The Sirius is a keyboard "groove-synth," featuring a subtractive hybrid-tone-generation synthesizer referred to as DTE ('Difficult To Explain') synthesis [1] introduced in 1997 by Quasimidi. The unit featured a both real-time and step sequencers sequencer with pattern- and song-modes, capable of acting basic drum machine, groove-box, or sound-module. The units is thus 7x multitimbral and has 28-voice polyphony across it's 7 tracks, with track selecting a (track-specific) sound within 96 sounds (per bank) of preset- or user-writeable sounds. Track/voice structure was organized as follows:
- Kick drum- 1 voice sample through synthesis chain
- Snare drum - 1 voice sample through synthesis chain
- Hi-Hat - 2 voice samples (open/closed) through synthesis chain
- multi-sampled Percussion drum kit - 12 samples with common synthesis chain)
- (chromatic) Synth 1
- (chromatic) Synth 2
- (chromatic) Synth 3
...where all 3 (chromatic) Synth parts loading mono- or polyphonic keyboard patches.
In addition to level and pan controls, each track could also be routed to 2 effects engines; 1 for reverb/delay, 1 for modulation delays (delay/chorus/flanger). It also features an 11-band vocoder with flexible track routing to its carrier and/or modulator, and an advanced/programmable arpeggiator/gate-sequencer for automatic rhythms controlled by its 37-note keyboard.
The DTE synthesis method combines a basic sample playback tone generator with virtual analog-like controls. While some of the waveforms are sampled instruments, most of them are merely single-cycle samples of periodic waveforms common in analog modeling synthesizers. The unit is technically a rompler, but the it's marketing and technical documentation refer to it as virtual analog synthesizer. This is not entirely accurate, given the lack of any options to shape, sync, or ring-modulate VCO's common to real- and virtual-analog synthesis. While sequencing tools and the (preset) sounds of the Sirius were tailored toward 90's techno- and dance-music sounds and workflow, and the models in production held some unresolved limitations and bugs (being the last model produced before the Quasimidi company folded), the synthesis architecture and user interface of the Sirius feature unique innovations (many of which not seen on more modern/successor synthesizers) that may contribute to its niche appeal. Among these, include it's particular handling of...
... the vocoder
In addition to routing the onboard mic- and line-inputs for traditional "talking-synth" applications, any (combination of) tracks could be routed to it as carrier, modulator, or both (for filter-bank effects) with the front-panel buttons.
the vocoder acted as an 8th track, with level, pan, and effect-send controls.
...the "ARP" (arpeggio,gate,etc.) for rhythm creation, similar to the "motivator" of the
any voice could be set to 1 of 16 arpeggiator settings.
in addition to storing the voice-synthesis settings, each program could store (inherent) arpeggio settings for one of 3 modes:
arpeggiator, with options for rhythmic value, constraining arp-length to pattern-length (or running free), and note-handling, using either simple monophonic sorting (up/down/random/etc) or one of 9 "UserArps;" user programmable 8-note poly-phonic note-sorting sub-sequences with per-note accent and swing.
"gater," which rhythmically chops the voice amplitude in a rhythmical tremolo
"chord," where the played keys will be re-triggered according to a fixed rhythmic rate, or according to the sequencing of a separate part (i.e. pattern of Snare track plays the notes keyed into Synth 2 part)
...randomization
the unit featured a dedicated button to randomize the settings for the presently-selected voice (within that tracks drum/snare/synth/etc architecture).
any track could be reassigned to a new randomly selected motif from memory, facilitating improvisational remixing.
...organization and inheritance
the 96 (per preset- and user- bank) voice programs available to each track were organized into 6-groups (of 16 presets) by character, both for chromatic voice ("clean bass," "dirty bass," "plucked," "pad," "lead," "SFX") and drum voice ("clean acoustics", "dirty acoustic," "electronic percussion," "808," "909," "SFX")
Sequencer was stored sequencer data as motifs-per-track within patterns, and songs as sets of (8) patterns. Editing any motif or pattern data would change any pattern or song that used it.
changing the character of any UserArp would change any voice that used that UserArp, and any Pattern- or Song or song that utilized that within it sequencing.
...knob Macros and multi-tasking
a single Wave Macro knob selects the starting waveform (single-cycle or drum rom-ple). Synthesizer voices instantiate 2 detunable instances of the same waveform, with no option to freely mix or match.
the ADSR envelopes for both the filter and amplifier could have all relevant times/levels set from the detailed edit-menu, but the front panel featured an Envelope Macro that morphed the ADSR settings continuously among plug, sustain, and
while synthesis knobs (tune, filter, etc.) adjusted such properties of the presently-selected-track's (drum, percussion, or chromatic voice) sound, the 8 dedicated mixer knobs allowed for controlling level/pan/fx-send across 8 tracks simultaneously (i.e. to turn down hi-hat and snare tracks maintaining keyboard and synth-panel focus on Synth track for melody).
