String parts are among the easiest parts to score; surprisingly little can be written on a staff that a good string player can’t perform beautifully. This is great news for composers, arrangers, and orchestrators, but it’s the sort of thing that makes synth players break out into a sweat. Trying to emulate all the nuances of a good string performance using a sampler or a sample-based synth is at best very difficult and at worst downright impossible, a situation that is not helped by the fact that many synths and sample libraries fall short when it comes to solo string samples.

Happily, performing string sections from a keyboard is relatively easy. Bunch those pesky string players together, and many of the finer points of articulation and expression tend to get lost in the crowd. They’re reasonably easy to sample en masse and forgiving when controlled from a keyboard. (For information on sampling string ensembles, see "All Together Now" in the June 1998 EM.)

But solo strings are an altogether different story. For one thing, due to memory constraints, the average synth only offers one kind of articulation on a solo string patch. "Many synths don’t even have different samples for violin, viola, and cello," points out composer orchestrator, and keyboard player Tom Zink, who specializes in programming orchestral emulations for individual clients and companies such as Coda, Kawai, and Yamaha. "They’ll maybe have one sample and just change the timbre a little bit and call it something else."

As a result, your choice of pitch range, articulation, and instrumental effects is considerably limited before you’ve even played a note. The obvious solution, Zink suggests, is to "use a very good sample library such as the Miroslav Vitous Symphonic Orchestra Samples [reviewed in the September 1995 EM] or the Denny Jaeger Master Violin Library."

In short, the situation isn’t ideal, particularly for keyboardists who can’t hire a real string player or buy a fully expanded sampler and high-end sample library. Nevertheless, there are steps you can take to improve your solo string parts to the point where you can achieve an acceptable level of realism.

Obviously, you always need to be aware of the pitch range of the instrument you’re writing for. Don’t go below the range of the instrument, but also watch out for your sample’s realism in the upper part of the instrument’s range. This may be a little restrictive with violin and cello parts, but not so much with viola, which is rarely played high in its range. "Stick to the low or middle portion of the instrument’s pitch range," offers programmer Terry Wilson, whose synthesized instruments are used by such film composers as Georgio Moroder, David Kitay, and Anthony Marineli.

Other than pitch constraints, however, it’s virtually a case of anything goes. "Strings are extremely versatile, especially cellos," explains Zink. "Because their instruments are tuned in fifths, string players can jump all over the place, and they can do incredibly fast lines, slurs, or very quickly repeated bowing on a note. It’s actually the opposite of what you might expect: there’s pretty much nothing that you can do on a keyboard that you couldn’t do on a stringed instrument—except play ten notes at once!"

Violinists and cellists might not be able to play ten notes at once, but they can play two, three, or even four notes with relative ease. However, cellist Matthew Brubeck points out that double-stopping intervals of a fifth is somewhat difficult for a string player, and 4-note chords must be "spread," or arpeggiated. (The son of famed jazz master Dave Brubeck, Matthew leads the Matthew Brubeck Quartet, plays with the Berkeley Symphony, and has backed up such pop singers as Jewel and Tom Waits.)

Keyboard players should be cautious when playing chords with string sounds, because the results often sound unnatural. "It’s strange, but you even have to be careful about synthesizing effects that are common on the real instruments," says Zink. "Double stops can work great if they’re short and bowed strongly [detaché], as is common in cello lines, but if you’re playing them from a keyboard, they tend not to work when sustained or played legato. I generally avoid playing more than two notes at once, and I try not to use double stops that often."

In any attempt to emulate another instrument, half the battle is working out how to translate the idiosyncrasies of phrasing and articulation on that particular instrument into an appropriate keyboard line. That usually means adjusting your playing technique. "With strings, it’s good to play with an even touch; dynamics should generally be controlled with the master volume rather than by how hard you strike the keys," suggests Wilson. "You can actually sound more realistic by prevention," adds Yamaha sound designer and style programmer Nate Tschetter. "For example, try playing with the articulation that’s sampled into the sound."

"The biggest thing that makes strings sound synthy is that keyboard players put both of their hands on the keyboard and just start playing," states Zink. "I recommend always playing or programming string parts individually, one line at a time. That way you can approach it like a real player does and concentrate on articulating it right."

In addition, with only one hand on the keyboard, your other hand is free to operate the modulation wheel or another assignable controller so you can make real-time changes to add natural-sounding expression to your string line. "I regularly use Controller 7 [Volume] and Controller 11 [Expression] at the ends of phrases to drop the volume slightly, like real string players do before they release or lift the bow," reveals Zink, who does a lot of orchestral emulation for the Yamaha Disklavier library. "Controller 11 is good for emulating swells," he adds, "and in XG synths, there’s a dedicated controller for changing the attack and decay rate, which can be extremely useful on solo string sounds."

Altering attacks from note to note can really help your string line sound realistic. Real string players alternate downbows and upbows, and they rarely play consecutive notes with the same attack. Subtle changes such as these contribute greatly to the expressive nature of the instrument. To slightly alter the attack rates on a Yamaha XG instrument, you can use CC 73 (Attack Time), assigning it either to a physical controller such as a slider or to key Velocity. If you’re using a sampler, you can do a similar thing with the Sample Start function, moving the start time slightly with a wheel, slider, or Velocity.

You could also emulate Terry Wilson, who loads up his Roland S-760 with several different, sampled attacks on separate MIDI channels. "There’s a feature in Emagic’s Logic Audio that allows you to randomize the MIDI channels. When I’m programming a string line, using that feature to switch between different sampled attacks really helps," says Wilson. "I’ll take maybe a couple of attacks from the Denny Jaeger library, one from Best Service’s Peter Siedlaczek’s Advanced Orchestra sample CD library [reviewed in the September 1998 EM], and perhaps one or two others that sound similar from elsewhere. Then I’ll trim them up so they’re the same length, but each with a slightly different tonal color. You have to randomly switch between them, though, because if you formed a pattern, it would sound almost as unrealistic as just playing the same sample over and over."

That’s a sophisticated way to go, but even alternating between two samples—say a heavy attack for a downbow and a lighter attack for an upbow—can work. "If your instrument supports a key rotation scheme, you can use that to switch between two samples from one key press to the next," suggests Tschetter.

The biggest telltale sign of a synthesized string line is when you hear an initial attack with every note. A real string player commonly plays several notes in the same bow stroke to get a legato effect, or slur, but this simple musical figure is difficult to achieve using string samples controlled from a keyboard. One possible solution is to alter the level of your sample’s attack in real time, using one of the aforementioned controllers.

Alternatively, if your synth or sampler supports Mono mode (whereby the attack portion of the sound only retriggers when you play distinct, detached notes), you can experiment with that. Beware, though, as even this can sound unnatural. "I have never had much luck with Mono mode," warns Zink. "For one thing, strings can play more than one note, and a lot of times they do. When performers phrase between notes on different strings, they might not lift their finger up from the last note, and there can be quite a lot of overlap. Mono mode has an unnatural cutoff; even when two notes don’t overlap, if the notes are slurred, your ear can fool your mind into hearing a very slight overlap."

Zink gets around this problem by purposefully using a short overlap, which has the psychoacoustic effect of diminishing the amount of attack on the second note. This is something that’s far easier to do in a sequenced string line than in real time, where simply moving a controller to reduce some of the sampled attack is perhaps more practical.

Some of the easiest things to play on a stringed instrument can be the hardest to emulate from a keyboard. Vibrato is a good example: most solo string samples include vibrato in the sample, which is by far the best way of achieving maximum realism. However, it can also be very restrictive.

"The problem with having vibrato sampled into the sound is that you have no control over it," notes Tschetter. "If you can control it yourself using an LFO, you’ll have more freedom to use and assign it as you wish, depending on the style of music you’re performing or programming. For example, country fiddle players have a very wide pitch depth on their vibrato, so you want to set that parameter accordingly. It’s nice to control filter, amplitude, and pitch for vibrato; if you can control the depth with the modulation wheel and the rate with Aftertouch, it can sound really good."

"It’s quite important to change the vibrato rate depending on the style of music and also the dynamic," agrees Brubeck. "String players in general vibrate faster when they play louder, and in earlier styles of music, such as baroque, vibrato would only be used as an ornament, whereas in later styles it’s more of a constant."

It’s also worth bearing in mind that the lowest note on any stringed instrument is always played sans vibrato. The same thing is true if you’re emulating a string player tuning up, where they would be bowing or plucking only the open strings.

On the subject of plucking, or pizzicato, most orchestration books note that string players generally can’t do fast pizzicato runs, and that’s something to bear in mind if you’re playing or programming in a classical idiom. However, there are exceptions to that rule, as Brubeck explains: "If you want something to sound orchestral, you can’t use techniques that nonorchestral players have worked on. Acoustic bass players who are schooled in jazz, for example, can play pizzicato a lot faster than classical players simply because they spend more time doing it. The same is true of cello; I can play pizzicato much faster than most cellists because I use a 2-finger technique, whereas most cellists are trained just to use one finger."

Another common string effect is tremolo, which is rapidly repeated, short-bowed notes. Because it’s impossible for a keyboard player to restrike the keys as fast as a string player can play tremolo, this is another effect that works best when recorded as part of the sample. But what if your synth doesn’t have that sample? "Try using an amplifier and filter LFO with a square wave shape going really fast and deep," suggests Tschetter.

Occasionally your string line may call for a glissando; simulating that is almost impossible. Depending on the style of music, there are two ways to go. Using regular Pitch Bend can work for the cliché, Psycho effect. However, string players often perform glissandi by fingering and playing each note individually. A standard keyboard glissando is the only way to attempt this, although you’re restricted to a diatonic scale along the white keys (and exact pitches), which isn’t altogether ideal.

"A glissando is something I wouldn’t even try to play live!" says Zink. "I’d program it into a sequencer and probably draw in the bend. If you listen to a string glissando or pitch bend, you’ll find they tend to start off slowly, and then they’ll go faster, so you should draw a curve that is not completely linear. There are often some other things going on, though, such as volume changes and tonal changes, so you want to use a filter as the bend occurs. On the XG, for example, I’d use Controller 74 [Brightness] to control the timbre on this effect."

One major stumbling block when performing stringed instruments from a keyboard is that, unlike keyboard instruments, fretless orchestral strings aren’t constrained to playing only twelve divisions of the scale. Fiddle players and jazz/blues players use this to great effect, so using Pitch Bend freely in those styles can be good. Classical players rely more on vibrato for this type of expression; however they misfinger notes sometimes, accidentally or deliberately.

"Most string players at least try to center the pitch if they miss, so you can assign a pitch EG that subtly detunes each note and slowly draws itself back in. The rate at which it does this can be your player virtuosity factor," suggests Tschetter. "Then you assign the pitch EG depth to Velocity to give you control over it. A simpler way would be to just assign small amounts of fine tuning to Velocity or Aftertouch, but that’s not as convincing."

"Good string players are actually very deliberate about their mistunings, and part of what makes strings unusual is that they don’t play in tempered tuning, like a piano or other keyboard does," remarks Zink. "There’s a cool program called Justonic Tuning Pitch Palette [described in the January 1997 EM "What’s New" column] that looks at a MIDI file, analyzes every chord, and then retunes each note within each chord to just tuning. It only works on synths that accept microtunings, though, because it uses System Exclusive microtuning information. I’ve used it with my XG module, and it’s really pretty happening—definitely the next step toward instrumental realism."

String players regularly employ an assortment of special effects to vary the tonal color and add musical interest. For example, they play near the bridge (sul ponticello) to create a pinched sound, rich in high frequencies. To simulate this, try using a sweepable bandpass or highpass filter. They can also mute the strings (con sordino), which produces a quieter, rather muffled sound. If your synth or sampler doesn’t have its own solo muted string patch, applying a lowpass filter with resonance might work. Another well-known special effect is snapping the string back onto the fingerboard (Bartók pizzicato), which you could attempt to emulate by layering an unpitched percussion sound to your pizzicato sample.

One common effect involves the use of harmonics. Finding a good harmonic sample could be a problem, and as with all of the aforementioned effects, you’ll have to come up with some convenient way to bring it in and out when required.

"I would use sample Velocity-switching and assign high-Velocity notes to play the harmonic sample and the lower-Velocity notes to play the regular string sound," says Tschetter. "You can then use yet another controller, such as a footswitch, to bring the samples in and out."

A little reverb can go a long way on a solo string line, but the easiest way to make a string sample sound unrealistic is to drown it in reverb. So don’t overdo it. "In a concert hall, the strings are close to you, and they don’t seem to put out the kind of timbre that bounces off the wall," explains Zink. "I always use a separate reverb for strings and use far less of it than I would on, say, a brass or woodwind sound. I also mute the high end of the reverb so it’s more midrangey, and I make it shorter."

"For country fiddle, an amp simulator with a very slight amount of distortion can sound cool," adds Tschetter.

Of course, the oldest trick in the book for achieving a convincing sound is perhaps the easiest and quickest one to implement: hire a real string player. "Think about how much time you actually spend getting a really good sound and all the effort it takes to do all that programming. If you just stick a microphone up in front of somebody’s instrument, that’s the closest you’ll get to realism," concludes Brubeck.

But hey, who wants to take the coward’s way out? Not me.

Sam Molineaux became semiproficient at the violin, viola, and cello before she reached puberty, at which point she switched to the synth, which was far more rock ’n’ roll. She’d like to thank Matthew Brubeck, Nate Tschetter, Terry Wilson, and Tom Zink for their invaluable help with this article.

Note: This article appeared in the October 1998 issue of Electronic Musician

This article presented courtesy of Electronic Musician magazine.