From a delicate shimmer to an earthquake tremor, harmonics, tremolo and trills can add vibrancy and drama to your string arrangements. Find out how to make the most of them in this SOS masterclass.
Welcome to the third part of my string-arranging series — and if you missed the first two, they're available online at /sos/jun12/articles/string-theory-pt1.htm and /sos/jul12/articles/string-theory-2.htm. This month, I'm joined by the talented composer, arranger and orchestrator David William Hearn, whose extensive CV includes (deep breath) music for the TV shows Dancing On Ice, American Idol and The Voice, countless television ads, programming orchestral mock-ups for the movies The Chronicles of Narnia, 2012, Lay The Favourite and Paul, for tours and albums by Katherine Jenkins, George Michael, Westlife, Kylie Minogue and Shirley Bassey, and for producers such as David Foster and Phil Ramone.
The inexhaustible Mr Hearn has contributed some great masterclass tips to help spice up your MIDI orchestrations and improve your workflow. Although intended primarily for experienced orchestral samplists, these tips can be applied by anyone with a modicum of programming skill. In addition to these useful pointers, I'll continue to share my personal views on the noble art of string arranging, including extracts from arranging jobs that have come my way over the last few years.
When I started out arranging for strings, I soon realised that keyboard and guitar chords rarely work well when transcribed note-for-note for a string section. The reasons for this are fairly obvious: a keyboard patch or strummed guitar makes a homogenous sound in which all notes blend together into a unified whole; a string ensemble is far less homogenous, since it features different types of instrument, as well as multiple players performing each note of a chord. The resulting sound is therefore richer, more complex and more expressive than a keyboard or guitar could ever be, so in order to avoid creating an overly dense sound, we need to think carefully about what notes in a chord should be assigned to the string players.
A simple illustration is the pair of chords you see in diagram 1: this classic A-minor to C-major movement sounds perfectly satisfactory played on keyboard or guitar, but when performed by a string section literally as written, it sounds too full. In order to let some 'air' into the string chords, we need to thin out the voicings. A good starting point is to identify any duplicated notes: as you can see in diagram 2, the top notes (A and C respectively) are doubled an octave down (marked in red); the second-from-bottom notes of E and G (shown in blue) are also duplicated an octave higher.
Although they add warmth and richness when played on keyboard and guitar, these mid-range octave doublings have a built-in musical redundancy — that's to say, they don't add any new harmonic information to the chord — and so can be safely omitted when arranging for strings, leaving us with the pared-down voicings in diagram 3. Alternatively, if you miss the comforting presence of a fifth at the base of the chord, you could try the alternative version in diagram 4, which reinstates the low fifths and omits the second-from-top notes.
Benefiting from wide, similarly proportioned intervals, the last version sounds elegant and would work a treat with its notes assigned to (from the top down), first violins, second violins, violas and cellos. The difference in sound between the diagram 3 and diagram 4 chords is rather subtle; it would be an excellent ear-training exercise to alternate them slowly and repeatedly, enabling you to accurately discern their different flavours.
As I remarked in my first article, two months ago, one of the interesting things about writing string arrangements nowadays is that songs usually arrive with a 'strings demo' attached, more often than not created by the song's composer. A set of demo string arrangements of unusual quality came from a client called Rob Reed, a genial Welsh keyboardist and TV composer who was in the process of recording his own long-form symphonic-rock concept album. (I realise that to utter the last sentence in a UK music magazine back in 1977 would possibly have led to death threats, but in the post-post-punk, chilled-out climate we now enjoy, I feel safe to tell you that the music on this album is of an impressively high standard.)
The composer, who earns his living working with live players as well as creating MIDI mock-ups round the clock, knows his musical onions, and had used some top-end sample collections in his string demos. The libraries included LA Scoring Strings and Miroslav Vitous' Symphonic Orchestra String Ensembles. The arrangements, therefore, sounded very good, and the musical ideas worked so well that I didn't feel the need to elaborate or rewrite much. For me, this meant that (unusually) the job was mainly concerned with the details of the orchestration, rather than dreaming up entirely new arrangements from scratch.
One song's intro was particularly effective, being a quiet, sustained chord with an eerie, dreamy atmosphere, played on high violins. I couldn't exactly identify it by ear, but thought it might be a major chord with some added intervals. Opening the MIDI file revealed the dense voicing you see in diagram 5. I'm assuming that Rob played it with two hands, rather than cheating and using his feet.
Orchestrating this dense event for real players required thought: while the sampled demo version sounded fine, I wasn't convinced that we needed the players to perform all nine notes. I feared that would sound too full and undermine the attractive, transparent quality of the demo. The solution was to omit the three pitches marked in red: the top note was so quietly played on the demo as to be practically inaudible, and I also felt we could live without the two Eb notes, since the Gb was already adding a nice fourth suspension to the lower Db-major triad. In the end, I orchestrated the chord as depicted in diagram 6. You'll notice that the top note of the violas is interposed between the second violins' two pitches, in order to help unify the two sections.
I should point out that although this chord contains a fair number of closely-positioned notes, the concerns about over-density noted earlier were not a problem. This is because, firstly, its high pitch precludes an excess of warmth or muddiness in the sound and, secondly, since it's played unaccompanied at the top of a piece, it's not competing with any other arrangement elements.
The transparency and fragility of high, quietly played violins can add a lovely atmospheric quality to a song. I had a chance to explore this area when I worked with the Liverpool band Anathema on Falling Deeper last year. The album is a collection of re-interpretations of earlier recordings, some dating back to the early '90s when Anathema were a doom metal band. Although a certain heaviness remains, the deathly overtones of yesteryear have been supplanted by a tender, more reflective, emotional and somewhat mystical approach tinged with a wistful, Celtic melancholy — in other words, the perfect setting for strings!
As several of the songs on Falling Deeper were played on unaccompanied piano, there was plenty of scope for lighter sonorities. One example was the song 'Everwake', a charming 3/4 ballad featuring guest vocalist Anneke van Giersbergen. The opening verse leads to an eight-bar instrumental break in A-minor, played on solo acoustic guitar; I orchestrated it with violins playing a sustained, high-pitched fifth interval of A and E (the lower note pitched an octave and a sixth above Middle C).
Note-wise, this is about as simple as it gets; clearly, adding a high fifth based on the home key of the music isn't going to affect the harmonic make-up of the piece one way or the other. However, what I was exploiting here was not note content, but sheer sound — the light, bright, clear and pure sustained tone of the violin's top string, stroked delicately into sensual motion by a horsehair bow. With eight violins playing the high note and six tackling the lower, this simple interval had a subtly electrifying effect on the track.
If you want an even more delicate effect, try harmonics. As every guitarist knows, these are obtained by playing a string while touching it lightly at the octave with the fingertip, producing an ethereal, quiet chime. String family instrument players can do this too, and in fact can go one better, being also able to play 'artificial' harmonics by pressing the string down and lightly touching a fourth interval above it, producing a harmonic two octaves above the stopped note. Using this method, you can write melodies consisting entirely of harmonics, although, due to the difficulty of the technique, it's sensible to keep the part slow and simple!
A famous example is the solo violin melody in the introduction of the second part of Stravinsky's Rite of Spring. Played entirely on harmonics over an eldritch series of eighth-note woodwind chords, this unearthly, disembodied tune sounds as if it's coming from another universe. God knows what the audience of 100 years ago must have made of it when the piece was premiered... oh, I remember now — they rioted.
I was able to feature violin harmonics on another Anathema track, the piano-based instrumental 'I Made A Promise' (originally released in song form as 'J'ai Fait Une Promesse'). After an intro featuring some mournful, full-strings chords, the piano picks out the vocal theme accompanied by the first violins playing quiet, sustained harmonics (see diagram 7). With each note taken by two players, this created a delicate, ethereal sound. As you can see, harmonics are notated by placing a small circle above the note. Artificial harmonics are often notated with diamond-shaped note heads, but in this case I left it to the players to work out how to play them!
Canny media composers will be wise to the fact that the strings' tremolo style (a very fast repetition of a note, usually played with the bow tip) is commonly used to denote suspense. When executed by an entire string section playing a diminished chord, this technique constitutes the classic, scary 'Behind you!' film music cue, and when played by double basses it can produce massive-sounding, shuddering vibrations on a seismic scale. There is, however, a more subtle application of tremolo that avoids such clichés: we in the trade call it 'div trem' (short for 'divisi tremolo').
The basic idea is that, rather than everyone playing tremolo, a section divides in two, with half the players doing the tremolo bowing while the others play in the normal arco style.
This proved to be the perfect approach for an instrumental piece called 'Belle de Jour' on Steven Wilson's Grace For Drowning 2011 double album. The music is light and romantic, reminiscent of '60s film soundtracks; Steven had specified the tremolo delivery, and had augmented the sampled strings on his demo with a real violin, played by himself in vigorous tremolo style.
Diagram 8 shows a short extract from the 'Belle de Jour' string arrangement. In the studio, we tried various ways of implementing the 'div trem' approach — and, as I recall, we ended up with only the first violins doing it, which added a very nice shimmering effect to the high melody line. The basic chord sequence oscillates between A-minor and Ab-major; I wanted to introduce some additional harmonic movement within the chords, so gave the cellos a melody line that adds a low major seventh to the Ab-major chord in bar three. I also deliberately introduced some harmonic confusion in bar six, where the melody line is briefly underpinned by the four adjacent white notes A, B, C and D. This breaks all the rules of harmony, but it sounded good to me!
The crescendo and diminuendo dynamic movements in this extract are a very important aspect of string playing. I noticed when I first worked with the superb London Session Orchestra players that they naturally added subtle, expressive volume swells and fades even though the parts didn't specify them. Coming from a rock background that recognises only two dynamics — loud, and louder — this was both a pleasant surprise and a valuable lesson. I was also amazed that the players collectively executed these expressive gestures as if they had performed them hundreds of times, when in fact they had never heard the music before — surely the sign of an ensemble who play a lot together.
Trills — love'em or hate'em? A vocalist of my acquaintance once asked me not to play any on the backing track of a song we were working on. When I asked why, he replied, "because I always think musicians only play trills when they can't think of anything else to play.” That certainly wasn't true in my case, but I can see how in certain circles a trill might be considered a bit namby-pamby — however, the technique can be very effective when used sparingly, and in my opinion sounds particularly good on strings and woodwinds.
I used trills on a string arrangement for Anathema's 'Sunset of Age', also from Falling Deeper. The original version had appeared on the 1995 album A Silent Enigma, back in the group's doom period, and consequently it's a slow and heavy, guitar-dominated affair. The re-interpreted, 21st-century version is more lyrical and romantic, but still retains an ominous quality.
One of the song's main themes is a big, half-time, see-sawing E-minor bass riff which goes up to F in the second half of every fourth bar. The chords played over it are E-minor, A-minor, C-major 7 over E (one bar each), and in the fourth-bar turnaround, two beats each on C-major over E and F-major 7. At the outset, the strings play the changes as sustained chord pads; in the second verse, I introduced more rhythm, by making the higher strings play half-time off-beats, almost like slowed-down reggae, while the cellos doubled the on-beat bass line. When it came to the instrumental, I wanted to hear something different playing over the chord sequence, so wrote the four-bar passage you see in diagram 9.
The hallmark of this particular section is the emphatic staccato rhythm played by violas and cellos; it's actually an orchestrated version of something I played on keyboard, using a combined strings-and-woodwinds patch from Project Sam's excellent Symphobia 2 library (the woodwind element is pretty subdued, but the marcato strings have a good, fierce attack).
To add more angst, I wrote a high trill note in bars 3 and 4, introduced by a dramatic run-up; to finish the trill with a bang, I appended a manic-sounding slide down from a high-pitched E-minor triad over the last two beats. The three notes in this chord were split between 14 violins in a 6:4:4 ratio. As you'll recall, the instruction 'div' (short for 'divisi') in the score tells the players to divide the notes of the chord between them, rather than attempting to play the entire chord on their instrument (theoretically possible, but usually inadvisable).
'Sunset of Age' plays out on a long, histrionic guitar solo (some things never change), doubled in places by high violins. The solo peaks on a high sustained note, at which point I pushed the boat out and wrote the chordal trills you see in diagram 10. The strings are basically moving in a kind of counter-rhythm between a D-major and E-minor chord, but the notes are almost irrelevant: the significant fact here is that each pitch is played as a loud trill, giving the strings a great, vibrant energy.
I hope some of the techniques, musical extracts and technical tips outlined above will be of use to you. I suspect that many musicians would like to try their hand at string arranging but are intimidated by what seems like a complicated process. Fortunately, we now have excellent string sample libraries that you can experiment with before calling in the session musicians. When you do pluck up courage to book them, you'll find that the new breed of studio string players are very helpful, and will happily advise you on the best way to perform your arrangements.
In my next article, I'll feature some of the less subtle aspects of string arrangement, and examine how a string section can hold its own against a rock band in full flight. Until then, I wish you a pleasant and productive month of music-making.
'Lily' (music by Rob Reed) is from the 2012 album Kompendium (www.magenta-web.com). 'Sunset of Age', 'Everwake' and 'I Made a Promise' (all composed by Daniel Cavanagh) are from the 2011 album Falling Deeper by Anathema (www.anathema.ws). 'Belle de Jour' (Steven Wilson) is from the 2011 double album Grace For Drowning by Steven Wilson (www.gracefordrowning.com). Thanks to the composers for permission to use extracts.
It's important to imbue digital mock-ups with as much semblance of human expression as possible, which to me means constantly varying dynamics and vibrato levels in a musically pleasing and realistic way. It can take many hours of study to learn when and how live instrumentalists vary their vibrato, but it can all add an extra dimension to your mock-up, and perhaps hold the illusion long enough to get a music cue approved by a director!
I invariably use the mod wheel (MIDI CC#1) as a dynamic controller, which means that in my orchestral templates, MIDI CC#11 (traditionally used as the Expression controller) is freed up for other functions. I've found it's very effective to use an expression pedal to crossfade between non-vibrato and vibrato versions of instruments: it works very well on solo woodwinds such as the flute, and also on string ensembles. Here's how to set up a vibrato crossfade using the Vienna Instrument and 'Flute 1' from Vienna Symphonic Library's Woodwinds I collection.
Inverting the second instrument's Expression curve has the effect of making the non-vibrato flute get quieter when you press down your expression pedal, while the vibrato flute simultaneously gets louder. Pull back the pedal, and the reverse happens. This means you can start a note without vibrato and introduce it at will via the pedal; a nice, expressive technique, and a big improvement on using 'progressive vibrato' instruments in which the vibrato always kicks in at the same point. And, of course, while varying the vibrato intensity, you can also add volume swells and fades with the mod wheel…
One additional tip: when working with a multiple computer setup, I recommend using VSL's Vienna Ensemble Pro host, a very cool piece of software which makes it easy to control multiple Vienna Instruments and VST/AU plug-ins over Ethernet. It's also a very good option for allowing you to use large sample libraries inside DAWs that still aren't 64-bit, such as Pro Tools. David William Hearn
Just as a good knowledge of instrument balance, section balance and range is essential when setting up a homogenous and realistic-sounding orchestral template, technical consistency amongst MIDI controllers, keyswitches, transposition and crossfade variables are just as important to provide an efficient workflow. One of the best ways to provide consistency throughout your template and across all your sample libraries (whether reprogrammable or not) is to use VST Expression — a Cubase-only feature I've come to rely on heavily to get work done on schedule, and without losing my sanity.
VST Expression (introduced by Steinberg in Cubase 5) is a very powerful tool, and a great help for composers on tight deadlines who need to compose, provide mock-ups and print out the parts for the players. It takes the concept of keyswitches and presents them in a musically relevant way, via Cubase's Articulations lane and Key Editor list, and as musical symbols within the Score Editor. As well as providing a great, at-a-glance overview in both editors, having the symbols show up in the score can save hours of tedious manual inputting later on.
Using this system, I can designate a keyboard note of (say) C0 to access arco (bowed) samples, C#0 for tremolo, D0 for staccato, D#0 for pizzicato, E0 for marcato, and so on. These notes don't have to correspond to the actual keyswitches on your sampler. If you like, you can remap them to different output notes, make them change the track's MIDI channel, or even send out CC messages. This is a godsend when working with samplers such as Play, which don't allow you to alter or create keyswitch patches; it also avoids the hassle of having to edit your sampler patches, which, firstly, isn't always possible and, secondly, can cause headaches when updating.
Equally cool is the fact that the input keyswitches don't appear in the sequence or score as actual notes (thereby avoiding tedious clean-up operations), and that you can easily change the key of the music without worrying about affecting the keyswitches. Another nice touch is that Cubase chases the keyswitches, so you'll always hear the right articulation, regardless of where you start playback. All these little benefits massively speed up the mock-up process, making it a lot easier and more enjoyable.
Additionally, to provide consistency between MIDI controllers, I often make use of Cubase's track-specific (or 'local') input transformer. This means that I can convert, for example, CC1 to CC11 on some tracks but not others. This is essential when working with multiple sample libraries that use different controllers to achieve the same thing, and is a perfect complement to consolidating keyswitches via VST Expression. Once it's set up, I only have to remember one main set of keyswitches and controllers for all the tracks in my template — regardless of which sample library it's triggering. This is great for my brain, allowing a bit more space in there to remember other things, such as my girlfriend's birthday, and, er… it'll come to me.
This intelligent way of handling MIDI data makes it possible to quickly create large-scale mock-ups and musically coherent scores at the same time. There are two drawbacks: the first is that Cubase's Score Editor's vertical positioning of the musical symbols leaves a lot to be desired, but at least the symbols are in there, and I can easily and automatically tweak their positions in Sibelius later, once I've imported the Cubase arrangement as a MusicXML file. The other issue is that VST Expression is currently only offered by Steinberg/Cubase, but we live in hope that it will be implemented by other developers in due course! David William Hearn
For more about working with VST Expression, see the video Cinematic Strings & VST Expression at http://player.vimeo.com/video/37377379 and read John Walden's SOS article at /sos/jan10/articles/cubasetech_0110.htm.
This simple technique is as old as the hills, but it works for me: when programming string parts, try adding a solo violin to the ensemble. The sound of the solo instrument (in particular, its highly emotive vibrato) introduces an extra expressive dimension. Using the mod wheel as the volume controller, you can subtly fade it in and out to add poignancy to selected phrases, or use it more dramatically to emphasise a climactic passage. Be careful not to overdo this, though, as it can have the effect of making large ensembles seem much smaller. David William Hearn