When your sequenced compositions start to become more complex and adventurous, you'll find it useful to know about the more advanced facilities provided by software sequencers — Stephanie Sobey-Jones presents a tutorial to help you get the best results.

Last month, I put together a short song using the very basic features of a software sequencer. This month, I'm going to look further into the use of some of the facilities and techniques which are designed to make life easier once your compositions start to get more adventurous. The features and terminology are once again based around the Cubase and Notator families, but could of course apply equally well to other packages.

Arrangements

Most songs can be broken down into several basic sections — Intro, Verse, Chorus and so on — which will perhaps be repeated and varied to form the overall song structure. Last month I introduced the concept of working in sections or patterns, and in order to make best use of your sequencer, it is worth a further look at how recorded material can be organised to give maximum flexibility, whilst at the same time providing scope for easy experimentation within the overall formal structure. Recording a part in a single 150-bar 'take' is an impressive feat, but it makes any editing and additional arrangement difficult, to say the least, so a bit of forward planning in the initial stages can save a lot of aggravation later on! Most sequencers allow you to construct individual 'patterns' of material in one way or another, and it is important to establish separate patterns for the sections which will form your verse, chorus and so on. A pattern can be as long or short as you like, and can contain any number of tracks of rhythmic and melodic phrases (subject to the limitations of the particular sequencing package). Furthermore, each track does not have to be recorded as a single 'chunk'; for example, the drum track of your verse may consist of a 4-bar pattern followed by another four bars with fill. You may then plan to extend this by copying the material, inserting different fills, and so on. You may even want to use material from another pattern, or from another Arrangement window (Cubase) in order to construct your track. This is fine, so long as you end up with a complete section which can be freely moved around, preferably as a single unit, when it's time to experiment with the actual arrangement.

Notator users can assign a different pattern for each section of the song, and build up tracks within those patterns by recording new material and by 'segment copying', which entails copying segments of existing material to new positions. Cubase users will end up with several 'parts' along the track, and it is a good idea to use the Group Track facility to organise those parts (and the tracks as well, if you wish) into one section called Verse (or whatever) which can then be moved around as a single block. Figure 1 shows how a typical section or pattern can consist of a number of small ideas on each track, organised into an easily movable unit. Track 16 contains the material from all the above tracks.

Ghosts And Other Weird Things

Within a pattern you may wish to 'double' material and perhaps have several instruments play the same part. An economical way to do this is to create 'Ghost' tracks or parts. These are parts which use the notes and MIDI information of the original part, but which can play back on different MIDI channels and with different parameters (such as Velocity or Program number — see later). The advantages over making a straightforward copy of the part is that Ghost parts, having no notes of their own, do not use up any more memory. Furthermore, any alterations made to the original part are automatically reflected in all the Ghost parts belonging to it. (Altering the contents of a Ghost part, however, will transform it into a 'real' part.) Figure 2 shows a combination of real and ghost parts within a song section.

Several voices playing the same material may result in slight differences of timing. This is because not all instruments, or even instrumental voices, have the same reaction time, so a sound with a slow attack will naturally appear to lag behind something else whose attack time is faster. This may generate some weird and wonderful effects, but if you don't like the result, try experimenting with some Delay. Adding a negative value to the track advances the start time of each note, whilst a positive value delays or slows down the start in relation to the other tracks. Setting different delay times over several tracks playing the same material can produce a thicker sound which may well suit the character of the song. (Warning: sometimes a delay setting will cause a note to move outside the start of a pattern or part, causing it not to sound. Your sequencer manual will suggest how to compensate for this.) Having put together a few basic sections, you can now experiment with the organisation of the overall song, adding any new sections — a 'break' for example — as appropriate. In an educational situation, this is an excellent way of teaching musical form, as it shows graphically how a composition is put together. Manipulating one or two pre-recorded patterns will demonstrate in practical terms a simple A/B/A form, or even the basic mechanics of a Rondo if you feel adventurous.

Hassle-Free Variations

Although the general outline of each verse and chorus repeat may be similar, you may well want to create some variety of rhythm, melody or instrumentation in each appearance of a particular section. This may be purely a matter of doubling the melody line on another instrument, or it may involve some fairly radical changes to a track. For example, imagine that Verse 2 is going to retain the rhythm, bass and chord tracks of Verse 1, but the melody track will be slightly different. One way of tackling this is to copy the three identical tracks into a new Pattern, then record the new melody line on top. This will work, but a more economical method is to make use of the Mute facility.

The trick is to record all the different options into the same pattern and then use the mutes to make sure that only the ones you want to hear are playing at the right points in the song. For example, you might have a brass track in the 'verse' pattern but keep it muted until the start of the second verse where the brass will be allowed to join in.

The Arrange pages of both Notator and Cubase allow you to mute a particular track or part at a set point in the song, so by recording the new Verse 2 melody line into the existing Verse 1 Pattern (Notator), or alongside the appropriate Group Track (Cubase), any unwanted track can be silenced when required. (In Cubase, the particular Group may have to be 'unpacked' at this point in order to access the unwanted track. Alternatively, if you have not organised your material into 'Groups', you can just mute the relevant track or part.) Unfortunately, some of the entry-level sequencers do not have this particular facility, so an alternative is to Record your mutes. The most common way of doing this is to select an empty track, start recording, then click the track mutes on and off when required. Or you may be able to pre-program and store several mute settings, which can be recalled at the desired points. Figure 3 shows a recorded Mute track as part of a repeated Verse. Note that certain tracks appear muted at this point in the song — the information in this example is recorded on Track 16.

This may sound involved when written down, but all you're really doing is creating alternative tracks, which provide the variations you need within a pattern, and then using the mutes to turn off the ones you don't need to hear. The same technique can be used if you have an instrumental solo which occurs over a verse pattern; every time the verse plays, the instrumental solo track is set to mute, except on the one occasion when you want to hear it.

Once the general outline of your arrangement is established, other important areas to consider are use of instrumental voices or patches; tempo settings/changes; and dynamics and expression.

Program Changes

These are MIDI messages/events which control voice/patch changes on a keyboard or sound module (or programs in a multi-effects unit). As with other types of MIDI information, there is an available range of values from 0-127, '0' corresponding to the first voice on the keyboard. (Just to confuse things, some instrument voice banks start with '1' rather than '0' so that selecting Program 15 will actually give you number 16 on that particular instrument.) As I briefly mentioned last month, using the Program Change facility on a sequencer is a quick way of trying out different combinations of voices on the various tracks of your song. But, more importantly, once you have decided on a suitable combination, those Program numbers can be saved as part of the song so that every time that song file is loaded, your keyboard/sound module will automatically set itself to the correct voices once it receives the relevant information from the sequencer. In a learning situation, where lots of people will use the same instruments, this will save a great deal of potential frustration; even if it is a short piece using a single voice, don't bank on being able to remember exactly which of the 128 (or whatever) sounds was actually used.

Not all keyboards number their voices with a straightforward numerical system, so — if your keyboard manual does not already contain one — it might be a good idea to make a chart showing the translation of your particular voice numbers to standard Program Change numbers. (My own students find this invaluable, particularly when working with two or three instruments simultaneously.) Of course, you're not limited to using one particular voice for the entire duration of a track — Program Changes can be inserted at any point in the music. (Cubase users can add them to the Arrange Page or the Controller Display in the Key Edit page, whilst Notator users will add them to the Event List.) One thing to bear in mind is that you should try to insert the actual change where there is a slight gap in the music; your keyboard cannot actually play a note whilst it is changing from one voice to another, so you could experience a slight 'hiccup' if you insert the change at the start of a note.

General MIDI

Until recently, problems have existed because of the fact that saving a series of Program numbers and changes in a song doesn't necessarily help if that song is later replayed on a different instrument from the one on which it was composed. It can be very annoying to hear Synth Bass, Acoustic Piano and Strings on a Korg keyboard manifest themselves as Koto, Organ and Telephone respectively on a Roland instrument! The basic idea behind the relatively recent General MIDI protocol is that voice/program settings will transfer from one instrument to another (provided the instruments are General MIDI compatible, of course.) In the General MIDI system, there are basically 128 voices or patches assigned to set numbers so that Program 1, for example, on a General MIDI instrument is always Acoustic Piano. Similarly, there is also some uniformity with MIDI channel mapping, so that drums are always accessed on Channel 10, with the same sounds attached to the same notes, thus preventing the current disasters which can arise with different makes of keyboard producing wildly different drum sounds on identical sequences of MIDI notes. Of course the sounds won't be exactly the same, as all instruments have their own character, but at least you'll have the right type of sound.

Changing Tempo

The basic tempo of a song can, of course, be set by selecting the appropriate value in the Tempo display. However, you may wish to have specific tempo changes at set points during the song, and these will need to be added and stored along with all the other bits of information. Tempo changes can be manually inserted — using Mastertrack in Cubase or the Event List in Notator — where you enter the different values at the required places. As soon as this point in the song is reached, the sequencer will switch to the new tempo. However, if you want a gradual tempo change over a number of bars — slowing down at the end, for example — a more practical method is to record the changes onto an empty track. (This facility can be enabled in the appropriate pulldown menu.) Start recording and alter the tempo values with the mouse as you go along. This way, the music tends to have a more life-like 'feel', and you can still manually edit the recorded information if you don't get it quite right first time!

Dynamics and Expression

Last month's basic project made use of the Velocity facility for adjusting the relative balance between tracks and editing extremes of volume on individual notes. This function is used to alter the dynamics of a part or track by adding or subtracting a certain value from the existing velocity for that part. One thing to bear in mind, however, is that some older instruments are not velocity sensitive and therefore will not respond to the value changes. Similarly, all velocity sensitive instruments do not necessarily respond to the same degree, so that a particular value change might lead to wildly contrasting results on different instruments.

A further method of 'streamlining' the dynamic range of a track or part is to use the Compression facility. This can be used to tone down any notes which might stick out, to produce an even effect over the track as a whole. Velocity basically affects the Note On and Off messages in a part, but other forms of expression and dynamics can be determined by MIDI Controllers. Controllers affect various functions of a MIDI instrument; for example, the use of MIDI Volume (which is not the same as Velocity) will change the overall volume of whichever instrument receives that controller information. This is a good way of balancing the relative volumes of different tracks and producing a basic 'mix' (provided that your instrument can receive MIDI Volume — check the manual). One thing to watch, however, is that the more you turn down the instrument using MIDI, the more background 'hiss' you'll hear so don't work with very low velocity settings if you don't have to!

Stephanie Sobey-Jones is Education Specialist at Systems Workshop, and a part-time Lecturer in Music Technology. She is currently running a series of courses for Primary and Secondary teachers, covering different aspects of technology in the Music curriculum.

Other types of controller which can affect the amount of expression generated into a piece of music include Breath Control, use of the Modulation Wheel and the Foot Pedal. Each type of controller has a specific control number and generates MIDI information which can be edited in an Event List (Notator) or Edit Window (Cubase). Controllers and other expressive non-note events such as Pitch Bend and Aftertouch often generate huge amounts of information which can take up valuable memory space and sometimes get in the way of the notes themselves. So anything which is automatically generated by your keyboard, but which is perhaps not required can be filtered out with the appropriate type of MIDI filter either before or after recording (check your sequencer manual).

Learn As You Go

In many ways, working with a sequencer is like exploring a bottomless pit, and most users only utilise a small number of the available features. This is no bad thing, as you can start getting results very quickly and then allow your knowledge to growas you start exploring new features. People tend to attach varying degrees of importance to certain facilities — a classical composer may use some features more frequently (and in a slightly different way) from a composer of dance tracks. For example, I like to see my work displayed in standard musical notation and I do most of my editing in a Score Edit page. On the other hand, I know several people who never use this facility at all, preferring different editing pages. Whatever the case, a sequencer is designed to make composition and arrangement both flexible and relatively painless and we have looked at some of the common facilities and techniques which can be applied to many of the current sequencing packages. However, as they say, the sky's the limit, and there's always much more to discover!