That Syncing Feeling (Part 1)
Lock your sequencer to audio and video with a little help from Chris Smith's definitive guide
With a sequencer at your disposal a whole wealth of possibilities can be opened up by synchronising your equipment to tape. Chris Smith begins a series on this topic.
This guide covers all the aspects of synchronisation that any modern musician is ever likely to come across. Because the subject is so large, the guide has been divided into two parts. Part one covers the basics of MIDI and tape sync, and looks at the various features available on synchroniser units. Part two will cover the finer points of recording and recovering time codes, syncing tape machines together, syncing music to video, VITC, MTC, and mix down automation.
What is Synchronisation? Synchronisation means making the right things happen at the right time - with respect to each other. While this is generally a good idea in day to day life, it is practically essential in music.
When musicians play live they are "synchronised" by some central rhythmic event such as the bass and snare drum pattern or a conductors baton. That is not to say that the bass guitarist won't play his next note until he hears the snare drum play. The human brain is pretty good at keeping time and if you give it four beats in time it will usually fill in the next four quite accurately. After about 12 beats however, that accuracy begins to waver, and after 120 beats it may be two or three beats out from the original clock (try this yourself with a digital watch). This is called drifting.
If you give each player in a band a four beat count in and let them play for 30 bars or so, without being able to hear each other, they would all drift apart quite wildly, some faster and some slower than the original tempo. This is where the drummer comes in. He can't necessarily play without drifting any more than any other band member, but because they all hear "his beat" they all make subtle adjustments to their own timing and all drift with him. So everyone is still playing together half way through the song (with the exception of modern jazz maybe, where they only seem to start and finish together). This is the essence of synchronisation.
In conventional music terms it simply means playing in time. In the domain of music recording, sequencing and linking sound with vision it becomes the single most important technical concept to understand and implement.
A sequencer, whether it be a dedicated unit or a computer with a MIDI interface and appropriate software, is both the conductor and the manuscript for each of the instruments it controls. It uses MIDI commands to tell each synthesizer or sampler what notes to play and at what time. The software itself makes sure that each instrument track is synchronised to the computers own internal clock. When you record a new pattern into the sequencer it will give you a drummer (metronome) to keep in time with or let you play along with what you have already recorded.
Quantisation helps here by pulling all your slightly late or early notes onto the right beat, or fraction of a beat, though there is a move towards humanisation recently, which means just the opposite.
Now let us introduce a drum machine, with its own clock and sequence of patterns, set to play at the same tempo as the sequencer. Even if you do manage to start them simultaneously the sequencer and drum machine will have drifted apart a little before the end of your song. This is because their tempo settings vary slightly by as much as one tenth of a beat per minute. To make them play in time you have to slave one to the other. Putting the drum machine in external MIDI sync makes it respond to MIDI clocks from the sequencer, as well as MIDI stop and start commands. Most modern drum machines also respond to Song Position Pointers, a MIDI code output by the sequencer when it starts playing past the beginning of your song.
The SPP tells the drum machine when and where to start playing to be in time (and place) with the sequencer. For the technically minded MIDI tempo clocks are transmitted at a standard rate of 24 pulses per quarter note (ppqn). One quarter note is a beat in 4/4 time giving 96 MIDI clocks per bar, or 48 clocks per second at a tempo of 120 beats per minute (bpm), or about one clock every 21 ms. A SPP codes the total number of clocks elapsed from the beginning of the song to the requested starting point (ie. continue point) and is followed by MIDI clocks at the correct tempo. Usefully, this locates the slave (drum machine) correctly, regardless of any previous tempo changes.
Of course your drum machine can just as easily slave your sequencer, but make sure both equipments send and respond to external MIDI sync and SPP.
Okay, you've got synths and drums all synced up, what about guitars and vocals (ie. non-sequenced parts). You could just play or sing along with the sequencer and record them all to audio tape together - but then you can't change anything later. That doesn't leave much room for error or experiment, and you have to record the sequenced parts before anything else. This is where tape sync comes in. The main advantages of tape sync are:
1) You can add, take or change sequenced parts at any time and stay in sync with music recorded an tape.
2) Playing from the sequencer, you can use the same synthesizer sounds more than once by overdubbing them with recorded parts on tape.
3) You can utilise far more tracks by playing sequenced tracks alongside tape recorded tracks and then mixing down the whole lot together.
4) You can synchronise other units to tape as well as the sequencer.
All tape synchronisers generate a time code (usually audible) that can be recorded onto one track of a multitrack tape recorder. When this tape track is played back the synchroniser reads the code and generates MIDI clocks (and other information) to synchronise the sequencer to music recorded on the other tape tracks. So you lose one tape track, but gain a whole load of sequencer tracks.
The simplest devices convert MIDI clocks into audible signals using Frequency Shift Keying (FSK) type codes. You play MIDI clocks from your sequencer (internal sync) into the synchroniser and record the FSK time code. Once recorded the time code is played back and converted back to MIDI clocks by the synchroniser which controls the sequencer (now in external sync). This basic arrangement is straight forward enough and is often "built in" to sequencers and drum machines (which therefore don't require external tape sync devices). It does however, have four major disadvantages:
1) Once recorded the time code tempo is set. You can put in as much tempo variation as you like while recording the code by varying the clock rate (ie tempo) of the sequencer, but you can't change it later on. So you need a pretty good idea how your song is going to end up (in terms of tempo changes) before you start.
2) The FSK signal itself is just series of pulses and doesn't contain any song position information. So to get the tape and sequencer to line up you have to play from the start of the pulse train on tape at the beginning of the song every time even if you just want to record a bit of guitar over the last 20 seconds. Some tape sync devices use a special code called "smart FSK" with SPP information. They may take a bar or two to lock up your sequencer to tape, but do save a lot of time.
3) Most sequencers only have one MIDI input. If you connect this to the MIDI output from your synchroniser that leaves nowhere to connect the MIDI output from your keyboard. If you want to record new sequencer parts while synced to tape you'll need a separate MIDI merge unit to merge the synchroniser output and keyboard output together before they go into your sequencer. Some tape sync units have a MIDI input and do the merging, internally a feature well worth having. MIDI merge is not usually a problem for sequencers with built in tape sync.
4) Hassle. FSK is not renowned for being the most reliable or universal of time codes and tends to be specific to the device that generates it. This is fine for the smaller discreet recording set up, but if you intend to take your homework into the studio, they'll expect to see SMPTE time code on your tape.
SMPTE (pronounced simpty) stands for the Society of Motion Picture and Television Engineers who standardised the code back in 1967. It consists of a binary code that identifies each point along the tape cue track with a specific frame, second, minute and hour. This time code was originally developed for the TV industry and is usually laid down on an audio track while a video camera is filming. Being able to identify each individual frame (picture still) greatly simplifies the process of editing a series of shots into a complete production. The SMPTE standard is for the American TV system which uses 30 frames per second (fps), corrected to 29.97 fps for colour systems (called drop frame).
The European TV (EBU) equivalent works at 25 fps and a 24 fps rate is reserved for film use. For audio work it matters not which rate is used, as long as it is used consistently, though in the UK it is best to stick to 25 fps just in case your work spills over to video. Unlike FSK tape sync, there is no direct correlation between the SMPTE time code and musical tempo, MIDI clocks, song bar numbers, or anything else. The code is recorded onto the audio cue track before hand and is standard and fixed.
To derive musical timing requires a SMPTE to MIDI converter, the subject of the remainder of this article.
Nearly all SMPTE synchronisers (ie. SMPTE/MIDI converters) perform a number of basic functions:
1. They generate, read and display the four basic time code rates and hence give a very accurate tape position count.
2. They can derive at least one user definable tempo from the SMPTE code and output MIDI clocks to synchronise a MIDI sequencer to tape.
3. They can provide at least one user definable cue point to start a MIDI sequencer at a specific time.
4. They incorporate a MIDI merge facility to enable them to be used in a MIDI/tape recorder set up similar to the diagram.
Firstly the code should be recorded (striped) onto one track (usually the outside track) of a multitrack recorder at a level high enough to be easily read, but low enough to avoid cross-talk onto other tracks. Between -5 and -10 VU is best, a SMPTE output level control is useful here, especially if your multitracker has no cue track input VU pot. Record the time code from zero and for at least twice the length of your song. Set the MIDI start time on the synchroniser well into the stripe on tape (about one minute to be safe) and put your sequencer in external sync mode.
Nothing should be recorded onto the other tape tracks from the sequencer unless it is synchronised in this way. If non-sequenced music (guitars, vocals, real drums etc.) is to be recorded, it must be played in time with music already on the sequencer or a metronome synchronised to tape. Follow these simple rules and your recorded music will always stay in perfect sync., wherever in the song you start the tape. Lock up time is usually so quick that the sequencer tracks become just an extension of the tape recorder tracks, under full control of the tape transport.
Physically SMPTE units come in three guises, desk top units, 1U rack mounted units, or as peripherals for computers driving music software. Peripheral sync units are now available for most Atari ST and MAC music software packages. Most of these have the inherent advantages of mouse control, infinite tempo variation, cue points and VDU SMPTE read out.
They are each specific to one make of sequencer though, and tend to be expensive. Discrete synchronisers have been around for some years now, but have recently become more affordable and therefore more popular. If you are in the market for any kind of tape synchroniser, the buyers guide table should help you. Most of the features listed have been covered already. The ones that haven't are listed below.
This can be very important when the synchroniser is the control unit in a dimly lit studio. LCD tends to be dull and difficult to read at angles. A large bright LED read out works best.
This usually means the old Roland standard of 5 Volt DC pulses and is useful for syncing those golden oldies like the TB-303 baseline or TR-606 drum box as well as some old Linn gear. There's no SPP, but if you are into House music it's a must.
This feature is a rarity on synchronisers but has tremendous potential for video applications.
A cue point is a time specified in frames, seconds, minutes and hours at which an event or trigger is activated by the synchroniser. This event can be the playing of a MIDI note; a program change; a start command; a voltage trigger output from the synchroniser; or some other user definable function. For example, imagine adding a sound effect, say a car door slamming, to a video sound track.
The easiest way is to trigger a sampled door slam sound from a sequencer. You could try watching the video and play the samplers MIDI note into your sequencer "on the fly", but you are likely to be that bit out. You could now use the sequencers track delay features to line up the sound and vision, but it's all a bit hit and miss.
A much easier method is to stop the video at exactly the right visual point, read off the SMPTE time and then program your synchroniser to output the event at that specific time. Your sequencer can record a series of cue events programmed in this way and use one sampler for lots of sound effects throughout the film. In fact many modern film sound tracks are made up almost entirely of sampled sounds (even speech!) put together this way, all be it with extravagantly expensive equipment. The important concept is that the synchroniser can define event times in real time, rather than musical time. Most synchronisers allow at least one user definable cue point, the sequencer start time. This time can usually be changed in frames or bits (1/80 of a frame) to introduce an offset between tape recorded and sequenced tracks - useful for echoing and ADT effects.
Typically your music will have a precise tempo, with a few specific tempo changes programmed into the synchroniser. What if you want to derive a tempo from a live drummer, or sync, up to music already recorded without a SMPTE track? For this you need a synchroniser that will take external synchronisation from either a MIDI clock, or a tap input, or an audio input.
The SMPTE code is laid down on tape as normal, or on a track alongside existing music. On play back the synchroniser reads the SMPTE code together with the external sync input, derived directly from a gated rhythmic sound from tape (eg the snare drum) or someone actually playing a "tap input" in time with the recorded music. The synchroniser then simply memorises this input timing (and all its variations from the theoretical true tempo) and in future outputs this MIDI tempo with appropriate SPP's whenever it hears the SMPTE code.
Clever eh? Well yes, but it is not often that easy in practice.
This is simply the ability to save all the parameters you set up on your synchroniser to your sequencer. It is best to save this system exclusive information to one of your songs sequence tracks if you can, rather than in a separate software dump utility.
Some synchronisers use tape dump instead of MIDI. The less said about tape dump the better.
Resolution is a mine field of misinterpretation, and is more a parameter of your sequencer than your synchroniser. It refers to the absolute accuracy with which a note can be played in time. A standard SMPTE unit throws out MIDI clocks at 24 ppqn. (This can be increased to 96 ppqn on most units, only useful if your sequencer responds correctly to the higher rate.) This accurate clock cues along your sequencer while it plays and records MIDI data from your keyboards.
Now if a new note comes along in between two clock pulses the sequencer has to determine exactly where it is in terms of fractions of a clock pulse. It does this by "sampling" its MIDI input to see if a new note (or any other MIDI event) has arrived. The frequency with which it looks at its input buffer determines how accurately it will record the timing of each new MIDI event, and hence the resolution of the sequencer. In effect this is the minimum quantisation value, and it is usually the same whether the sequencer is in internal or external sync.
Tempo resolution is quite different. This is the accuracy with which the bpm value can be set on a master clock. Most devices can be set to say 120 or 121 bpm. Now 120 bpm can be derived directly by dividing a 6 or 2 MHz clock (usually the quartz clock frequency used in this equipment). However, to derive 121 bpm requires a far more complex calculation, and a certain amount of rounding up or down.
Because different clock devices calculate tempos to different levels of accuracy (ie. numbers of decimal places) and round up or down inconsistently, their definitions of 121 bpm may vary, causing drifting between two unsynchronised clocks over a period of time. So some devices offer the ability to set the tempo to say 121.01 bpm to avoid drifting against other clocks not synchronised to them.
To put all this in perspective, high tempo resolution is only really useful for accurately syncing without slaving. The whole idea of synchronisation is to slave everything to a single clock, and the absolute accuracy of that clock is not particularly critical. The average listener can hardly detect a tempo variation of less than 1 bpm through the whole song anyway. The golden rule is to use the same synchroniser throughout the recording whenever you add audio tracks on tape in time with sequencer tracks. A particular SMPTE unit will always convert SMPTE code to MIDI clocks in exactly the same way.
This is a new standard being incorporated into some sequencers and synchronisers that offers some advantages over MIDI clocks and SPP. The synchroniser actually converts the audio SMPTE code into a MIDI code containing frame/second/minute/hour information instead so that the sequencer can respond directly to SMPTE. It seems unlikely that any sequencer system will support MTC exclusively in the near future.
There are many different makes of synchroniser around offering many different features at very different prices. Ask yourself which features you will actually need in your recording set up and whether it is worth paying extra for the features you may hardly ever use. On the other hand one particular feature may make a more expensive unit invaluable in the studio.
The buyers guide cannot possibly cover all the pros and cons of each unit and it is always best to read a review or get a demonstration before handing over the reddies. Only one thing is for certain. If you have any kind of multitrack tape recorder and sequencer, the power and flexibility of your music system will greatly increase by adding a synchroniser.
Feature by Chris Smith