Real time is an illusion caused by not working in a studio. In fact time, as mere mortals outside of control rooms know it, ceases to work in the normal way once you pass into any studio. Once inside, time is twisted; a three minute track will take a whole day to record, a 12 inch single lasting six or seven minutes will not be finished for a few days and the mere 40 minutes of an album can actually last weeks or months. At no time during the recording process is this time warping more evident than at mixdown where every second of a track will be subject to many minutes of attention over and over again, while the engineer tries to get this snare or that bass note to sound that little bit better.

Meanwhile, out in the real world, the show goes on. Out here a minute is a minute and an hour is an hour, and if it isn't right then you have had your chance so forget it and go on to the next song. I'm talking about live music.

Anyone who has ever mixed a live concert, a theatre show or musical will probably know just how useful a scene setting mix system could, or can be. Back in the studio the ultimate in remix technology, where every movement of fader and every channel mute is stored in perfect synchronisation to the music, becomes practically useless in a live situation.

An automation system for remix must be linked to a timecode track on the tape. The display of the timecode bears no relation to the clock on the wall, whereas for a live concert, in 'real' time, there is no way to generate a timecode linked to the music. Every live concert or theatre show is going to vary in length from performance to performance so setting a preset mix sequence running would certainly not keep control changes in sync with the show.

A scene setter gets around the timing problem by allowing a series of 'snapshots' of the fader and mute positions to be stored along with a number or a name. Each scene or section of the show will then have a mix set-up which can be recalled at any time. Once a complete show has been built up in this way the sound engineer can simply step through the preset scenes at times indicated by a script or cue sheet. In effect, the sound engineer becomes the 'clock' for the system.

Compared to an automated mixdown system, a scene setter or 'realtime' automation system is comparatively simple. A remix system must be able to store any number of fader moves and mutes to an accuracy of a few milliseconds and so scans every automated control during those few fractions of a second. Along with this fast scanning system there must also be some way of processing enormous amounts of data. For example a slow fade on a handful of faders will produce anything up to a hundred bytes per fader per second which, even with fancy data compression, will require quite a large memory for a 15 minute mix.

On the other hand, each fader set-up on a scene setter system only needs one byte per fader and one bit for each associated mute. Even allowing for some space to store a scene name, one 'snapshot' of the desk will occupy less than a hundred bytes. A complete two hour show with say, two hundred cue points will fit in 20k of memory. Not only is the memory required smaller but data in/out can be slower. Anything up to around 20 milliseconds is going to be faster than the reflexes of the operator hitting the cue point but still fast enough to be an inaudible delay and 20 milliseconds is an enormous amount of a time to a modern micro.

The hardware required for a basic fader scene setter is relatively small. Each channel fader is replaced by a VCA in the audio path and the existing fader is used to generate a voltage varying from zero to five or 10 volts. When a scene is stored a multiplexing system scans each channel in turn and passes the fader voltage to an A to D converter. This converts the fader voltage into a byte which can be stored by the computer. Meanwhile, a D to A converter converts each piece of channel data into an analog voltage and this is returned to the relevant channel by a demultiplex system. The system can act as a normal desk by simply returning each fader voltage to its own VCA immediately and this is the way that a scene setter system would work during rehearsal when the scenes are being set up.

Having introduced scene-setter systems as being totally in the realms of live music, there would of course, be a place for scene setting in a small studio as a cheap automation system. It is in fact with us in part in the form of mute grouping where each channel has several mute switches to allow sets of channels to respond to a master mute switch. However, anyone thinking that they would like a scene setter system is going to end up having to buy a big expensive desk. Although there are several companies promising or looking at scene setter systems there are no widely available, cheap systems on the market.

THE BIG TIME

The professional end of the market in scene setters divides neatly into two areas; assignable consoles and standard consoles with either automation or specially designed add-ons. In the second category are companies such as SSL and Cadac, both with systems for their existing desks, whereas the first comprises the new assignable consoles from Amek, Harrison, Audix, Neve and soon Trident.

SSL's Real Time System is mainly a software add on for their existing Studio system. It allows the storing of a mix along with a relevant name and editing of pre-stored scenes into a sequence list for a complete production. The controls for changing scenes are interesting in that SSL has adopted a system similar to that used for lighting and video consoles. Two seven segment displays display the present and the next scene numbers. There is a push button switch to cut instantly to the next scene but also a fader which allows fading into the next scene, a feature not found on most other scene setters.

Assignable consoles are completely different altogether. A fully assignable console uses a completely new and different type of architecture to most modern desks although some manufacturers are now trying to bridge the gap to make assignable desks more acceptable.

The basic assumption behind assignable desks is that you only ever use the controls on one channel at a time, i.e. you can only work on one EQ section or auxiliary section at a time. This means that you only need one channel's worth of controls, and this one set can be moved to or 'assigned' to any channel. There is still a small control area for each channel, usually with a fader, mute and solo switches and an ASSIGN button to assign the main controls to that channel.

There is an immediate saving in that instead of having 100 switches and pots on each of 32 or so channels you have just one set of 100 controls. However this saving is totally lost under the enormous spending that has to be made on building a mixer that is completely controllable by the computer. For example, every one of the Auxiliary send pots on each channel (which used to cost about 20 pence) now has to be a voltage controlled amplifier (VCA) costing anything up to five pounds. Add to this the cost of the conversion circuits to allow the VCA to be controlled by a master processor and instantly the price of a desk is multiplied by anything up to ten times.

So why is it done?

Because the position of every pot, switch and fader has to be stored and manipulated by the processor in charge of the desk it is no problem to put the present set-up of every control on the desk into a file and store it on a disk. Instantly your very expensive, but average, desk becomes an enormously powerful scene-setting system, with not only faders and mutes being stored but also all the other controls on every channel of the desk. This is more or less the system that Audix and Trident have gone for. The Trident system displayed at this year's APRS has 255 scenes or 'events' as Trident like to call them. The Neve DSP offers what, to the user, is an identical form of control system, giving one small set of controls and a fader for each channel and then one large set of detailed channel controls which is assigned to each channel as required. There the similarity ends since the Neve DSP handles most of its internal signal processing in digital form.

From the basic assignable idea there have been two offshoots, both trying to close the gap between what many see as the obvious situation of a big 'one control per function' desk like the SSL4000 and the more expensive but smaller and more powerful assignable consoles. To this end Amek and SSL have both gone to desks with a full complement of channels and channel controls both with the reset of push switches from memory but with the rotary controls, such as pan and auxiliary sends still having to be reset by hand. On these desks scene setting is not an inherent feature so would have to be added as an option via an automation system. Similarly Harrison has gone almost all the way to linking assignable hardware to a complete desk. On the new Harrison 10 series console all controls both switched and rotary are totally resettable from memory both in real-time and synced to SMPTE timecode on tape. It promises to be big in studios for the ultimate in automated mixing.