 |
|
|
|
FS Synthesis |
|
|
The term “formant” refers to the distinct spectral patterns which define the recognizable sounds of human speech, such as the vowels “a” or “i.” In traditional speech synthesis systems this is simulated by using an oscillator to perform the function of the vocal cords, and a series of controllable bandpass filters to create the required formant shapes. Consonant sounds such as “k” or “t,” and fricatives such as “f,” are based on slightly different principles, requiring a noise generator rather than an oscillator. Formants play an important role in defining the sound of many acoustic musical instruments as well as the human voice. Rather than a cumbersome system of oscillators and filters to synthesize the effect of formants, the FS1R employs 16 formant “operators” — 8 “voiced” operators, and 8 “unvoiced” operators (3 to 5 formants are generally considered to be more than enough to synthesize speech). Each operator digitally simulates the effect of both the driving source (oscillator) and filter in one easily manageable unit. The voiced operators produced pitched sounds which can be played on a musical scale via a MIDI keyboard or other MIDI controller. The unvoiced operators can be used to produce noise components of speech-like sound, or they can be used in much the same way as noise generators in more orthodox synthesis systems (e.g. to produce percussive sounds or sound effects). Each formant operator has a range of parameters which determine its shape and therefore contribute to the overall sound: center frequency, level, width, and skirt (the shape of the “flare” at the bottom of the formant’s bell-shaped response curve). All of these are programmable, and independent envelope generators are provided for the center frequency and level parameters of each operator. The envelope generators make it possible to produce a wide range of time-based timbral variation, such as smoothly shifting from an “a” to an “i” type sound. Also, any of the available controllers can be assigned to the formant parameters — the control knobs, a modulation wheel, foot controller, etc. — to allow realtime variation while playing. |
|
|
|
|
|
|
Formant Sequences |
|
|
The FS1R features a range of 90 preset FSeqs (Formant Sequences) which can be used to “sequence” the formants to produce voice-like phrases, rhythm loops, and more. FSeq playback speed can be set at a fixed value, controlled via a MIDI clock signal, varied by note velocity (e.g. keyboard dynamics) or varied manually. And since the sound of formants is not pitch dependent, Fseq playback speed and pitch can be varied over an extremely wide range without altering the basic timbre of the sound (something that even the best samplers cannot do). |
|
|
 |
|
|
FM Synthesis |
|
|
The FS1R’s operator-based architecture is what allows it to integrate both FM and Formant Shaping tone generation in one compact, easy-to-use module. To create FM (Frequency Modulation) voices, the operators are rearranged into a variety of “algorithms” with carrier/modulator relationships between certain operators. This basic architecture means that the FS1R is fully capable of producing the same type of FM sounds that placed the Yamaha DX-series synthesizers and TX-series tone generators firmly at the forefront of music production throughout the 1980’s — and makes them an indispensible part of the musical arsenal of so many discerning artists and production professionals even today. In fact, the FS1R actually comes with a complete set of 1152 original DX voices pre-programmed in preset memory! If you have other DX voices you have programmed yourself or obtained from other sources, they can be loaded into the FS1R too, and used with virtually no change in sound.
The FS1R actually takes FM synthesis to new levels of musical versatility and control. Unlike Yamaha’s original FM synthesizers and tone generators, the FS1R provides a choice of 8 different waveforms for each oscillator, thus significantly expanding the range of sound which can be produced. |
|
|
