A Short History of the World (as it applies to video editors)

[Archive]

by John Leonard

After some recent discussions with other members of the committee over Jack Swart’s recent award, I have decided to produce this unsolicited, highly biased potted history of the mechanics of video editing as I have seen it in a long and chequered career. No responsibility is accepted for any errors of fact, since we are relying here on an increasingly unreliable personal memory. It is also unashamedly Melbourne-centric, since that’s where I’ve spent most of my working life. If you know better than I do about anything in these articles I would be delighted to hear from you.

Part 1: Early videotape
I joined the industry in 1963 at GTV-9 in Melbourne. Videotape as a production tool was then in its infancy, and editing was to say the least primitive. In fact, videotape recording was really seen more as a means of time-shifting and program distribution than as a creative production tool. At that time ATN-7 in Sydney had Ampex videotape machine serial number 5, and we had serial number 8. Both machines had been bought to allow networking of programs between Sydney and Melbourne, then the only cities in Australia with television.

To those who don’t know, until Sir Frank Packer bought the controlling interest in General Television (Victoria) the network’s links were crossed. Seven Sydney was paired with Nine Melbourne, and Seven Melbourne was paired with Nine Sydney. And there was no direct video connection between any stations. By the time that I joined the takeover had happened, and we also had a coaxial cable connection between Sydney and Melbourne allowing the Nine network stations to talk to each other.

Part of the agreement done with the government to allow the cable to be laid required that the ABC should have regular access to it and the GPO (later Telstra) should also be able to use it for general telephony. This meant that bookings had to be precise and limited, so that in practice we had regular bookings for news and big budget live variety programs, but not much else. Tape was still used as the primary networking medium. Only prestige programs used tape editing.

Manual tape editing
In those days precise videotape editing was only possible if you physically cut and spliced the tape in roughly the same way as audiotape is often edited to this day. The tools of the gun editor were a chinagraph pencil, cotton gloves, iron dust, benzine, aluminised mylar tape, a splicing block and of course a razor blade. Later, an Ampex and RCA both produced a more or less precise videotape splicer.

The methodology for cutting videotape was in principle quite simple. First you identified where the cut was to be made and marked the tape on the backing side with a chinagraph pencil, removed the tape from the heads and placed it backing side down on your jointing block, with the chinagraph mark centrally placed. We all had techniques for identifying the frame on which to cut, and we all kept our own particular method as secret as we could.

My method sounds absolutely horrendous nowadays, and I really don’t know how I was allowed to get away with it. I used to wedge the capstan pinch wheel away from the capstan, put the machine into play and manually wind the tape through the heads to identify the frame. When you consider that a video frame was recorded as a series of strips across the tape, you will understand that at no time using this method did you actually see a locked frame. You saw a 16-line frame fragment repeated vertically across your monitor. A little experience helped enormously and the potential for tape damage was huge!

See it here: http://www.sssm.com/editing/museum/smith/smith.html

Whatever method you used to find the cut point, you then developed an image of the recorded tracks using a suspension of iron dust in benzine. This required a fair degree of skill, even when magnetic developer pens came on the market. If you applied too little of the mixture you couldn’t see the tracks, and if you applied too much you obliterated them. Either way you had to start again, wiping down the tape and redeveloping until you could reliably read the track information.

Where do we cut?
Once you had marked and developed your tape, you then had to work out where you physically had to cut. The developed tape was examined to find the vertical interval – in film terms, the frame line. But you couldn’t just cut there. You then looked across to the bottom edge of the tape to where the control track was recorded. There, close to your vertical interval if you were lucky, you would find a bright spot, which was the edit pulse. If you were unlucky you had picked up the vertical interval halfway through the frame, and had to move to the next vertical interval. Back in those days there was one edit pulse recorded every frame, so I guess that you could regard it as the electronic equivalent of a sprocket hole (16mm, of course).

At this point a steady hand helped, because the next step was to cut the tape. If you were using a jointing block you cut at an angle of 90º 33’ across the tape precisely between the developed tracks and to one side of the edit pulse. For this purpose most jointing blocks had a straight edge built in. Later when the videotape splicer came on the market it had an in-built blade that was engineered to cut at this angle. Because drawing a blade across a 2” wide strip of tape almost inevitably distorted it one of GTV’s in-house engineers,Al Kleeberg, rebuilt ours replacing the blade with a precision guillotine. This dramatically improved both cutting and splicing accuracy. Al was an early example of the innovators that the Australian television industry has regularly produced, but this (as far as I know) was the only editing tool that he worked on.

The tape was then flipped over, cleaned, and spliced using very thin aluminised mylar splicing tape. The ends of the tape were brought into precise alignment and held there while the splice was performed, and your edit was complete! With our splicer it was possible to join the ends without turning the tape over. You next cleaned your tape carefully on the oxide side with a tape-friendly solvent – remember, there were no dropout compensators then, and any dirt or magnetic particles on the tape would most definitely be seen – and put it back on the videotape machine for playback.

Admiring the results
If you were successful, the tape played and looked roughly the way that you had envisaged it. If you had slipped up somewhere you might not have the cut that you expected or worse still the edit could reframe, or play unstably. A reframe is when the tape goes through a playback relock which on 2” tape could take anything up to eight seconds. During this time the videotape would be totally unviewable.

This all seems like a system doomed to failure, but in fact a surprising number of programs were successfully created in this way. Variety programs and sporting programs predominated, since such drama as was produced here was routinely shot and cut on film. It was also too slow a process for news, and portable video recorders in any case were well in the future.

An early big-event program produced at GTV-9 was the BP Super Show, which produced specials with Louis Armstrong, Liberace, Bob Crosby and his Bobcats and others too many to itemise. But one BP Super Show in particular sticks in my mind. The Borovansky ballet had recorded a special which the videotape editor Peter Dare sweated blood over. The end result was a masterpiece. Because of GTV’s heavy production load it was decided that TCN would do the network dubs, and the master cut tape was very carefully wrapped and shipped by air to Sydney.

What got off the plane at the other end was a beautifully spliced perfectly blank tape.

Fortunately GTV had a policy of dual recording all master material, and we had copies of all the ballet originals. With two and a half days to go, Peter set out to recreate the master. It made it to air. At that time American tape editors earned a $2.50 loading per splice performed, which was big money. Peter received his normal shift payment as a videotape supervisor.

And yes, I still have that pair of cotton gloves.

- John Leonard.

The next article in this series discusses the beginnings of electronic editing.

[Archive]

by John Leonard

The story so far: In last month's instalment of my biased and personal history of videotape editing I taught you all how to splice 2" videotape. However, splicing videotape can make it useless for further recording and in those early days videotape was very expensive. Some non-destructive means of editing creatively would obviously be better.

Part II - Electronic editing

In the early 1960's Ampex introduced an electronic editor for 2" quadruplex videotape. This was a manual device - you pressed the record button while playing back the edit tape, and 16 frames later the machine went into record. Similarly when you pressed stop, there was a 16-frame delay before the VTR stopped. Why 16 frames? That was the distance between the erase head and the video record head.

Despite this, a good editor could place his or her edit within a 4 to 6 frame window, and big production numbers using electronic editing were in fact done this way, usually built "live." The first sequence would be set up and recorded to audio playback from the edit master. A new sequence would then be set up; the edit would be performed and that sequence recorded, and so on until the production was complete. Any mistakes made meant that re-edits were deliberately done earlier than the edit they replaced to avoid the risk of flash-frames from the over-recorded bad take.

Unfortunately, for precision, manual electronic editing just couldn't cut it.

More precision! More precision!
There were various work-arounds adopted by different organisations to circumvent these accuracy issues. A technique that I used for musical productions was to lay down a master track with, ahead of the leader, a short piece of music. This tape was then dubbed and alternate shots were insert edited onto the original and copy to make an A and B chequerboard pair. The two tapes were then played from ahead of the first piece of music and manually pulled into sync while listening to that music. We then cut between the two tapes on a vision mixer to make the final master on a third tape. It was time consuming but we were able to overcome to a degree the accuracy at the cost of a further tape generation.

The alternative, splicing a tape for musical sequences, was very difficult and if a lot of cuts were required there was a real risk of losing sync. This could happen because after all edits were done it was advisable to restripe the audio, and if your edits were all on the tight or loose side you could end up frames out. Manual splicing was used for comedy productions where timing precision was critical - Rohan and Martin's Laugh In would have been impossible without it - or for joining longer sequences in other types of production.

What the industry really needed was a precise means of controlling the electronic editor.

Editec - or, can I preview that cut again?
In 1963 Ampex stunned the television world by demonstrating single frame animation recorded directly onto videotape. At about that time I can remember seeing on a blackboard at Swinburne Art school (later to be Swinburne Film and TV school) a comparison chart between film and videotape. And there, along with issues of portability and resolution was the statement - videotape is unsuitable for animation: it is impossible to record single frames. (I think that Jill Bilcock - at that time Jill Stephenson - may have been in that class. Ted Mason, later the editor of Breaker Morant amongst other films, also probably was.)

Ampex had broken this rule by building a frame-accurate edit controller they called Editec, which could be retrofitted to any Ampex videotape machine with an electronic editor. It permitted edits as short as one frame to be recorded, and won them an Emmy for technical innovation.

Operationally, it was a very easy system to use. To set up an edit the master tape was played and where the edit was required a cue button was pressed. The tape was then played back and the edit point was previewed and its position adjusted as necessary by means of a dial calibrated in single-frame increments. When the edit point was acceptable, the tape was cued back ahead of the proposed edit and put into record. The videotape machine would then edit precisely where the editor (human) required it.

Ampex had done this magic using an under-used track on 2" videotape - the cue track. On an Editec equipped VTR pressing a cue button placed a precise burst of 1 kHz tone onto that track. This cue pulse was subsequently played back by a new head specially added for the purpose to the video erase stack, upstream from where it had been recorded by 33 frames. This made it possible to adjust record in and out times by plus or minus 16 frames. RCA subsequently produced a similar device for their VTRs.

It was possible for Editec to trigger external events using a second different type of cue pulse (here's where GPI events all started, folks!) While potentially this meant that a playback VTR could be started by a pulse placed ahead of the edit cue on the record machine, in practice very few people did.

Practical precision
Various approaches to obtain total edit precision were used. Cuing at this time was exclusively manual, and was limited in accuracy by the accuracy of the tape counter. These were mechanical and had a resolution of tenths of a second, and were driven by a drum pressed firmly against the tape. We would cue to the edit point, reset the counter to zero, and roll back by ten seconds. Then, with one operator on the record machine and another on the playback we would count to three and hopefully both press play at the same time. All around the country cries of "One, two, three, Roll!" echoed through videotape production facilities.

(At GTV a side effect of this technique was that on edit machines counter failure rate due to wear and tear was very high. Fortunately Peter Chaffey, one of the two supervising technicians in videotape at the time, was able to source a local supplier!)

If greater precision was required once both tapes were cued they were marked on the back with a chinagraph pencil and the edit was previewed. Any adjustment needed was made by moving the chinagraph mark by a calibrated amount and trying again. Although primitive, this meant that suddenly precise electronic editing was a reality with record accuracy to the frame and playback accuracy to around plus or minus 2 frames if you were good.

At GTV-9 we achieved greater roll-up reliability by building what we called the buggery box. This was a device that took the remote control play line from each tape machine and allowed a selected machine to be put into play mode by the same button that started the Editec equipped machine. At this point we had repeatable frame accurate edit capability. I know that others built similar devices.

Changes to drama production
We were making Hunter, a cold war spy drama for Crawford Productions. The technique used was to shoot studio sequences multi-cam onto predefined positions on our master insert tapes, and later record location film scenes from telecine onto the same tapes. The edit master was then produced by rolling individual scenes "live" from the insert tapes and switching them manually onto the master. If we missed a cue we would go back to the last commercial break and start again. All audio spot effects and music were rolled in from audio cart and if any of those failed we also had to restart. On program integration day the two tape operators rostered began at 4 am and ended totally exhausted at around 3 pm.

As you can imagine space wastage on the original tapes was considerable, since we had to allow room on the tapes for overruns and alternative takes. Also if the predefined window limit on an insert tape was reached there would be the agony of deciding what retakes we could discard to make things fit.

Editec changed all that, and in the process altered the way that TV drama was made in Australia. Now it was possible to go back to the cut prior to a fluff and pick up the scene from there - we could now fix it in post! Nor was it necessary to predefine positions on insert tapes for scenes. We recorded them as they came, and - we fixed it in post! If we needed stunts in studio, we shot them single camera, and - we fixed it in post! If we needed a cutaway or POV we shot it, and - we fixed it in post! It was a world first for television drama production and variations of the technique are still in use.

How we taught the Poms to edit
The BBC at that time was producing drama on videotape, but doing it as if it were a live production. They would rehearse intensively (a good thing - modern directors please note), then go for it. If they blew it and couldn't recover without being obvious, they did it all again (a bad thing). Somehow, they had heard that there was this little television station out in the colonies that had a radical new way of doing drama, and they flew out some engineers to find out what was happening. They took Aussie methods and adapted them for their needs, and in that way we changed how the Brits did things - something that was to become a bit of a habit in the later Grundy years.

John Leonard.

The next article in this series talks about early timecode based edit systems.

[Matthew]

The story so far: Last month’s article took us through the introduction of electronic tape editing up to early edit control systems. In the process we introduced GPI events and revolutionised the production of drama in this country. Of course, Crawford’s had something to do with that also. So without further ado in this month’s instalment of my personal history of videotape editing we’ll simplify our techniques.

Part III – Accuracy and repeatability

By the time that Division 4 came around at GTV-9 we had bought our first Ampex AVR-1 videotape machine. They had a guaranteed lock up time to stable playback of under 200 milliseconds, or less than five frames. They also had a frame-accurate counter that ran directly off the capstan – another first. Clearly this machine was magic but it was very expensive – around $250,000 fully optioned – so it was the only AVR-1 we bought! We weren’t alone there. Because of the cost only 500 were built by Ampex.
Jim Higgins, one of GTV’s two tape supervisors, had the idea of using the advanced edit cue read on the Editec machine to put the AVR-1 into playback. What we could then do was this: once the edit cue had been placed, the start point of the next playback segment was found and the tape stopped. The capstan was then manually wound back by one second and ten frames. This was 33 frames plus a correction for reaction and run-up times. From then on integration became simple, and the average episode now took between four and five hours, down from eleven hours. It was the fast, efficient, and precise method that we used until timecode editing came along.

(Jim was responsible for a lot of the videotape post-production methodology used at GTV. He should have been an editor, regarded himself as a technician, and subsequently became an engineer. In contrast, my family thought that I was going to be an engineer, I started work as a technician, and became an editor.)

Sound sync and editing

Timecode had been around for a while before it was used for video editing. The first paper that I read on timecode related to uses of a new time signature data format for synchronising audio to video. I can’t remember the company that was doing this, but I think that they were German. I also can’t remember whether they had a product or if it was just a think piece. What I can remember is that the proposal was published around the time that Editec surfaced. So but for an accident of history we might have gone to timecode editing about six years or so before we did.

For our part we had dealt with problems of sound sync by a variety of methods. Two that I wish to single out were either custom developments at GTV or built to our brief. Why those two? Because they’re the ones that I had personal experience with.

Precise sound lock

The first was a means of locking Nagra recorders to video. Bob Henson, (no relation to Jim – not a muppet in sight) was a particularly bright engineer employed by Nine. He designed and built a device that produced an external lock reference for a Nagra from television frame pulses. We could now record locked to video, and playback the same way. We preferred not to use crystal lock, because back in the days of black and white sync pulse generators were less precise than they are now and the potential for drift was actually quite high. This tool solved that problem and was used not simply for location recording but also for location playback of prerecorded tracks.

The second was a tool that we used for the big event productions that GTV specialised in at that time. It allowed an eight-track audiotape machine to be precisely locked to video frame pulses. Now when we sent an OB van to cover a live concert we sent at least two – one was the audio van with its own dedicated eight-track recorder. This device was designed and built by Graham Thirkell, of Sontron fame, and to say that it revolutionised big event production and post-production for us is not overstating things.
Note that neither of these systems had any means of automatic synchronisation. They were tools that maintained lock after synchronisation had been achieved.

Timecode arrives

The first timecode edit system that I can recall seeing was the Ampex RA-4000 in 1970. It was a very sophisticated device well ahead of its time, but it suffered the quite serious disadvantages of expense and an inability to control anything other than AVR-1s. I don’t think any were sold in Australia despite the aggressive marketing bundles that Ampex put together. Apart from the problems outlined above they also used their own unique timecode format unreadable by other systems, despite the fact that timecode editing was already becoming established internationally.
(A little history here: timecode was actually invented as a means of synchronising data loggers for the US aerospace industry and for other scientific use. The earliest use for television was by EECO in 1966 in the US, using a data format modified from that early scientific code.)
The first timecode edit system that I used was made by EECO. In the early days of timecode these systems had become so ubiquitous that timecode was often called EECO code in the same way that another generation referred to aspirin as Aspro. However, neither the EECO 900 nor the RA-4000 was a computer edit system, as we know it today. They were dedicated digital systems built with discreet components – there was not a CPU chip in sight.

At about the same time as we moved to EECO, the Melbourne branch of Videotape Corp (remember them?) and Channel O (later channel Ten) also bought systems. I understand that EECO had similar penetration throughout Australia, so as you can see they got off to a good start. Oh, and colour television production was also beginning.

This means change (again)
Because these systems were separate stand-alone devices they were not usually built into videotape machines in this country. At GTV we solved the problem by constructing our own dedicated post-production facility attached to the tape department, but separate from it. We also, with money from our colour upgrade budget, bought three RCA TR-61 recorders that were dedicated to post and a Grass Valley 1600 vision mixer. Division 4 from that time on was post-produced in that department, by myself until I left to join AAV Australia, and subsequently by Gerry Stack. My recollection is that Jim Higgins was also one of the EECO drama editors but he assures me that he was not.

Another advanced feature that we introduced into the post-production chain at GTV-9 was a Fernseh colorgrade system. This allowed us to grade already recorded material – unheard of at the time – as we edited our master. As a result at various times we had other teams visit to find how we were achieving our consistently high quality output. They looked at studio recording, they looked at telecine transfers, at editing as well, but the one thing that nobody seemed to pick up on was the Fernseh colorgrade system. The only other company that regularly used one that I can recall was AAV when they started – but they used it purely to get around the limitations of their Rank Cintel telecine colorgrade system.
The staff driving the colorgrade system was dedicated to that task and rostered from amongst the technical directors. At various times Mike Bannister, Colin Johnson, Peter Fisher, and Alan Leslie all sat in the hot seat. Anyone else that I’ve missed I whole-heartedly apologise to.
The final post enhancement was to provide a means of timecode synchronising audio to the EECO system. This technique was only ever used on big-event productions. For Division 4 Chris Eichler was still manually rolling in spot effects and music from carts supplied by Crawford’s sound editors. This and the on-line video editing was unsupervised by Crawford’s – the directors would send in their marked up scripts and Crawford’s would receive finished dubs back on Umatic for approval.
In line with Nine network policy none of the on-line post team at GTV were ever credited at this time.

John Leonard.

The next article in this series introduces the computer to editing.

[Matthew]

A short History of the World as it applies to Video Editors

by John Leonard

The story so far: Last month I talked about how the introduction of timecode improved precision and flexibility in video post. But just when we thought everything was damn near perfect it all changed again. With it my working life along with every other video editor in the country was to change radically.

Part IV – Computers? Who needs computers?

In 1971 (two years before Australia officially converted to colour TV) I read about a ridiculously complicated device that a small company called CMX had invented for video editing. Built around cutting edge technology, it recorded video in poor quality black and white onto packs of rhodium plated 14” aluminium alloy computer disks. Marketed as the CMX-600, it was capable of storing and editing up to 27 minutes of NTSC originals. Tying it all together was a Digital Equipment PDP-11 minicomputer.

(A little history here: CMX was a company set up by CBS and Memorex to develop editing systems based on pioneering work done by CBS Labs. The inclusion of Memorex bought with it their expertise in magnetic media for data recording. Later joint development work done by CBS Labs and Sony in the 1980s was to lead to tape-based non-linear systems like Montage.)

The world was not ready for this machine. They were just too expensive and for the average post house there was simply no way that you could recover your outlay in a reasonable time. Only six were ever built, but I saw one of the then two remaining machines still in operation in San Francisco in 1984. At that time it was being used exclusively as a playback source for audio sweetening. It had been modified to store an hour of program material.

The specification of the machine makes interesting reading. It recorded in drop-field mode, just like modern non-linear off-line systems; it used two monitors for display purposes, and was totally random access. A nicety was that the user interface was drag-and-drop, using a light pen, but recording quality was rather inferior to Avid’s AVR2 mode. If anyone asks you when non-linear video editing was first produced, now you know – there’s nothing new under the sun.

Also supplied was a companion piece of software called the CMX-200, designed to conform on-line masters using the off-line EDL generated by the CMX-600 on papertape. The CMX-200 was in no way an on-line editor – while it was possible to manually edit the EDL once it was loaded it wasn’t a simple process. The EDL format created by CMX in 1971 is still used today with only slight modifications.

The birth of AAV

In 1972 Ranald Macdonald at The Age newspaper in Melbourne in a visionary move decided to invest in video and audio production as part of his desire to move his newspaper into electronic media. He started by buying a collection of small companies, two of the key ones being Armstrong Audio in South Melbourne and Video Tape Centre, a small video company based in Moorabbin, a suburb of Melbourne. He also contracted an American adviser, C. Robert (“Bob”) Fine, known amongst other things for the setting up of the Mercury recording label in the US, to tie it all together.
The new company was called Armstrong Audio Video, with Bob installed as the CEO. The Video Tape Centre name was quietly dropped, since it was regarded as too similar to Videotape Corporation and the mnemonic abbreviation for both, VTC, was quite frankly an embarrassment. Premises in Bank Street were renovated and the combined entities were moved there, and as part of the setting up process The Age injected a massive amount of cash.

I was still at GTV at this stage. Our chief engineer Tom O’Donohue left to join AAV, and was followed by Jim and Elaine Higgins, from the videotape department. I managed to wangle myself a position as well. Other ex-channel 9 staff that I found already there were Andy Chapman, like myself ex-videotape, Richard Berriman who had come from CCUs, and Ted Gregory, who in the past had been a director at GTV.

What do we do now?

Bob as head of a new cashed-up company looked around for smart ways to spend the money. He was shown a new product built by CMX as a means of getting a quick return on the massive R & D effort that they had put into the CMX-600. This new product was called the CMX-300, and was the world’s first computer controlled on-line videotape edit system. At the time that he saw it, it was NTSC only.

He wanted it.

Armstrong Audio Video bought the CMX-300 on the understanding that it would be adapted for PAL. All modifications were ostensibly done in the US, and the system was shipped to Australia. An American programmer, David Morgenstern, was supplied by CMX for shakedown and debugging purposes. It was intended that he be in Melbourne for around three weeks, but six months later he returned home. The PAL conversion wasn’t quite as ready as we had all hoped!

Bill O’Donnell had been poached from Videotape Corporation’s Sydney operation, and was installed as senior editor. He worked alongside Dave during the debug, and became our first resident CMX software expert. Chris Littlehales, an ex-ABC video technician, later took over that role having taught himself programming on the job.

Launch time

Armstrong Audio Video officially opened its doors to paying customers on June 3, 1973. A big media event had been planned with industry guests flown in from all over the country. I can remember amongst other things telling one of the guests at the opening in great detail why the Philips domestic video recorder was technically inferior to the Sony Umatic, only to discover that he was the sales and marketing manager for Philips Australia. Since Philips had given us great support during the setup phase, he was rather less than amused.

A more positive claim that I can make for that first day was developing in detail a post system with film editor John Scott that was identical to the method subsequently used by Armstrong Audio Video for finishing film commercials. John didn’t really take the theory anywhere, but Mike Reed most certainly did. It would be nice to be able think that I was responsible for the system that helped establish the company, but I can’t claim to have invented the method, since it was such an obvious way to do it. In any case I think that Geoff Baxter or Bill O’Donnell have at least equal or even stronger claims.

The great method

The technique to be used was quite simple. A film commercial would be cut in the normal way on a flatbed from workprint. After client approval the film editor would bring into telecine a roll of spliced neg or one-light workprint consisting of all of the shots used in the commercial, camera stop to camera stop, with a standard leader spliced at the head. In those days there was still a fair bit of nervousness about neg being handled anywhere other than the lab, and there was also quite heated discussion about quality of neg vs. pos transfers.

In telecine the film was graded shot by shot from the roll and transferred to tape. If dissolves or other multi source effects were required often the telecine transfer would be dual recorded; failing that the 2” transfer would be dubbed with timecode.
The tape or tapes would arrive in the CMX edit suite and the timecode address of the two frame on the leader would be identified – a later improvement was to transfer the film with a predetermined two frame. The film editor would have armed themselves with a negmatch list of his or her cut in feet and frames referenced to the two frame and a calculator. Seated beside the on-line editor they would calculate the start point of each shot in sequence in frames.

The on-line editor would trim the two-frame timecode by the calculated start time to obtain the timecode address of each shot, and record it into position on the master tape for the required duration. At the end of the process we would have a videotape release master without having gone to the lab for a film finish. In later years Chris Littlehales rewrote the CMX software to allow us to directly input 16mm and 35mm feet and frames. This improved accuracy and took a lot of pressure off the film editor.

Yes, there were shortcomings: for instance on 2” videotape and also early 1” it was impossible to edit to every frame. You had a 50% chance of being able to edit to the required frame due to the technical constraints of the PAL television system. Routinely about half the edits in any commercial were offset by a frame from where they were supposed to be. Film editors who knew that this was happening (not all did) weren’t happy but either chose to ignore it or would attempt to pre-correct their telecine roll in advance. The big advantages of cost and flexibility overrode that limitation.

John