40 Sequencer Essentials
In this lengthy article, Craig Anderton explains all the features found on sequencers, from the simple to the esoteric. It is intended for those who are trying to make an intelligent purchasing decision about a sequencer and don't fully understand all the options that are available.
This article explains all the features found on sequencers, from the simple to the esoteric. It is intended for those who are trying to make an intelligent purchasing decision about a sequencer and don't fully understand all the options that are available.
One sequencer cannot be all things to all musicians, so different sequencers include different features. Unfortunately, the sheer number and complexity of these features can be confusing - how many tracks do you really need? What type of synchronisation works best? How important are different types of quantisation and 'humanisation'?
This article describes the most common sequencer functions so that you can best evaluate which sequencer most closely meets your specific needs. To further de-mystify the subject, we'll mention typical applications for some of these functions.
No sequencer contains all of the features mentioned below and, of course, more features usually translates to a higher purchase price. But don't automatically assume a low-cost sequencer can't do the job - there are many 'budget' sequencers that do almost as much as the 'big guys'. There is one important caution, though: as with synthesizers, not all manufacturers refer to the same feature by the same name. Where possible, alternate names for features are given below in parentheses. Ready? Let's go...
Nowadays sequencers seldom have less than eight tracks, and software packages offering 64, 200, or even an unlimited number of tracks are available. For very basic applications eight tracks might be adequate, but with the proliferation of multitimbral instruments, 16 tracks (corresponding to one for each MIDI channel; see 'Tracks & Channels' sidebar) is a virtual necessity.
With only 16 MIDI channels available, having more than 16 tracks may sound like overkill - but there are valid reasons for such things as 64-track sequencers. Even though you probably won't want to play back data to 64 different synths, extra tracks can be used for organising sounds. You could record each section of a part on its own track, which makes it easier to modify, say, just the first verse while leaving the rest of the part untouched (editing can also be done with punch-in and punch-out features, but separate tracks are often easier to manipulate). Separate tracks can also store MIDI controller data; you might want to punch-in a difficult pitch bend without disturbing the notes themselves, which would be recorded on a different track. The controller and note tracks could then be bounced (merged) together to create a single, composite track containing the pitch bend and note data.
But what happens when 16 MIDI channels are not enough? Clever software manufacturers have come up with a variety of ways to circumvent that problem. The most common approach is to have two (or more) completely independent MIDI outputs, and route some track data to one output and other track data to the second output. Thus, data from one track could be assigned to MIDI channel 1 and routed to output A, while data from another track could also be assigned to MIDI channel 1 yet be routed to output B. This essentially doubles the number of available MIDI channels. The Southworth JamBox interface, used in conjunction with their MidiPaint sequencer, provides four completely independent outputs and yields the equivalent of 64 MIDI channels - 1A-16A, 1B-16B, 1C-16C, and 1D-16D.
Another way to double the number of channels is to simply synchronise two sequencers together: one sequencer provides one set of 16 outputs, and the other sequencer provides another set of 16 outputs. Note that since data is divided among the two sequencers, there is less of a tendency for the 'data clogging' that can occur when a single sequencer is forced to handle every single piece of MIDI data.
It's frustrating to have MIDI functions on your synthesizers that you can't access with your sequencer, so the ability to record pitch wheel changes, Aftertouch, Program Changes, and Continuous Controller data is vital. Being able to record System Exclusive (SysEx) data, while not possible on many sequencers, offers several advantages - specifically, the opportunity to reconfigure an instrument's status (including loading in a new set of patches, if need be) in the middle of a tune. One enterprising colleague of mine, Jim Johnson, even uses Dr. T's Keyboard Controlled Sequencer (which can record SysEx data) to transmit messages and prompts to the display of his Oberheim Xpander.
Anything that can be recorded should also be editable. Suppose, for example, that you play a part perfectly but mess up on some of the pitch bending. If you can't edit controller information independently, you would need to play the entire part over again. With controller editing, you could change just the pitch bending itself.
Quantisation rounds off timing errors in your playing to the nearest note value you specify - quarter notes, eighth notes, triplets, etc. Some sequencers correct only during playback, which is a useful feature since you can correct an already recorded track and change quantisation at will. Generally, a high resolution or real time mode will also be available that turns off auto-correct (see next section).
Typically, a sequencer's high resolution mode gives a resolution of 96 pulses per quarter note (in other words, your parts will still be quantised, but to such a fine degree of quantisation - 1/96th of a quarter note - that it will appear not to be quantised). However, musicians who are very sensitive to subtle time shifts feel that 96 ppqn resolution is not sufficient, so some sequencers offer 240 ppqn, 384 ppqn, or even higher resolution (eg. C-Lab Creator, which runs at 768 ppqn) to ensure that notes are recorded exactly where you've played them. Another advantage of high resolution sequencers occurs with track time-shifting (described later).
Advanced programs will offer different types of quantisation, such as quantising only those notes that fall outside an acceptable range of quantisation (for example, if you specify a 10-clock pulse 'window' around a particular beat, then events that occur more than 10 clock pulses away from that beat will be quantised, but events within 10 clock pulses of the beat will remain exactly as recorded). This helps you keep a good 'feel' and still trap the most significant timing errors. Some sequencers let you quantise only notes within certain pitch ranges. Adjusting these parameters can really add a new dimension to a part.
Note that there are several kinds of auto-correct function. Some quantise Note-On information but not Note-Off, which can distort the note duration; others quantise both equally, but arguably the most natural approach is to quantise Note-Ons and adjust Note-Offs to maintain the same duration. (Note that Mark Of The Unicorn's Personal Composer provides several variables for adjusting quantisation to taste.)
Sometimes quantisation leaves a track sounding overly-mechanical, which is where humanising options come in. These let you add subtle timing variations to note attacks or durations, or small level changes to velocity, to make a track sound more 'human'.
With primitive sequencers, editing is often done on a track basis - eg. if you want to quantise, you have to quantise the entire track. More advanced sequencers let you define a region of the sequence for editing so that, for example, you can quantise two measures of a track but leave the rest of the sequence alone. In many cases, the defined region can include multiple tracks as well as multiple measures but the catch is that these must generally be contiguous - in other words, you can't pick a measure, skip two measures, then select the next few measures to be part of the region including the first measure selected. The same goes for tracks; you would be able to define, say, tracks 1-6 as a region, but not tracks 1-3, 6, and 8. With non-contiguous editing, you can in fact specify isolated sections of the sequence as being part of the same region for editing purposes. At the time of writing, Performer 2.3 is the only sequencer I know of that allows for noncontiguous editing.
This is mostly of concern with stand-alone, non-computer-based sequencers; in fact, many stand-alone sequencers include a built-in disk drive. Saving to disk is a lot faster than saving data on cassettes, and also more reliable.
Note that with computer-based sequencers, the speed of disk operations depends very much on the machine being used. The Commodore 64 is notorious for its incredibly slow floppy disk operations; Macintosh, Atari, IBM, and Amiga disk drives work much faster. The best option for data storage, though, is the hard disk. Compared to floppy disk drives, hard disks use a different type of technology that offers fast access and the ability to store lots of data (typically 20 to 80 Megabytes). Few stand-alone sequencers include hard disks, but the Emulator III and Emax HD samplers can store sequences on their built-in hard disks, and most computers can be outfitted with an external (or sometimes internal) hard disk. While hard disks used to be quite costly, prices are falling rapidly; concerning the future, IBM recently announced a process that could increase hard disk densities 50-fold! However, it's important to remember that hard disks can (and will) fail someday, so in any event, hard disk data must be backed up on some other medium - usually floppy disks.
Being able to load and save individual tracks (not just songs) to disk is also a useful feature.
Real-time recording works like a tape recorder - put the sequencer into record and play away. Step-time recording lets you move one-step-at-a-time through each and every step of the sequence, deleting or adding notes as you see fit. Some sequencers let you do only one or the other. Modular recording lets you create individual patterns which are then linked into songs (like drum machine programming). These patterns may usually be recorded in realtime or step-time. Note that step-time note entry, while useful, can often be simulated on real-time-only sequencers by simply slowing the tempo way down. Some sequencers allow all three types of recording.
There are two major thoughts on how to edit MIDI data: graphic editing, and MIDI event lists. Figure 1 shows a screen from Passport's Master Tracks Pro, which uses mostly graphic editing - you can even see, and modify, controller and tempo data graphically. Each note is represented by a solid block, with height indicating pitch and length indicating duration. This is a great way to see your music 'at a glance', but is not extremely precise - does that note start right on the beat, or a few clock pulses later? And does it include Aftertouch or other controller data?
"A reliable form of synchronisation to tape and other devices is a necessity if you want to get the most out of a MIDI sequencer in a studio context."
With this particular sequencer program, double-clicking on a note calls up a mini-screen that contains more precise information on the note - starting time, duration in clocks, channel number, velocity, and so on (Figure 2).
Figure 3 shows a screen from Mark Of The Unicorn's Performer. As you can see, each and every event - notes, Pitch Bend, Program Changes, etc - are all individually listed with a great deal of detail about each event. This is a lot of data to wade through, but this particular program lets you remove types of data from the display only (not from the track!) so that you can look at, say, note values without having to see Aftertouch information on screen. Some people feel event lists are less 'user-friendly', but do offer the advantage of a great degree of precision.
Graphic editing and event lists offer both advantages and disadvantages unique to each type of note representation. Nonetheless, many people claim that one particular method is the 'only way to go'. My theory is that right-brain oriented people, who supposedly think more graphically and abstractly, probably gravitate towards the graphic displays; left-brain people, who are reputed to think more analytically and linearly, are most likely those who prefer the event lists. I like having both available - graphic editing for doing broad edits, and event lists for detailed editing. And I might add that graphic editing of controller data is exceptionally helpful, particularly when using MIDI guitar synths since they generate huge amounts of data.
This is similar to punching with a standard tape recorder, but there are some subtleties when applying this technique to MIDI data. If you punch right after a Note-On command and don't programme anything to turn that note off, the note will stay stuck on indefinitely. A pre-roll feature is also handy, where you can programme a section of music to start playing a couple of measures before the punch occurs.
This is even better than regular punching. You simply programme the measure and beat in the song where the punch-in and punch-out should occur, and the sequencer takes care of the rest. No more missed punches! Other features may also be automated; with automatic rewind, for example, upon reaching the end of the punch the sequence will return to the beginning of the section to be punched.
You can radically change the feel of a song by, say, pushing the verse by one extra beat per minute or laying back the chorus by one less beat per minute. Being able to change the tempo (speed) of a song is very useful; unlike tape, speeding up and slowing down a MIDI sequencer doesn't affect the timbre of the instruments. Even better, some sequencers let you specify a start tempo, an end tempo, and the number of measures over which the tempo will increase or decrease. There is one caution, though. If you have a densely-recorded track that is on the verge of clogging the MIDI data stream, increasing the tempo may send the program over the edge (or at least into a nervous breakdown).
Some sequencers let you set the tempo by simply tapping a button; the computer reads the time between taps, takes an average (or just gives you the time for the last two taps), and converts this time period into a beats-per-measure reading that sets the sequencer's tempo.
If you want to drive your D50 instead of your DX7 from track 5 without having to do any re-patching, this option will let you do it. Simply assign the track to output its data over the appropriate MIDI channel for the chosen instrument, and you're ready to go. Almost all modern sequencers have this feature.
This MIDI message can be recorded at any point on a track to tell a device (or devices) tuned to the channel assigned to that track to change programs. Optimally, you should be able to edit and manually insert Program Changes if desired.
Probably more than any other single factor, this affects how much you will enjoy using a sequencer (or any other piece of software, for that matter). You want to do the least amount of typing or mouse-clicking necessary. A program that requires only single-letter commands or mouse clicks and lets you move a cursor around to make selections is better than one which makes you type in stuff like 'Save: Composition #1 in B-minor: Disk A'. The type of 'help' messages built into the program (if any) are also important. Some products aim to be so 'user-friendly' that help messages are sprinkled throughout the program, which can be handy while learning but add unnecessary clutter once you know your way around the sequencer.
Following the example set by the Apple Macintosh (and prior to that, Xerox), many computers have a standardised 'user interface' where various operations (save to disk, load from disk, edit, etc) are performed in exactly the same way for different makes of programs. This drastically reduces the learning time for a program, since you need learn only the differences between programs, not the similarities. The downside is that these standard user interfaces make extensive use of mouse operations, and you might get a little tired of all that pointing and clicking. As a result, many programs that rely on the mouse also provide keyboard command equivalents. For example, to start a sequence using a mouse, you would need to move the mouse and point to the Record option, then click on Record; some programs will let you start the record or playback process by simply pressing the computer keyboard's space bar (which provides a large, convenient target). Some accessory computer programs even let you define your own keyboard equivalents. For example, if you often find yourself quantising parts to 16th notes, you might want to define a keyboard equivalent which, when pressed, quantises all tracks to 16th notes.
Another important aspect is the clarity and appropriateness of the display screens - is related information grouped together? Do you have to constantly switch between screens? Is it easy to inadvertently erase tracks or, worse yet, an entire song?
Many sequencers store their data files in a common file format so that these files can be transferred among different programs. Probably the most common use of this feature is to 'export' a sequence from a program optimised for sequencing to a program optimised for printing music transcriptions. The process of turning a sequence into a printed score is difficult and requires very sophisticated software; in general, trying to make a program do everything - from sequencing to scoring to who knows what else - will require more trade-offs than optimising a program for a specific application. File transfer protocols are a particular boon to notation programs that do not support real-time recording.
Another example of the usefulness of file transfers involves M and Jam Factory, two composition programs marketed by Intelligent Music. The compositions created by these programs conform to the standard sequencer file format, so they can be exported to sequencer programs should you want to edit the algorithmically-created compositions produced by these programs.
It's much easier to remember a song title than a number. Naming tracks is also handy; that way you know which instrument is supposed to be driven from each track. The more letters the better - 'Bass' says a lot less than 'Bass: DX7 Patch 18'.
I don't know about you, but I always need a few beats before a song starts in order to prepare myself for the recording process. Being able to start playing or recording from any bar in a song is also useful.
"If a track is on the verge of clogging the MIDI data stream, increasing the tempo may send the program over the edge (or at least into a nervous breakdown)."
Being able to programme fast metronome times (ie. 16th notes) means that you'll still have a solid click reference if you slow the sequence way down when overdubbing.
Early sequencer programs for the Commodore 64 and Apple II could typically remember somewhere between 5,000 and 10,000 events (with Note-On, Note-Off, Pitch Bend, and so on being considered as 'events'). This isn't as much as it might appear to be, especially if an instrument generates a lot of Pitch Bend or Aftertouch data. In fact, using any dynamic controllers eats up tons of memory. One of the advantages of computers with 1 Megabyte or more of memory (Mac Plus, Atari 1040ST, etc) is the ability to store just about as many events as you're likely to produce under normal circumstances. If you buy a stand-alone sequencer, check if memory expansions are possible. The more memory, the less you'll have to access a disk drive to shuttle data back and forth between the disk and the sequencer's internal memory.
You should be able to check how much memory is left, as well as how much disk space is left.
It's really helpful to have some kind of indication of whether data is contained in a track. For example, with a sequencer whose main screen displays the name and number of each track, a symbol might show up next to the track number whenever data is playing back from that track. This shows at a glance how much general track activity is going on.
This takes the signal coming into the MIDI In port and re-transmits it over the channel of your choice. This feature is vital for owners of the original DX7, which transmitted only on channel 1; with a channelisation feature, you can leave your DX7 hooked up and use the computer to change its MIDI signal to any one of the available 16 channels. However, channelisation also saves a lot of time in any situation where you're using a single keyboard to control a bunch of expander modules, as you can direct the flow of MIDI data at the sequencer itself, rather than having to constantly reset channels on the master keyboard to access different expander modules.
This safety feature lets you protect individual sequencer tracks to prevent accidental erasure.
Although few sequencers let you record System Exclusive data as part of a sequence, many let you store this data (typically, patch data from synthesizers) on disk under its own file name, or perhaps as a separate sub-file of the main sequence file. With this feature, if you develop a specific set of patches to go along with a particular sequence, you can store the data for those patches on the same disk as the sequence - very convenient. When you come back to the sequence at a later date, you can reload the patch data back into the synthesizer, and you're ready to go. The System Exclusive storage functions found on most synthesizers work similarly to conventional patch librarian programs, albeit with less sophistication since the sequencer versions will be more 'general purpose'.
How would that piece of music sound played backwards? Or with the lowest notes substituted for the highest notes? How about creating a random two-bar melody line? Some programs (such as Dr. T's Keyboard Controlled Sequencer Version 1.6) now include functions that let you modify music according to some compositional algorithm (created either by the machine or by you).
This feature lets you call up sequences by typing keys on the computer keyboard, and is great when using a sequencer in live performance where you may want to vary the length of particular sections to allow for improvisation (Opcode's MidiMac Sequencer is designed with this type of application in mind). A variation on this theme is to call up a sequence from within another sequence, which opens up all kinds of outrageous possibilities. In fact, you could create a sequence that contains no notes of its own but simply calls other sequences (this is similar to the modular programming technique described earlier, but more flexible).
Make sure the person writing the manual is trying to instruct you, not impress you. If the first few pages make good sense, the rest probably will too.
When editing a sequence, some sequencers create a copy which you edit. This preserves an unedited version of the original in case you end up not liking the edited sequence as much (don't you wish tape recorders would save a previous track when you did an overdub?). When you have an edited version you prefer, you can then overwrite the original. Even if a sequencer does not offer this feature, you can usually approximate this function by saving a track to disk (program permitting) so that you always have a safety backup of the original, or by copying to another track and muting the original track as you work on the revised version.
It's fun to hear the sequence whiz by as you look for a part towards the beginning or end of a song.
This looks for a particular part of the sequence, or places you a certain number of measures into the sequence.
You should be able to bounce data from one track to another, and combine tracks together. In many cases, tracks that are bounced together cannot be unbounced. However, there are some exceptions to this (see the section on filtering data).
Typical bouncing applications include playing individual sections of a complex part for one instrument over several tracks, then bouncing them all down to create one composite part on one track. Another option is to play several different solos on multiple tracks, then use punch-in and punch-out to erase those parts of the solo you don't like and bounce the sections that remain into one track, thus freeing up the other tracks for more instruments.
Some sequencers let you offset one track from another with respect to time, in single clock pulse increments. As you might expect, sequencers with higher resolution (and, therefore, more clock pulses per quarter note) let you shift by smaller amounts of time. If a sequencer doesn't have the ability to offset tracks, you can often fake it by inserting a short bar (1/32nd note, for example) in front of a copy of a non-offset track. Some programs have 'time offset filters' rather than an overall track offset, which let you change only selected parts of a track.
Track-shifting also lets you remove tiny timing differences between devices to obtain perfect synchronisation, as well as create special effects such as doubling, chorus, echo, and canon harmonies by copying a track to another track and offsetting the new track by the desired amount compared to the old track. Another application is to create different 'feels' with this function; a track that is a little ahead of the beat gives a very different teel compared to one that is a bit behind the beat. (For more information on the subject of creating different 'feels' with sequencers, see 'The Feel Factor' in SOS October '87.)
So you can't sing that song in D# after all? Then transpose it until you hit the right range. Some sequencers offer both individual track transpose and overall song transpose.
Note that if you are driving a drum machine from your sequencer, you will probably not want to transpose the track containing the drum data, since each note corresponds to a specific drum, and transposing the track will trigger entirely different drums. For this reason, some sequencers (eg. Akai ASQ10) let you designate one or more tracks as non-transposable.
"Many sequencer programs now include functions that let you modify music according to some compositional algorithm."
The filtering function eliminates selected data from the MIDI data stream - for example, all notes within a certain pitch range, or all notes associated with a particular MIDI channel (assuming that a track can hold data from several MIDI channels at once, which is often the case). In many cases, this filtered (cut) data can be 'pasted' to another track.
As one example of how to use filtering, suppose you played left and right hand parts on a single keyboard and wanted to split off the left hand part to a different keyboard. You could copy the track and filter the low notes from the original, thus sending the right hand part to one instrument; the next step would be to filter the high notes from the copy and send the low notes (the left hand part) to a different instrument. (I first saw this particular low note/high note filtering option in Roger Powell's Texture program.) Filtering also lets you 'unmerge' merged parts, assuming that those parts have some distinguishing differences (eg. occupy different note ranges, are recorded on different MIDI channels, etc). For example, assume that you're transferring sequencer parts from one sequencer to another in real time and that one sequencer contains parts recorded on eight different tracks, corresponding to eight different MIDI channels. Playing this data back into another sequencer will often end up sending all this data into a single track. The Strip Data (filter) command can then 'strip off' the data by MIDI channel, and paste the data associated with a particular MIDI channel into its own sequencer track.
Yet another use for filtering is to conserve memory. Aftertouch, Pitch Bend, Continuous Controllers, and similar real-time data take up a fair amount of sequencer memory. Filtering out data that is not used in a performance makes more memory available for other functions and sequences.
Filters aren't just for conserving memory, however. If you record several tracks full of controller data, you can quickly end up with a song that is not playable through a single MIDI Out socket. Imagine what happens when our complex tracks are recorded singly then played back together... MIDI data stream clogging is a real problem, and filters are about the only good cure.
There are two main types of filters: one blocks on record (for example, it ignores all controllers) while another removes data from already-recorded tracks. This is similar to record vs. playback auto-correct - record filtering is useful when you know you don't want to record a particular type of data, but once filtered, you can't get it back again. Note that playback-only track filters can take a lot of time to remove data (if you're going to filter out 300 bars chock full of Aftertouch data, it might be time for a coffee break). Many sequencers implement both types of filtering.
Mute lets you selectively silence tracks while recording or playing back. Solo selects one track but silences all others. Some programs let you mute/unmute/solo during playback, while others let you make these changes only while the sequence is stopped. (Note: the more features that are available while the record or playback process is in progress, the better.) One good use for muting and soloing is to record four or five different solos, and listen to each one individually before deciding which one to keep. For live use, this means you can play different solos at different performances so you don't get bored with hearing the same sequenced part over and over again.
According to Anderton's Law of Hi-Tech Equipment Purchasing, never buy anything that says 'Version 1.0' unless the company will upgrade you to the next software revision for a reasonably low fee. Initial software offerings often have bugs; by the time you get a few versions into the program, the bugs are pretty much all gone. Note that software upgrades are generally much more common, less expensive, and easier to install if you have a computer-based sequencer as opposed to a built-in or stand-alone type.
Copy protection is designed to prevent people from making unauthorised copies of programs, but it also prevents legitimate users from making backup copies to guard against damage to a master disk. Also, you sometimes can't use a hard disk with copy-protected software because the software can't be copied over to the hard disk. Although copy protection is a pain, remember that writing software takes a tremendous amount of time and money, and people who rip off a program are no better than shoplifters or car thieves.
Copy protection takes several forms. Sometimes, a hardware 'key' (called a 'dongle') is provided that you plug into one of the computer's ports. Unlimited backup copies can be made but they won't run without the dongle inserted in the computer. Don't lose the dongle, though, or you're in trouble! Another method uses the 'key disk' principle. Again, you can make unlimited backups; but upon booting the backup version of the program, you will be asked to insert the original master disk - the key disk - into any available drive so that the computer can verify your ownership of the backup. Yet another form of copy protection (which happily is on the wane) is to disallow copying of any kind. You are therefore forced to always run the program from the master disk and, of course, the more you use the master disk, the greater the media wear and the greater the odds of damaging the disk. This highlights one of the advantages of the key disk approach: since you need insert the key only for a few moments on booting, it should last a very long time.
Check the manufacturer's policy towards providing backup copies in case your master gets trashed. You shouldn't have to pay list price twice for the same program, although the level of manufacturer paranoia varies widely.
Recognising that most professional users rely on hard disks in order to save time, some software sequencers provide utilities that let you install the program on hard disk. Different programs implement this process in different ways; check with the manufacturer about the hard disk installation procedure before you lay down your hard-earned cash, and also investigate how easy it is to de-install the software should you need to reformat your hard disk.
A sequencer should have at least two synchronisation options: internal and external. With internal synchronisation, the sequencer has its own tempo reference and this timing data will also appear at the sequencer's MIDI Out connector so that other devices can synchronise to the sequencer. When set to external, the sequencer will expect to receive MIDI timing data at its MIDI In connector, possibly from a drum machine or other sequencer.
The situation described above assumes that in external mode, some source of MIDI timing data is available. If you need to sync to tape, this implies that you have a device capable of storing MIDI timing data on tape, which can then be read by the sequencer. Devices such as the J.L Cooper PPS1, Tascam MTS30, etc, can provide this function. If SMPTE timecode is the master timing reference recorded on tape, then a SMPTE-to-MIDI convertor can convert the SMPTE data into MIDI timing signals suitable for driving the sequencer. However, not all sequencers require an external device to do tape sync, and instead include a sync-to-tape option (usually some proprietary system that is not necessarily compatible with other sequencers) similar to the systems used with drum machines. This will usually be in addition to the two main sync options mentioned above.
If you are planning to drive the sequencer from MIDI data, make sure that the sequencer recognises MIDI Song Pointer data, as this will allow the sequencer to autolocate to devices that produce Song Pointer data.
Finally, some sequencers can read SMPTE directly, thus obviating the need for a SMPTE-to-MIDI convertor for synchronisation to tape. Most professional audio/video facilities are SMPTE-based, and having a sequencer that can read SMPTE directly is considered the best way to go. Of course, you pay extra for the privilege. In any event, a reliable form of synchronisation to tape and other devices is a necessity if you want to get the most out of a MIDI sequencer in a studio context.
© 1988 Mix Publications Inc and reprinted with the kind permission of the publisher. (Portions of the above article are excerpted from the book MIDI For Musicians with permission of the publisher.)
About the author: Craig Anderton is editor of Electronic Musician magazine and author of several books on musical electronics, including MIDI For Musicians and The Electronic Musician's Dictionary (available from Music Sales). He is currently working on a book about choosing, using, and occasionally abusing samplers, and plans to release his next solo album some time prior to the turn of the century.