Roger Bunce wondered when it first become possible to do a “Rollback and Mix”.
Roger Bunce
I was looking at old episodes of “Doctor Who” and effects like the disappearance or reappearance of the Tardis. Everything in the early years would have been done ‘As Live’, with minimal editing. But a disappearance can’t be done ‘As Live’. There must be a shot of the set with Tardis, and a shot of the set without Tardis, and a mix between the two. But the two shots can’t have existed at the same time – and I don’t think the mix was done in post production. So, way back in the mid-nineteen-sixties, would it have been possible to do a Rollback and Mix? – or did that sort of thing only become possible later?
Graeme Wall
Aa I understand it, it was done by recording the first shot on a seperate VT and playing it back in as an insert and then mix to the locked off shot. I seem to remember rollback and mix being available from the early 1970s when I was doing my inlay training, so circa 1972.
Bernie Newnham
Yes. You just need to be able to sync the ‘play-in’ VT to the studio, and that was being done in the mid-nineteen-sixties. I remember seeing it being set up on TOTP when it was in G.
You’re recording on 2 machines up to the break. A camera is locked off for the last shot. Then you stop and remove the Tardis. One of the machines rolls back a few seconds – say 20 seconds. The other records. Then the play-in machine plays back in sync, and the vision mixer mixes. The record machine records the result. Then the play-in machine resumes the studio recording. At some point you have to check the single ended recording to make sure its ok before the studio breaks.
Of course, if you’re rich you can book a third machine, but they probably wouldn’t have one anyway because "…it’s being used by sport…".
Chris Woolf
A critical element was the ability of the VT machine to provide truly synchronous playback. That required a basic (analogue) TBC, since the raw output of a quad machine was far too jittery to do any mixing. Amtec, and a little later, Colortec, were introduced to provide a signal with sufficient stability that it could be mixed or wiped, rather than just cut to.
I think the first production machine with this facility was the VR2000 (1964) which also had Editec (1963). Given that the UK had to wait till the 625 line versions to be built (UK UHF also started in 1964) I suspect "rollback and mix" could only have begun here in the late 1960s at best. Digital TBCs with a few lines of memory became available right at the end of the 1960s.
Timecode didn’t come in till after 1968.
When one thinks of the problems involved it was amazing what the 1960s’ engineers came up with.
Albert Barber
I done know if anyone remembers but in Bristol we did a similar Rollback and Mix thing for “Think Again” (Photography) where we had our presenter (Johnny Ball) standing on top of himself several times all done as a rollback and mix, or copy, on two machines whilst talking to himself. It worked pretty well although the quality suffered. I think the first pass was against a background and subsequent passes against CSO. maybe thinking about it may have used a separate background on another camera against a Bluescreen (CSO).
We did a similar thing with CSO and again the copying process degraded the picture quality.
Alec Bray
There was something very similar done on "Nationwide”, with Richard Stilgoe doing something about the "Statutory Right of Entry to your Home" – along the lines of "when the gas man came to call" . It started out with one of him, and as each verse was added, another Stilgoe as a different character joined him. There was a problem for the last "take", with Stilgoe dressed as a Policeman, who was supposed to sit on a chair at the right of frame (camera right). Unfortunately the overlay (as existed at that time) could not let him actually sit in the chair, so he had to sort of crouch.
(Click on the picture below to see larger version:
use your Browser’s BACK button to return to this page)
The full video is here, https://www.youtube.com/watch?v=UgEgBsiiUGk: starts 49 seconds in.
Just look at the fringing or halo effect … was this CSO or old-style luminance overlay … ?
With luminance overlay – or in my day (mid 1960s) ’twas just overlay – the image to be "overlaid" was used to switch the line from the original source to the new source. In black and white days, this was tricky, as you only had a grey scale to use – and it was the grey scale that made up the overlaid picture.
So, to "overlay" a camera (or other source) onto a picture from another camera, Vision Control had a big part to play. Let’s say that the image to be overlaid is on its own against a black cyc or flat. Ok, parts of the image are black too, but the lit black of the subject is not the same as the unlit black of the cyc. Vision Control would do something like this … change the gamma response of the tube (or change the gamma response of the signal from the source, for example, VT) so that the dark parts of the picture are "compressed" and then sit down the black level so that any creases from the cyc or flat would be below black level. Then the Inlay/Overlay operator would set the "cut" level somewhere between black level and the black parts of the image to be overlaid.
As with early CSO, the time taken to switch between the two sources could cause fringing of the inserted picture.
Bernie Newnham
I’d like to show off my best ever trail idea.
James Burke made a series about the nature of reality. I looked for clips to make the trail, but it really was a bit opaque, so had an "in the bath" idea. Here are two James Burkes arguing about which one of them is real, and a third coming on and saying that none of them are, as they are all just video tape. The first two were done with a roll back and wipe spilt screen, and the third was yellow overlay so that the fringes weren’t blue. The recording I still (just about) have is terrible, because when I copied it from Philips to VHS the Philips was dying. It’s been through miniDV and various codecs since then, though from DV on, there’s no appreciable degradation – that was all analogue.
(Click on the picture below to see larger version:
use your Browser’s BACK button to return to this page)
Roger Bunce
The Tardis disappearance and appearance wasn’t done as an overlay: it was just a mix between the two scenes: I remember what luminance overlay looks like. Although we did use luminance overlay later in that story – combined with a ripple effect generator.
I thought it must be a rollback and mix, but couldn’t work out how a VT machine could play and record at the same time. Bernie and Graham’s description of going single ended, playing from one machine and recoding on the other, explains all. That would easily have been possible back in the 1960s.
I’ve got another ‘Doctor Who’ commentary to do, and desperate to find something to talk about, for an episode I can barely remember, I noticed that there were a number of appearances and disappearances – which started me thinking, "How the hell did we do that, when we were ‘As Live’?"
Pat Heigham
I spent many hours in LG ‘D’ Sound Control, playing in the sound FX for “Doctor Who”. I must have cued in the Tardis disappearance/appearance noises, but cannot remember if the visual effect was electronic, or maybe off telecine, in which case it would have been produced in the lab as a lap dissolve (or wound back) on film.
The Production Guide for Dr. Who (an absolute labour of love) lists all the TK inserts to see if it was a filmed effect.
David Brunt
Almost 20 years ago I co-wrote the ‘Production Guide’ book. I read through a lot of camera scripts.
There were a few examples where the fade was done on film – the model shot that opened and closed “The Keys of Marinus” story (the one with George Coulouis), for one. That’s April 1964.
Other times it would be done by fading out an inlayed photo caption.
I think the first roll back and mix was used in studio at the start of the second recording season – “The Rescue” in December 1964.
By the time of May 1967 and “The Evil of the Daleks” it was RB&M for several sequences in the story. This one stuck out as as it was using the technique more than others.
Roger Bunce
And I didn’t even tell you that it was "Evil of the Daleks" that I’m supposed to be commenting on. Only episode two survives.
Chris Booth
There was another variant – the Roll Back and Cut. This was a favourite of Dennis Main Wilson who used it to keep the performance going over a set change (say).
Graeme Wall
Rollback and mix (strictly – and cut) was used at AP for single frame animations for Open University programmes, advancing the cut point a frame at a time. Not sure it did the tape much good. Again this would be around 1973.
Spent a very nice 3 months at AP that summer, nominally as a cameraman but, as I’d done my inlay training which they were using quite a lot, spent much of the time at the back of Studio A gallery watching the Test Matches on my monitor.
John Howell
Gosh I remember those days, I was a Sound Sup at AP in July 1973. Pop a mic in a boom so VT could hear the animators, plug VT reverse talkback to the foldback speaker and retire to the prep room (opposite the Sound Control room) and spend the rest of the day on the Myford lathe making things such as a new piston for Barry Bonner’s hot air engine. I had to bring my own tools though, every one around the lathe was broken!
Editec
For more about VT editing and Editec see http://tech-ops.co.uk/next/editech/ Editec
Here are a couple of new bits of information:
Martin Eccles
I remember plugging the Editec tone blip from VT to the director’s intercom loudspeaker so they could cue things off just before the edit took place and end up with a finished programme on the tape.
Here is a piece by Don Kershaw about Editec from this article about television recording…http://www.vtoldboys.com/donk01.htm:
“… AMPEX came up with a controller called EDITEC. Two versions were produced which involved recording a blip of tone on the tape cue track. whilst playing back at the point where the edit should take place. The edit could be rehearsed as before and if incorrect could be adjusted using a frame calibrated knob until correct.. The Mark1 version also had an animation facility. The Mark2 version used a variable knob to adjust the cue and didn’t have animation…”
John Howell
I recall a sound feed that you could plug to a foldback speaker that went silent at the edit point so artists could take a sound cue. Of course if you got the wrong feed on the speakkkkkkkkkkkker…….!
Dave Plowman
From the Editec relay on the machine?
I left the BBC in 1976 and went to Thames at Teddington.
Editec hadn’t been much used at TC, so was surprised to find that most of the kids’ progs (and some others) at Thames used it. But the vision mixer put the dot on, and had proper talkback to VTR to adjust things if needed.
Very different from how I’d seen the cock-ups at TC in the early days of trying it.
John Howell
"That’s Life" used it at the TV Theatre for quite a time. The show was recorded ‘as live’ and Editec was used to re-start after a cock-up or if the lawyers had spotted something.
By the time the show was complete the transmission was only hours away, but to be fair it was often not needed due to that old fashioned concept: rehearsal!
Peter Neill
And there was the time the Theatre was evacuated halfway through recording due to a fire alarm or somethings. Eventually we and half the audience finally returned and we did a roll back and carry on live.
Bernie Newnham
I think that ‘through the studio’ editing went away pretty quickly at TC because people didn’t like it, especially after the great Sherlock Holmes debacle. It’s a whole lot easier to sit in VT and do the same job and get the result better, but of course takes more resources.
Dave Plowman
The one in my mind was “Troubleshooters”. We ended up going backwards because of (I reckon) those doing it hadn’t a clue how to. It’s never a good idea to practice on a real programme. I’m not sure I ever saw it in use again at TC.
