Harm Visser''s Creative Physical Modeling Toolbox for Reaktor 5 includes more than 130 Ensembles that use physical modeling to create real and virtual instruments.
Noted sound designer Harm Visser has released an impressive set of Reaktor Ensembles that use physical modeling in a variety of ways. Creative Physical Modeling Toolbox for Reaktor 5 ($189) includes instruments from all the traditional symphonic families (woodwind, brass, percussion, and string), several prebaroque and non-Western instruments, and many other imaginary sound-making devices. The models are based on analyses Visser performed on recordings of actual acoustic instruments, for which he determined the various resonant frequencies of the instruments, then built Ensembles tuned to those frequencies. The results are often stunningly realistic.
Most of the roughly 130 Ensembles have only a single Snapshot; perhaps some of the Ensembles that cover similar instrumental families could have been grouped. And Visser won't win any awards for graphic design, as the Ensembles' interfaces are very basic and don't use any modern-looking graphic elements. But the sonic potential of this set is vast, and the opportunity for experimentation is unlimited.
The collection is divided into 14 groups. In the acoustic family, you'll find three separate woodwind sets (Western, Medieval, and Ethnic); two percussion groups (Mallets and Percussion); Bowed, Plucked, and Plucked Ethnic Strings; Keyboard and Dulcimer sets; and a Brass set. The FOF group focuses on vocal models, and Bounce Roll Scrape includes Ensembles that model a variety of playing techniques on both real and virtual instruments. The MovieSounds set mixes components from the other groups, and a miscellaneous category called Various includes a number of bowed and blown models. (You can find a complete list of the instruments in each group at Visser's Web site.)
Most of the individual Ensembles include several tunable resonators that are fed by either a noise source or an excitation source, as well as one or more ADSRs, an occasional filter, master volume and tuning controls, and other parameters specific to each model. In some cases, Visser gives intuitive names to the individual Modules, such as Lips, Breath, Mouthpiece, and Overblow, but in others, there is no indication as to what aspect of the instrument's makeup you are adjusting (Timbre, for example). Rollover help explains the functions of some Modules, but again, this is not consistent across Ensembles or over parameters within an Ensemble. (A manual in PDF format gives a good basic explanation of the parameters of most of the models.) But with very little effort — a few twists of a knob — you can determine the role any parameter plays.
If you're up for a little experimentation, you can easily change the included excitations for many of the Ensembles to one of your own. For example, I changed the exciter in one of the MovieSounds from the default to a sample of a human voice; the result sounded similar to what you might get using a cross-synthesis method such as LPC or convolution (see Web Clip 1). And by simply lowering the Pulse Transposition value and raising the Breath Noise amount in the FOF synth, I morphed a pitched female-vocal sound into a repeating series of low glottal pulses with a steamy hissing ambience (see Web Clip 2).
Keeping It Unreal
Visser's technique for creating physical models has nearly unlimited potential: any sound that can be analyzed can be modeled using his approach. Though a few of the acoustic-instrument Ensembles are not particularly compelling — the brass models are often too buzzy and overly thin and are the least effective overall — the vast majority are very musical and could be useful in numerous settings. Moreover, his convincing medieval and non-Western Ensembles bring editable models of often rare and exotic instruments to the desktop.
But even more exciting are the hybrid virtual instruments that explore realms well beyond the world of traditional instruments. There aren't too many other places you'll find a plucked flute or trombone! (To be fair, Visser has created similar models for Applied Acoustics' Tassman.) You can also easily create your own hybrids by mixing and matching the Resonance Ratio values of the individual instruments (real or imaginary). Visser's Toolbox is an outstanding demonstration of the potential of physical modeling and one of the best I've ever seen. This collection fills in a huge gap in the otherwise excellent Reaktor sound universe.
Value (1 through 5): 4