Reality is a funny thing, especially when you try to emulate it with MIDI instruments. What is it that makes one performance sound more real than another? The most popular approach to rendering the real world of musical performance has traditionally been to create ever-more-accurate samples of instrument sounds and combine them with a mechanism for switching quickly from one sample to another. That often yields mammoth libraries with a comparable appetite for processing power. Synful's Eric Lindemann, however, has decided to buck the current trend. His nimble Synful Orchestra charts a new course in an effort to impart greater realism to MIDI performances.
FIG. 1: Synful Orchestra''s front panel provides slots for assigning instruments to different MIDI channels. The sliders in the lower half of the screen let you adjust several of the parameters in the additive synthesizer.
Synful Orchestra is a cross-platform plug-in instrument that supports Audio Units (Mac only), VST, and DXi (Windows only) formats (see Fig. 1). Unlike sample-based plug-ins, which focus on producing individual notes, Synful Orchestra concentrates on the transitions between notes in an effort to capture the elusive real-life qualities of musical performances.
How a player moves from one note to another affects the notes and the performance in many subtle but vital ways that are often overlooked in other software instruments. For example, when two notes are played, the first note may linger slightly before moving to the next note; it may slide into the next note with varying degrees of portamento; or it may become louder, become softer, or change timbre during the transition. Each note may also introduce vibrato or add different amounts of breath or bow noise. Those context-sensitive transitions are what Lindemann calls the “connective tissue of musical expression.”
Synful Orchestra is based on a sophisticated multipatented technology called Reconstructive Phrase Modeling (RPM), which grew out of Lindemann's dissatisfaction with more-traditional approaches to synthesis. Here's how it works: Synful Orchestra contains an extensive RPM Phrase Database that includes musical phrases representing the myriad ways in which notes lead from one to another on different musical instruments. Phrases include a variety of articulations and other transitional elements, such as whether notes are detached or slurred, move precisely or with portamento, or incorporate various kinds of attacks.
When you play a series of notes into Synful Orchestra (directly from a MIDI controller or from a sequencer track), the software analyzes your performance and searches the database for the appropriate kinds of transitions. It then isolates the transitions and the associated notes from one or more phrases, adapts them, and recombines them into a new phrase that represents your performance.
Say, for example, that you want to play two notes on the violin in a highly expressive manner. The first note is louder than the second, it leads into the second note with a long portamento, and the second note has a distinct accent followed by vibrato. The RPM Phrase Database includes phrases with all of those transitional elements, but not necessarily in the exact form as your performance. The notes might be at different pitches, the portamento might be too long or too short, the accent might be too strong, and so forth.
FIG. 2: Drawing on a database of -transitional elements, performance data, and other components, Synful''s -Reconstructive Phrase Modeling (RPM) technology realistically re-creates your expressive gestures as you perform.
That's when the RPM technology steps in. It automatically pitch-shifts and time-stretches the notes and transitional fragments and interprets such things as overlaps (or lack thereof), crescendos, timbre changes, and Pitch Bend data. The program combines these “musical gestures” into a realistic rendition of your performance with the right pitches and the proper timings (see Fig. 2).
Extracting various transitional fragments and characteristics from several different phrases and seamlessly splicing them together in real time would, of course, never work if you had to rely on sampled sounds. (You'd soon be buried in clicks, pops, and strange-sounding artifacts.) Instead, Synful Orchestra is based on a form of additive synthesis in which sounds are generated from combinations of sine waves and noise elements (such as bow scrapes and flute chiffs).
All of this fancy computational footwork takes place behind the scenes, and you don't need to understand the RPM technology to get good results. Nevertheless, Synful Orchestra lets you control (to some degree) how the plug-in processes the input that it receives. The main point to keep in mind is that this is not a sample-based instrument, nor is it a form of physical modeling. It's a unique software instrument that enables you to perform expressively in real time without having to jump from sample to sample or use keyswitching to change articulations. In essence, it interprets and responds to your playing technique before it outputs the notes.
Synful Orchestra's basic layout is similar to other multichannel software instruments. Its front-panel Channel Grid offers a slot for each of the 16 MIDI channels. You can assign one of the built-in instruments to a channel by clicking on its drop-down menu. Each menu lists a modest assortment of solo orchestral patches such as flute, oboe, clarinet, English horn, bassoon, violin, viola, cello, bass, French horn, trumpet, and trombone. If you need more than 16 instrument slots, you can have as many instances of the plug-in as your CPU can handle. If you prefer having only a single instrument per instance, you can switch to Single Channel mode, which displays only one instrument at a time (see Fig. 3). In terms of processing demand, it doesn't matter if you have a single 16-channel plug-in or 16 Single Channel instances.
FIG. 3: If you prefer to have only one instrument per front panel, you can switch to Synful Orchestra''s Single Channel mode.
The plug-in provides real-time control over the note transitions by using CC11 (Expression) data. Buttons in the upper-right area let you use your Volume (CC07) pedal to send Expression data, or you can use separate Expression and Volume controllers. The Expression controller adds a unique kind of spontaneity to a performance by enabling you to alter the dynamics and timbre of notes and transitions to more closely mimic the responses of acoustic instruments. The changes that you make affect which phrase fragments are selected by the RPM engine, yielding different performances from the same series of notes. In fact, changing the Velocities of notes or introducing mod-wheel and pitch-bend data also changes how the RPM technology selects and assembles its fragments.
Speaking of pitch bend, Synful Orchestra provides a specialized Synful Pitch Wheel mode that is designed to produce more realistic slides and portamentos in string and wind instruments (something that a standard pitch-bend control has never been particularly good at). It takes a while to get used to it, but among other things, it lets you play a note, slide into a second note with the pitch-bend wheel, and play the second note without having its pitch displaced by the wheel.
Sliders and Delays
A bevy of front-panel sliders (all of which respond to controller data) lets you control the additive synthesis part of the program. Harmonic Tilt is a type of EQ control that lets you change the relative emphasis of the higher- to the lower-frequency harmonics. Release Trim affects the length of note releases; Sustain Noise Trim controls how much sustained noise (such as the bowing noise of a violin) is added to an instrument. Transient Gain lets you control how much transient noise is added to attacks or transitions. Harmonic Parity allows you to change the relative balance between even and odd harmonics. For example, I made the flute sound like a clarinet by dragging the slider all the way to the right to boost the odd harmonics. Smaller adjustments let you make more subtle changes to an instrument's tone color.
Regardless of how you set the sliders and play the notes, you won't get the best possible performance from Synful Orchestra if you don't record your MIDI tracks into a sequencer before playing them through the plug-in. That's because with a live performance, the RPM engine has no way of knowing what the next note in a series will be or how it will connect to the previous note until they're both played. As a result, the software has to guess at your intent, which generally results in more-conservative interpretations.
If you first record your tracks into a sequencer, however, you can activate Synful Orchestra's Delay For Expression (DFE) function. It adds a one-second delay between the sequencer's MIDI output and the plug-in's response. With a little extra time to think, the RPM engine can look ahead to the upcoming notes and transitions and make more-sophisticated choices that produce a more expressive result. You could activate DFE and perform directly into the plug-in, but according to medical experts, playing an instrument with a latency of one second has been shown to cause insanity in musicians. I'd advise against it except for experimental purposes. A better solution is to turn off DFE when recording tracks (you'll still get some expressiveness), and then turn it on for the final mix. You'll have to compensate for the delay before mixing Synful Orchestra tracks with tracks from other sources.
Synful Orchestra's additive synthesizer is a strength and a weakness. On the positive side, it offers control over a wide range of sound-shaping parameters. More importantly, it enables the RPM technology to work its magic. On the negative side, the individual instrument sounds, when played in isolation, don't match the stunning realism found in many of the newest high-end sample libraries. My favorite patches were flute, violin, and viola. Oboe, English horn, and cello were a bit weaker, and the brass instruments were a mixed bag. (According to Lindemann, a new and improved set of instrument sounds will be available in the coming months, and further improvements to the program should yield better instrumental renditions in future updates.)
The premise on which Synful Orchestra is based, however, is that note transitions and connections — coarticulations — are what make a performance seem real. And in that respect, Synful Orchestra is surprisingly successful. Its expressive output is natural sounding and varied, and the resulting performances are often convincing in spite of the synthesized sound of some of the current instruments. The listener is drawn into the performance because it doesn't sound like it's coming from a MIDI keyboard. (Check out the classical excerpts in Web Clip 1 at the EM Web site, www.emusician.com.)
With its extensive support for controller data and responsiveness to a wide range of performance gestures, it'll take some practice to truly get the most from this program. But in the short time that I've spent with it, I've found Synful Orchestra to be an amazing instrument, a real ear- and mind-opener, and well worth a look.
David M. Rubin is the author of the recently released Power Tools for Peak Pro (Backbeat Books, 2005).
Synful Orchestra 2.2.0
PROS: Wide range of expressive capabilities. Extensive controller implementation. Responsive real-time articulation changes. Modest processing demands. Cross-platform support.
CONS: Limited number of instruments currently available. Additive synthesis less effective at emulating some instruments than others. One-second delay required for optimum performance.
EASE OF USE