Keyboard Magazine "In The Studio"
Article # 27
by Bobby Nathan
History of Automation
As multi track recording grew from 4 and 8 track to 16 and 24 track and 48 track the need to automate the mixing process became inevitable. Even the best producer and engineer team were only as good as the four hands involved muting faders and riding the levels. Even with the aid of a qualified assistant which usually added another hand or two it was impossible to get the same feeling twice without many passes of practice. What's more important with the producer and engineer watching their cues for cuts and fader levels rides, their attention was taken away from concentrating on the full intensity of the music and the mix. Those that were around and used early primitive automation systems had to pay the price but many felt it was well worth the aggravation of using systems that were prone to crash and did not always play back the mix the same. Today automation has been perfected to a level where it is still not transparent but is advanced enough so that's its reliability has made it part of the every day mixing procedure.
Data Storage on Tape
The first automation systems required two tracks to store the computer's data on tape. When working with less than 24 tracks giving up two tracks was a real sacrifice. The procedure was as follows: You would print your first pass of automation data of the mix on track A. Usually in this pass if you made any mistakes you could roll back and punch in but you would have to play the entire pass back after you were finished to be sure there were no glitches. Many times later after additional passes, you would then discover that there were glitches there at the punch in/punch out points even if they did not seem apparent upon close inspection after the initial track of automation data was recorded. So even though you could punch in on the initial data track, it was not a good idea if you were planning to do many additional updates.
The next pass would be recorded on track B while reading the data from track A. Additional mutes and/or fader level changes could now be written as data on track B while the engineer and producer listened to what had been previously done as the console's computer read the data on track A. For the third pass you would read the data off track B and write new data over the previously recorded track A. This whole procedure had a number of downfalls.
The first being the engineer always had to keep a conscious account of which track he was reading from and which track he was writing to. This was further aggravated by the fact that patch cords had to be repatched at the end of every pass. For instance from tape out to processor in (for the track you were reading from) and from processor out to tape in (for the track you were writing to). Now if this wasn't enough of a pain, a bigger problem was the fact that the computers scan time was slow. Every time that you read data from the tape and added new data bouncing to the other data track, the mix would be delayed by a few milliseconds. Critical mutes would eventually become so late that they would have to be written over again. The same held true with critical fader movements.
Finally smpte based automation came into play. Smpte based automation only required one track to be given up on the multi track recorder. Many of today's SMPTE/MIDI devices use the same track to sync up sequencers and drum machines as the automation. In Smpte based automation systems, the data, instead of being stored on tape, was stored in RAM memory and on floppy disk. Since smpte is a constant, mutes and fader levels never became late. Also, because each pass or a portion of was stored as a separate mix you could always go back to one of your mixes anywhere along the way. Many times an engineer and/or producer will lose the feeling after eight, ten, fifteen or twenty hours of mixing and then decide to go back to mix number 7 for example which was done in the sixth or seventh hour of mixing. With data storage on tape automation (as mentioned above) you only had access to the last mix and the mix before. Another important feature of smpte based automation is the ability to join or merge pieces from any different mix such as the intro, first verse and first chorus of mix twelve joined with the second verse and second chorus of mix 7 and etc.
Moving Faders Vs VCAs
In the early days of automation voltage controlled amplifiers (VCAs) were very noisy and 36 of them would add an cumulative amount of noise that made their use prohibitive. Some of the early console manufacturers such as API and Neve came up with the idea of motorized or moving faders. The computer would send plus or minus voltages to the fader's motors which would move the faders physically. A lot of engineers preferred this visual approach to mixing. Watching 36 or 48 motorized faders dance on the console was very much like seeing a player piano move its keys in a ghostlike fashion. The only drawback to this type of automation is that it was very expensive. It could not be retrofitted in just any console and from a maintenance standpoint was much more of a nightmare than VCA faders. I should point out though that Yamaha's DMP-7 console has moving faders and is realtively inexpensive. Getting back to back then, DBX was one of the first companies to offer a reasonably priced low noise VCA fader. This brought automation costs down by one third. Unfortunately fader movement was tracked by only two LED null lights, a + and a -. If you wanted to punch a fader level change in at a certain point you would have to roll the tape a few bars before and set the fader levels so that both the plus and the minus null LEDs would light. This would indicate that the fader was set at the correct position. This was not necessary with motorized faders because they were always at their correct position. With the advent of smpte/disk based automation onscreen video display of fader levels became a visual player piano type effect on a video monitor screen but faders still had to be moved manually to match their previous position before any punch in's could be carried out.
Reading, Writing, Groups - Fader Levels and Mutes
Up to now we have been talking about Reading and Writing but there are other modes of automation to discuss. For instance, if you liked the way you had written a certain track's mutes but not it's level's you must specify to the computer that you only wish to change that track's fader levels. But let's say for instance that you had three tracks of background vocals. You had to spend about an hour to balance out their levels. You had to mute certain parts that they extended too long. And then you discover that the four tracks of background vocals are too loud. You would then assign them to a group. You could then adjust the group's master fader to adjust all their volumes. If you wanted to write a mute to several tracks at once for a breakdown section, group masters come in real handy for this feature also. By muting the group master's mute you would be muting whatever was assigned to that group. Many times several groups can be assigned to another group as their master. It is customary to assign drums to a group. Lead vocals to another, Background vocal tracks to another, Horns tracks to one group and on and on. Even though a track is assigned to a group master, it still can be controlled separately. Having control either way is to the engineer's advantage and makes making the moves a lot faster.
Building a mix
The art of mixing is as varied as the technique's of a skilled chef. There is no set formulas but here are some. Most engineers start by getting the sound of all the tracks together first. This means equalizing the tracks to sound as envisioned. Compressing problem tracks such as vocals, horns lead instruments. Choosing the panning assignments. Noise gating problem tracks to quiet up the mix. Adding the desired amount of ambience to the drums, guitars, synths, vocals and horns. You might be asking right now "How can the engineer know?" If the engineer has been working on the project since scratch, he already has a clear picture of what the artist and producer are striving for. If he hasn't been around since the beginning and was just brought in on the project, it's because he usually has a "trademark" type sound. This means the drums and instruments will have a certain predetermined equalization and ambience. The producer usually lets the engineer have time to get all the tracks ready and then will make comment's about the sound. At this point the engineer will make changes in Equalization, panning, ambience and etc. When they both feel confident the mixing will begin.
Writing mutes is sort of the next standard procedure. With the help of the producer and or Artist, the engineer will write mutes to the automation. The arrangement can thus be made more definite. A lot of times during the overdubbing, parts were played all throughout the song and after close inspection only really belong at certain parts. This writing of mutes sometimes takes a couple of passes to get all the mutes right.
All during this time the engineer is writing levels. After the mutes are set, dynamics of levels can be concentrated on. Certain tracks may need to be louder at different parts in the song. Equalization and compression may have to be changed due to the fact that there are less instruments at a given part or their dynamics have changed. All in all the constant tweaking and re- tweaking goes on and on until a final mix is reached.
Recalling a Mix
Now up to now we have only been talking about automation controlling fader levels and fader mutes unfortunately there are many more sections of the console that automation does not remember in real time. Some of these sections are the line input trim, the equalization section, the echo sends, the output buss assignments and the dynamic sections which can include a compressor and/or noise gate. One of today's more elaborate consoles include a feature called total recall. Total recall can take a snapshot of every knob on every channel and store this to disk. Then one by one the engineer can compare each channel's knobs and buttons to a visual image of how each was set when the recall snapshot was last taken. Even though each button and knob has to be turned manually, with the visual aid of the computer the engineer can rest the board exactly as it was before.
As wonderful as total recall sounds it is still only part of what has to be notated and reset to make the recall valid. Every digital delay, digital reverb, outboard compressor, equalizer, noise gate, and any other effect that was used that is not part of the console, must be notated. To save time setup sheets are made usually by each studio for this very purpose. It is the usually the assistant engineer's responsibility to correctly notate correctly how each piece of outboard gear was set and on what track and channel is was used on. That big mess of patch chords in the patchbay which is reminiscent of a heaping plate of spaghetti at your grandmothers house must also be correctly notated. Many times the engineer will cross patch tracks returning from the multitrack to group certain instruments on certain faders at the console. For example, rhythm section instruments together , drums and other percussion tracks that do not need a lot of attention during the mix are usually placed outside the engineers immediate reach. while critical track's such as lead vocal , lead guitar, horn and string and synth lines are more conveniently located. All this cross patching again much be notated. If it sounds impossible it's not. Using Automation for large midi setups
With todays elaborate midi setups, automation can be and should be used to track the intricate balances of large midi textures. For instance it is common to have sometimes 3 4 or as many as 8 midi synthesizers or an entire TX rack to create a string or horn texture. Even though we can store and recall all the synthesizer's and sampler's patches to either cassette, cartridge or disk, it is the audio balance of all these sound producing modules that has to be stored and recalled if one wants the ability to come back another day and punch in and out on the same overdub track. By printing an automation pass of the fader levels at the console will help insure recalling a large midi setup.
Using Automation for effects
By patching a track into two channels, panning effects can be achieved. Each channel is panned left or right (or nine o'clock and 3 o'clock). Equalization and reverb is set for each of the channels (either the same or different for more dramatic panning). By moving the fader levels alternately up and down on the two channels, you can write superb panning effects.
You can use a channel to feed a delay unit. Ever want to catch one word of a lead vocal into a repeating delay unit? By assigning the track to a particular output bus and then returning that output bus to another track, you can then control or interrupt the signal before it gets to a delay unit. You would assign the new channel to another unused output bus and patch that into your delay unit. By working that new channel's mute and fader you can catch a word or phrase into a delay unit and have the automation do it for you.
I hope you've enjoyed this basic discussion about automation. The topic is virtually endless and if anything, I hope I've got you thinking! Until next time ......
Back to Bobby "Guitar" Nathan's Keyboard Magazine Articles Page
copyright 1985 - Bobby Nathan - May not be used without consent!