Polyphony
Definition ⚓︎
In a synthesizer, polyphony refers to the possibility of playing many different notes simultaneously. Each playing note is said to be a voice. Polyphony makes it possible to play chords, and allows sounding notes to decay naturally when new notes are played.
In contrast, synthesizers with a single voice are said to be monophonic.
This manual section only applies to the Multiphonics CV‑3 Synth plug-in and the standalone application. The Multiphonics CV‑3 Effect plug-in lacks MIDI input, and therefore cannot be polyphonic.
Patches on hardware modular synthesizers are usually monophonic. While a single modular system can be patched to play different sounds at the same time—like drums, a bass line and a plucked sequence—each of these sounds is monophonic and incapable of playing chords. The reason is simple: each voice of polyphony needs its own set of oscillators, filters, envelope generators, LFOs, etc. Moreover, all voices need to be set up the same way; otherwise, they will not sound like they belong to the same instrument. Needless to say this requires a massive investment in time, money and rack space.
Luckily, software synthesizers are not limited by physical space or by the fact that patch cables can only carry the signal for one voice.
Polyphony in Multiphonics ⚓︎
Multiphonics patches can gain 8 voices of polyphony simply by clicking on a switch in the Mode module. If you create or load a monophonic patch, clicking on that switch will instantly turn it into a fully working polyphonic synth.
When polyphony is enabled, all modules visible in the rack actually represent 8 modules that share the same settings. In other words, turning a knob on a module will also turn the same knob of 5 other identical modules that are hidden from view. Same with patch cables: connecting two visible modules will make the same connection between those modules on all voices.
Why only 8 voices? It’s a compromise between CPU Usage and playability. We wanted to allow for reasonably complex two-handed play while minimizing the risk of CPU overload.
Display ⚓︎
When polyphony is enabled, the LEDs, level meters, and displays in various modules will show the state of the most recently triggered voice. For example, if you play C3, G3, and C4 in succession, the modules on screen will display the state of C4 until a new note is played.
This behavior may lead to some confusion if the last-played note is released while earlier notes are still playing. For instance, if you play and hold C3 followed by C4, the visible state will be that of C4. If you release C4 while C3 continues to play, the modules will still display the settings for the C4 voice. As a result, although C3 is still sounding, the LED for the Keyboard Gate will be dark and ADSRs connected to the gate will be in their release state, reflecting the now-released C4. Therefore, if you’re looking at module LEDs and displays to understand the behavior of a patch, you should only play one note at a time.
CPU Usage ⚓︎
Playing more than one note on a polyphonic patch will necessarily use more CPU power than a monophonic patch. Given the modular nature of Multiphonics where anything can be routed anywhere to modulate anything else, it uses more CPU power and memory per voice than a single-purpose synth that would implement a similar patch.
If you intend to design or use a very large polyphonic patch or a polyphonic patch with many CPU-intensive modules like the Objeq Oscillator, Objeq Filter or 3-operator Digital FM, we recommend running your DAW at a sampling rate of 44.1 kHz or 48 kHz.
Multiphonics will attempt to minimize the waste of CPU time by shutting down voices that produce no sound. However, the voice that is currently visible on screen will always run in order to correctly update the state of the LEDs and other displays.
The exact rules for voice activation are as follow:
A voice is activated when a new note is played on the keyboard or on a MIDI track in a DAW.
The voice remains active for as long as the MIDI note is held.
The voice remains active for as long as an audio signal is detected on any cable connected to the Output Mixer.
The voice will only be deactivated when its audio output remains below the threshold for a short period of time.
Of course, in patches with a long enough decay, it’s easy to activate all voices at the same time even without playing 8 notes simultaneously.
Tips for Polyphonic Patch Design ⚓︎
Always trigger the sound from the Keyboard Gate or Trig outputs; don’t create a polyphonic generative patch.
Make sure that the sound completely stops shortly after the note is released.
Avoid using too many CPU-intensive modules. Limiting yourself to two Objeq Oscillator modules at maximum density or four 3-operator Digital FM modules shouldn’t prevent you from creating great sounding patches!
If you need a portamento effect, use the Glide or Slew Limiter modules and connect them directly to the Keyboard Pitch. They are polyphony-aware and will automatically glide from the previous note that was played. Patching a custom portamento by filtering the pitch with a low-pass filter will not work.