your marks landed. Find the average of those three marks and that should be pretty close to your actual starting point.

If you don't like reaching into a running projector with a marker in your hand, you can do your test starting at any reasonable spot in the film path. For instance, right at the point where the film enters the top of the projector. Just make one mark for the start point and another for the change point.

[Steve Guttag]

The acceptable range is between 3-seconds and 5-seconds from 0 to 24fps. Most American film projectors have a ramp up of around 3-3.5 seconds. It depends on the motor, the amount of drag/friction in the system.

Kinoton went for 5-seconds...which works...barely, for changeovers. For changeover houses, we normally dropped it to 3-seconds...for platter houses, we left it at 5-seconds. And, really, you can make it an even slower ramp up on a platter (what's the rush?).

For changeover, you have two opposing forces...if you don't ramp up fast enough, the sound drum will not be at speed and wow/flutter free at changeover. If you ramp too fast, the take up reel/drive mechanism may have difficulty responding. So, as the motor turns on, the reel stands still until the gears mesh, belt tensions, friction is overcome and clutch stops slipping. You are at the point of risking that film went slightly slack and as the reel comes up to speed will tug film (hard) as the film will now be holding the reel, that is now spinning faster than it needs to) back and breaking the clutch friction. Mind you...this all happens in a a fraction of a second but the film/perforations still feels it with each start.

If you keep it between 3 and 5 seconds, you can normally avoid most of the pitfalls and have the sound at speed in time. This is even more critical for Dolby Digital penthouses. Remember, that film has to be at speed and steady over a second sooner (than a basement reader) so it can be delayed and ready when the sound is changed over. If you look at your Dolby Digital decoder (DA20, CP500 or CP650)...look and see how fast it can go from a "-" to giving you a number that doesn't flash "F" while it is walking its way down to your nominal error number. If you choose to have a longer ramp up...you will also increase that time since it is constantly trying to find the nominal film speed and if you are continuing to change it with the ramp up...it will still be ramping too.

[Dudley Marshall]

A fun thing I did to make sure I was starting at the right number in the countdown was to put some markings on the reel I’m changing over to, and see where it lands. Using a grease pencil some old trailers, I put four X’s, followed by four circles, followed by 4 lines starting at the first frame of the changeover reel. I found when starting at five on the countdown I would land on an X, so I figured that was good. What do you all think? What is the exact frame to frame ratio you’re aiming for in a changeover? In the booth, it is not always easy to tell how smooth it is.

[Ryan Gallagher]

A fun thing I did to make sure I was starting at the right number in the countdown was to put some markings on the reel I’m changing over to, and see where it lands. Using a grease pencil some old trailers, I put four X’s, followed by four circles, followed by 4 lines starting at the first frame of the changeover reel. I found when starting at five on the countdown I would land on an X, so I figured that was good. What do you all think? What is the exact frame to frame ratio you’re aiming for in a changeover? In the booth, it is not always easy to tell how smooth it is.

That was kinda where I was headed with the thought of inserting a snippet of SMPTE Universal leader onto my test reel, because of the numbers and sweeping radar, one could potentially see/measure exactly how far past the 1st frame the changeover took place. But your grease pen blocks of frames with unique marks method could be done with any print as long as you clean up after yourself.

Thanks Randy and Steve for the additional thoughts! I'm trying to imagine how one stabs a piece of moving film with a sharpie blind while staring out a the port glass, but I'm sure it's doable. I like the idea of doing it a bunch and taking an average. Maybe multiple colors or just clean up between each. I'll have to try that as validation of my stopwatch/onscreen method sometime when I'm bored.

I wasn't even considering changing our ramp yet, and all of Steve's input is valid there about the balancing act to keep all the audio drums, flywheels, and take-up reels happy. I've certainly seen plenty of trashed/repaired heads and have always wondered how heads are so prone to getting torn. Steve alludes to one example of how.

We risk getting into the weeds for sure, this isn't something one has to do all the time, ideally after motor or ramp timing changes, perhaps even drive belt changes?, when walking into a new booth that is poorly documented, or so you can be intimately familiar with your start frame ideals so that you are more prepared to improvise correctly when prints are not standard (such as adding cue marks or altering your thread up framing number), or missing leaders entirely and you are tempted to "make" some out of other stock available.

[Steve Guttag]

The number you thread on depends on the ramp-up speed of your system...how much the film has been butchered so you don't have the full amount of film between the last changeover cue and the last frame. It also depends on the leader. Are you using Academy (12-feet) or SMPTE (8-seconds, with the sweeping second hand)? The most popular leader I contended with was Academy and I would usually start at "8 in the gate" for an unknown machine and then work my way out closer to 9 as I got a sense for the ramp up of the machine. With Async motors, you could even have different ramp ups depending on which machine you were on.

At the Uptown, in DC, where I worked for 18-years, I could pretty reliably put "8" at the top of the upper loop and never botch a changeover. They were async motors.

A pair of Kinoton DP75 with Async motors (circa 1993) had such fast start ups that you'd need 9 in the gate to not cut off the first foot of the incoming reel. And, to make it even more exciting...the changeover dowsers opened fast and closed slow. So, you were going to double-image if you didn't time your changeovers perfect. It was a screening room so people would notice if you had issues.

To prove to one person about how accurate I could be on a changeover...I put a slash on the last frame of the outgoing reel and a backslash on the incoming reel (old trailers and with grease pencil). I could reliably put an "X" on screen. The more you know your machines the more accurate you can be.

Of course, Brad ruins it all by using "automations" to cue his projectors for "frame-accurate" changeovers. In the end though, if the audience is aware of what you did, you failed. So, whatever works.

[Ryan Gallagher]

Yeah "8 in the gate" is certainly the industry terminology for Academy leaders. And since that was our practice I had no idea if it was simply due to industry norms being passed down in this booth, or real motor ramp times.

Where one really benefits from knowing motor specific timing is when you are missing footage after the c/o cue mark. Without knowing how accurate your changes are, a typical "fix" would be to put your own motor and c/o cue marks earlier on the print. But perhaps not needed if you know you know what "minimum" footage beyond the c/o you need reaction times, and can be confident about the next reel's picture being in the gate too based on your start frame practices (assuming complete leaders).

[Jim Cassedy]

It IS possible to over-think this sort of thing, ya know. . . .

That being said, whenever I occasionally have a really important show, with
maybe the director or cinematographer in the house, and I want be sure of
doing "frame accurate' changeovers, what I like to do is thread up R1 and
R2 (or any two random reels from whatever I'm showing) and when I see
the motor cue on the outgoing reel, I start the other projector and I manually
open the changeover dowser on the incoming projector, which puts the
countdown leader on screen, over the picture from the 1st projector. So,
if there is any 'black' on screen after after I see the c/o cue, then I need to
thread up my incoming projector a bit shorter. On the other hand, if I see
two overlapping images on screen after the changeover cue, then I need to
give myself a little bit more leader when I thread the incoming reels. Easy!

[Steve Guttag]

That presumes you have the time to run through the reels and rewind. Surly, you wouldn't thread up the film mid-reel and, on top of the handling, scratching...letting the film run up mid-reel. Then, what are you going to do on machines without a manual dowser (that doesn't close the outgoing machine). For example, Norelco AA2, Kinoton...etc. In fact, you worked the Dolby screening room in Potrero. Did their DP75s have original changeover dowsers or were they modified to use something more American?

[Jim Cassedy]

You are correct Steve- - this does take time, and threading/un-threading, or even
stopping the projector mid reel is a definite no-no for me, so I need to run two full
reels. That''s why if I do resort to this, I'll pick the two shortest reels. Fortunately,
99.5% of the machines I've worked with did have manual changeover dowsers, and
you're correct, the Dolby Kinotons did not have a way to do this, but I never ran a
show there that I felt the need. And, as I said up front, this is not something I need
to do very often. I've probably done it twice in the past two years, more to mollify
someone else than to appease my own anxiety. I've been doing this long enough
that I know my machines and I have a good sense of 'timing' . It's just one more
'trick' I've resorted to in my projection life.

[Ryan Gallagher]

I think Jim’s approach is quite a clever visual check, but it does rely on those things mentioned. It is probably a valid technique if you are running a screen test anyway, or if you have house prints handy with correct leaders and time to use them.

It also relies on the first frame of the subsequent reel not being dark or some kind of a cross fade.

We ran Breathless recently, one of its c/o cues is at the end of an iris close effect, and next reel irises open!!! Real easy to see if your timing was off with that one!!

All the print specific variation can be determined during inspection, really just discussing how to go from unknown timing to known motor timing.

But for what it is worth, my original method with a test reel also relied on a manual dowser somewhat, although I was doing it on one projector, not watching for c/o cues, so one could just change early if you did not have a manual option to reveal the leader.

[Ryan Gallagher]

...I start the other projector and I manually
open the changeover dowser on the incoming projector, which puts the
countdown leader on screen, over the picture from the 1st projector....

We have classic "zippers" on our JJs. I've never even considered trying to manually pull one open in addition to the already open one. Is that even possible?