Kawai K5000R

The Kawai K5000R is one of the only dedicated additive synthesis hardware synthesizers ever commercially produced. Released in 1997, it is the rackmount version of the K5000S keyboard. It was Kawai's second attempt at additive synthesis, following the K5 from the late 1980s, and represents the most sophisticated implementation of additive synthesis ever put into a hardware instrument.

Where every other synthesizer starts with a fixed waveform and sculpts it, the K5000R lets you construct the waveform itself from the ground up — and then animate every component of that waveform independently over time. The results are sounds that no subtractive synthesizer can produce: pads that breathe harmonically, textures with genuinely organic movement, bells that decay with mathematically precise overtone behavior, vocal formant shapes that sweep the frequency spectrum.

All sounds in nature are combinations of sine waves — a fundamental frequency plus harmonics above it. Additive synthesis works by building sounds from scratch using individual sine waves rather than filtering a harmonically rich waveform down. The K5000R implements this as Advanced Additive — "advanced" because each of the 64 harmonics in a source has its own independent amplitude envelope.

The Harmonic Series

Harmonic 1 is the fundamental pitch — the note you play. Each subsequent harmonic is a specific musical interval above it:

A sawtooth wave contains all harmonics at levels proportional to 1/n (harmonic 1 loudest, harmonic 2 half as loud, harmonic 3 a third as loud, etc.). A square wave contains only odd harmonics (1, 3, 5, 7…). A sine wave is harmonic 1 only. On the K5000R you can build any of these — or anything that has never existed before.

Why This Matters for Sound Design

In subtractive synthesis, when you sweep a filter the entire harmonic structure of the sound changes together — everything gets brighter or darker as a unit. In the K5000R's additive engine, you can have harmonic 3 fade in slowly while harmonic 7 decays quickly while harmonic 12 loops between two levels — all simultaneously, all independently. The result is a sound whose tonal character is literally changing at the harmonic level throughout the life of every note. That's why additive pads have that impossibly alive, organic quality that subtractive synths can never fully replicate.

A K5000R patch consists of up to 6 Sources, each running independently through its own synthesis engine and combining at the output mixer. Think of each Source as a complete synthesizer voice with its own waveform, filter, and amp.

Per-Source Signal Flow (ADD mode)

The Internal Mixer

Each of the 6 Sources has its own DHL Common level — the master volume for that Source's entire waveset. This scales all 64 harmonic levels proportionally, preserving their relative relationships while adjusting how loud that Source sits in the overall patch mix. Sources can also be muted individually for isolation during editing.

Source Types

The DHL section is where you set the static amplitude of each of the 64 harmonics — this is your waveform. Think of it as drawing an oscillator waveform from scratch instead of choosing one from a menu. The combination of all harmonic levels is called a Waveset.

Edit Groups — Your Shortcuts

Editing 64 individual harmonics one at a time would be impractical. The K5000R provides group shortcuts that adjust multiple harmonics simultaneously while preserving their relationships:

DHL Common

The DHL Common section sets the overall volume for the entire Source waveset, plus velocity curve response and keyboard scaling. The Velo Depth parameter controls how velocity affects the blend between soft and loud harmonic content — a powerful expressiveness tool. Use this to balance how loud each Source sits in the patch mix without disturbing the internal harmonic balance.

Waveset Library Strategy

There is no standalone Waveset save format — Wavesets live inside Sources, and Sources live inside patches. The standard workflow is to maintain a dedicated bank of template patches — patches that exist purely as a Waveset library rather than finished sounds. Build a great waveset, save it as a patch, and use the Source Copy function to copy that Source into any other patch. Over time this template bank becomes a personal oscillator waveform library.

This is where the K5000R does something no subtractive synthesizer can. Each of the 64 harmonics has its own independent amplitude envelope — not just a static level, but a full time-varying shape with Attack, Decay 1, Decay 2, and Release segments, each with its own Level and Rate.

Envelope Structure per Harmonic

With 64 harmonics each having 8 parameters, you have up to 512 envelope parameters per ADD Source. This is why the Multiview and group editing are essential — managing this at the individual harmonic level is done only for fine-tuning, after establishing the basic shape using groups.

DHE Multiview — Work Here First

The DHE Multiview shows all harmonic envelopes simultaneously and uses the same BRIGHT/DARK/ODD/EVEN/OCT/5TH group system as the DHL. Select a group, adjust the Attack Rate — all harmonics in that group shift together. This is how you build envelope shapes in practical time. The book strongly recommends: establish your spectrum in DHL using groups, establish your envelope shapes in DHE using groups, then use EACH for individual fine-tuning.

DHE Loops — Perpetual Motion

Individual harmonic envelopes can be set to loop between Decay 1 and Decay 2 levels, creating continuous movement in a sustained sound. Two loop modes are available:

Having two or three harmonics with their own independent loops creates a sound that seems endlessly moving — the ear will eventually detect the cycle, but it takes far longer than with a sampler loop, and layering LFO or Formant Filter modulation on top makes it practically imperceptible.

The Formant Filter is described by Bellingham and Gorges as "the blade of the sound butcher" — and this is not an understatement. It is the most distinctive and powerful feature of the K5000R and what most distinguishes it from any subtractive synthesizer.

What It Is

The Formant Filter (DFL — Digital Formant Level) is a 128-band graphic equalizer where each band is one semitone wide, spanning the entire audible frequency spectrum. You draw a shape across these 128 bands — peaks, dips, notches, lowpass curves, anything you can imagine. That shape is then applied to your ADD source's harmonic content.

Unlike a conventional EQ, the Formant Filter has a Bias control that shifts the entire drawn shape left or right across the frequency spectrum. This is what makes it extraordinary: you can draw a formant peak pattern and then sweep that entire pattern up or down the spectrum with an LFO or envelope — creating the vowel-like, vocal formant sweeps that give the K5000R its characteristic "alive" quality.

Key Parameters

DFE — Digital Formant Envelope

The DFE controls the Formant Filter over time, specifically animating the Bias parameter. It has the same envelope structure as the DHE (A/D1/D2/R with level and rate per stage) plus loop modes. The DFE also has an LFO mode — essentially a simplified looped envelope — which is the most commonly used method for ongoing filter movement in pads.

Simulate Flanging with the Formant Filter

Set the Formant Filter to a comb-like pattern (alternating peaks and dips) and sweep the Bias slowly with an LFO. The result sounds convincingly like a flanger but is happening at the harmonic level rather than through a delay line. This can free up the effects processor for other uses.

Each of the 6 Sources in a K5000R patch can alternatively run as a PCM source — a conventional subtractive synthesizer using one of the 123 onboard PCM samples as the oscillator (DCO), processed through a digital filter (DCF) and amplifier (DCA).

PCM Signal Flow

DCF Tips

• Always let the DCF Cutoff respond to velocity and key tracking on instrument-like sounds — this gives a more natural, playable feel
• The K5000R's digital filter resonance is extremely hot — high resonance values (above 4) cause distortion. For every 2 steps of resonance increase, decrease DCF Level by 1 to compensate
• You can bypass the DCF entirely but the book recommends keeping it slightly active even on ADD sources to smooth out very high frequency content
• Heavy use of the DCF belongs on PCM sources. For ADD sources, the Formant Filter is almost always the better choice — it can fully simulate a DCF and much more

When to Use PCM vs. ADD

The K5000R has a 4-bus effects processor. Each Source can be routed to a different effects bus (set via the Effect Path parameter in each Source's Control page). The effects chain ends with a global reverb and 7-band GEQ that affect the entire patch.

Effects Reference

Realtime Effects Control

Effect parameters (depth, speed, balance) can be assigned to the mod wheel, aftertouch, or other MIDI controllers via the Effects Control page. The K5000R's Macro Controllers (available via the MCB-1/10 expander with the R model) allow assigning two parameters simultaneously to a single controller — one knob turn can change effect depth, source balance, and Formant Filter LFO speed all at once.

The K5000R rewards a methodical approach. Bellingham and Gorges recommend this sequence for building ADD patches from scratch:

Recommended Build Order

1. Start in DHL using groups — establish the basic tonal character of your waveset. Use BRIGHT/DARK/ODD/EVEN/OCT/5TH to shape the spectrum. Get close before touching individual harmonics.
2. Balance Sources — if using multiple Sources or mixing ADD with PCM, set relative levels using DHL Common on each Source. Mute/unmute to audition individually.
3. Shape time in DHE using groups — add harmonic envelope movement. Decide which groups of harmonics should have slow attacks, which should decay quickly, which should loop.
4. Fine-tune individually — only after groups are dialed in, switch to EACH mode for surgical adjustments to specific harmonics.
5. Add Formant Filter — set up the DFL shape, then add DFE animation (LFO or envelope) to bring the filter to life.
6. Add DCF/DCA — final overall envelope shaping. Use DCA for the master amplitude shape; use DCF only if needed (more useful on PCM sources).
7. Add effects last — build the patch strong without effects first. Add effects as enhancement, not compensation.

Morf Mode

Morf mode creates crossfades between the Wavesets of up to four different Sources, generating intermediate harmonic shapes as it transitions. It's fun for exploring but the book is candid: "For you serious programmers, invest your time in other areas of the synth. Morf mode is fun but the results are generally unspectacular." Noted.

Source Copy

Any Source (including its complete Waveset, all DHE envelopes, and Formant Filter settings) can be copied from any patch to any other patch. This is the primary mechanism for building a Waveset library — copy your best Sources into a dedicated template bank and pull from them as needed.

Avoiding Aliasing

Aliasing occurs when high harmonics exceed the Nyquist limit and fold back as distortion artifacts. It's most noticeable on the upper keyboard range. To reduce: set harmonic 64 to zero (the 64th harmonic is the most likely culprit), or use a positive KS to Gain value in DHL Common to gradually reduce upper harmonic levels as you play higher up the keyboard.

Simulating Natural Sounds with Additive

Real-world instruments don't have harmonics at static levels — the harmonic balance shifts during the attack, sustain, and decay. A real piano's attack is brighter (more upper harmonics) than its sustain. Simulate this by setting upper harmonic groups to have faster, higher-peaking DHE envelopes than lower harmonics, which sustain at a more consistent level. The same principle applies to any acoustic instrument emulation.

Create Bell-Like Spectrums

Start with a sparse waveset — just a few widely spaced harmonics rather than a full sawtooth-style spectrum. Set those harmonics to have a fast attack and a long, slow release. Inharmonic spacing (detuning some harmonics slightly from their integer positions using the fine tuning) gives that characteristic bell shimmer.

Create Biting Leads

High EVEN harmonic content (nasal/brassy character) + moderate Formant Filter with slow upward Bias sweep + hard velocity response on DCF and DCA = cutting synth lead that opens up the harder you play.

Optimal EQ Settings

The global 7-band GEQ at the end of the effects chain can compensate for the K5000R's output stage characteristics. The book notes that cutting around 2–4 kHz slightly and adding a touch of high shelf above 8 kHz tends to give a more modern, airy sound to heavily programmed ADD patches.

Memory and Polyphony Management

Each ADD Source uses more polyphony than a PCM Source due to the computational load of 64 harmonic oscillators. A 6-Source all-ADD patch is polyphony-intensive. Strategies: mute unused Sources, reduce the number of active harmonics in less important Sources (ALL group to minimum, then set only the critical harmonics using EACH), or use PCM Sources where ADD complexity isn't needed.

The MIDI Timeout Setting

When using MIDI Quest Pro or any SysEx editor, set the MIDI timeout to 750ms (default 500ms can cause communication failures). This was a known issue with the K5000 series and SysEx transfers.

Firmware

The latest K5000R firmware is version 4.04, available free from Kawai. Check your current version in the system settings and update if needed. Version 4.x added significant bug fixes and improvements over earlier releases. Note: there is a known firmware bug where deleting the last patch in a bank corrupts memory organization. Solution: never delete the last patch — leave at least one patch in every bank.

The K5000R has DIN MIDI only — no USB. It cannot connect to the mioXL's USB host ports or via Network RTP-MIDI. The following routing uses the iConnectMIDI2 (already owned) as a dedicated interface for the K5000R.

Hardware Chain

Why This Setup

• The K5000R receives both MIDI note/patch data (from Mac via iConnectMIDI2) and tempo clock (from Nome II) merged into its single DIN IN
• The return path carries SysEx patch dumps back to MIDI Quest Pro via the iConnectMIDI2's DIN IN
• The iConnectMIDI2 is the same interface used for 3rd Wave and Hydrasynth firmware updates — proven SysEx reliable
• The Nome II has two DIN outputs — Out 1 already feeds the mioXL for all other instruments; Out 2 is dedicated to the K5000R merge chain

Cubase Integration

Follow the same External Instrument process used for the Moog One (documented in studio-notes.html). The only difference: select the iConnectMIDI2's port instead of a mioXL Network RTP session when assigning the MIDI output. Cubase sees both the mioXL ports and the iConnectMIDI2 ports simultaneously — a port is a port regardless of which interface it belongs to.

MIDI Quest Pro

MIDI Quest Pro has a dedicated K5000R editor/librarian module — this is the recommended editing interface given the difficulty of the onboard hardware UI. Run MIDI Quest Pro as a VST inside Cubase. Set MIDI timeout to 750ms in MIDI Quest settings. All patch dumps and SysEx transfers route through the iConnectMIDI2.

The K5000R's stereo outputs can be routed into the Modal Electronics 002R's external audio inputs, turning the two units into a single hybrid synthesizer. The K5000R acts as the oscillator and harmonic generation stage; the 002R acts as the filter, amplifier, and modulation stage. Together they produce sounds neither instrument could achieve alone.

Signal Flow

MIDI Setup for Hybrid Mode

In Cubase, create a MIDI track with two simultaneous outputs — one to the K5000R's port (iConnectMIDI2) and one to the 002R's port (mioXL Network session). Both instruments receive the same notes at the same time. The K5000R generates its additive texture; the 002R receives that audio through its inputs and processes it through its own filter and VCA stage.

002R Patch Configuration for Hybrid Mode

Alternative Routing — Modal 001

The Modal 001 also has external audio inputs and can be used instead of or in addition to the 002R. Running K5000R audio through the 001's filters creates a different character — the 001 has only 2 voices of polyphony, but the external input processing is independent of voice count, so the K5000R's full polyphony passes through the 001's filter as a single audio source.