Unison and Detune Explained: How Synths Get Wide and Thick

What is unison on a synth?

Unison plays several copies of the same oscillator for every note, like pressing the same key on several synths at once. Serum's manual describes it as stacking oscillators in a way that is similar to playing multiple notes of the same pitch, but slightly detuned, and it allows up to 16 voices while calling 7 the classic magic number.^[1]

More voices do not automatically mean louder. Serum keeps the level roughly constant as you add voices, so you judge thickness rather than volume.^[1] Other synths vary, so match levels by ear when you compare settings.

How does detune make a sound thicker?

Detune makes a sound thicker because slightly different pitches drift in and out of phase, so their sum constantly swells and dips. Physicists call this beating, and the beat rate equals the difference between the two frequencies.^[2] With many voices the beats overlap into the shimmering, chorus-like motion people hear as width and size.

Because detune is set in cents, a fixed detune produces faster beating on high notes and slower beating on low ones. The table shows the beat rate between two voices at two detune amounts. The low rows explain why detuned subs sound like they are fading in and out.

The practical rule: detune that sounds lush on a lead at A4 will make the same patch lurch when you play it two octaves lower. Use less detune on low parts, or automate the detune knob by register.

Unison voices, detune, spread and blend settings by sound

Set unison per role: pads and leads can take several voices and wide spread, while anything carrying the low end stays narrow. These values are starting points for Serum-style controls; scale them to your synth's knob ranges.

Blend decides how present the detuned outer voices are compared with the center voice. Serum treats it as a wet and dry mix between the unison and non-unison sound, with 75% as an even blend of all voices.^[1] Pull blend down when a patch sounds washy, and push it up for the full wall. The supersaw tutorial takes the lead row further, and the Reese bass tutorial covers detuned basses in detail.

Why does unison phase matter?

Phase decides where each unison voice starts its waveform cycle, and it changes the attack of every note. If all voices start together they are perfectly aligned for an instant and then drift apart, which produces a sweeping zap at the start of each note.

Serum's manual demonstrates this: with phase at 0 degrees, random phase at 0%, unison at 8 and a small detune, you hear a laser-zap phasing sound as the voices start in phase and slowly drift apart. Raising the RAND control offsets each voice randomly, and at 100% the zap is gone.^[1]

• Random phase high (70-100%) Pads, leads and chords. Every note sounds slightly different, which feels more analog and hides the stack.
• Random phase low (0-20%) Basses and plucks that need the same punch every time. Accept a little zap, or keep detune small.
• Phase retrigger on Useful on the single-voice sub under a unison bass, so the low end starts identically on every note.

When does unison hurt a mix?

Unison hurts a mix when it lands on parts that need to be solid, centered and consistent: subs, 808s and the fundamental of a bass. The same beating that makes a pad shimmer makes a sub rise and fall in level, which a limiter or a club system exaggerates.

• Subs and 808s Beating below about 100 Hz reads as uneven volume, and wide low end can partly cancel when summed to mono. Keep the sub a single, centered voice.
• Mono playback Voices panned left and right with different pitches can partly cancel when summed. Check wide patches in mono before you commit.
• Dense arrangements Several wide, many-voice synths fill the stereo field and blur each other. Give one part the width and keep the rest narrower.
• Tuning perception Heavy detune on a lead that plays with a vocal can sound out of tune. Reduce detune or blend the outer voices down.
• CPU and polyphony Each unison voice is a full voice. Serum's unison field changes color as a CPU reminder, and in Ableton's Drift the Unison mode uses four voices per note, so 32 voices give only 8-note polyphony.^[1][3]

Some synths offer a middle path for basses. Serum's Stack options include Center-12 and Center-24, which drop the center unison voices one or two octaves.^[1] Vital has similar Center Drop 12 and 24 modes, according to a community-written Vital user guide.^[4] Even then, a dedicated single-voice sub on its own track is easier to control.

How to check a unison patch for mono compatibility

Check unison in mono by summing the track and listening for level drops, hollow tone or notes that vanish. Sound On Sound's advice on stereo-widening processing ends the same way: always check the mix in mono so no unpleasant phase cancellation slips through.^[5]

1. Loop the busiest bar
   Pick the section where the unison part plays its lowest and highest notes.
2. Sum to mono
   Put a utility or imager on the track, or on the master temporarily, and switch it to mono.
3. Compare levels
   Toggle stereo and mono. A drop of more than a few dB, or a part that turns thin and phasey, means the width is costing you.
4. Reduce width first
   Lower the unison width or stereo spread before touching detune; width is usually the cause of mono loss.
5. Split by frequency
   If the patch must stay wide, high-pass the wide layer around 150-250 Hz and add a narrow, single-voice layer underneath.
6. Recheck in context
   Mono-check again with drums and bass playing. Cancellation that hides in solo can show up once the mix is full.

For a mix-stage view of the same problem, see how to mix wide leads without losing mono and why mixes fall apart in mono.

Unison vs chorus, spread and layering

Unison is not the only way to thicken a synth, and it is often not the cheapest. A chorus effect delays and modulates copies of the whole signal after the synth, so it costs one effect instead of extra voices per note. Built-in spread modes sit in between: Ableton's Simpler and Operator use two detuned voices per note panned left and right, and the Operator manual flags Spread as CPU-intensive.^[3]

Layering two patches gives the most control: a narrow layer for the body and a wide unison layer for air, each with its own filter and level. That is how you keep the size of a supersaw without handing the whole frequency range to a stereo effect. For the waveform choice under all of this, see oscillator waveforms explained.

CPU cost varies a lot between synths. Vital's developer says SSE optimizations keep its oscillators easy on the CPU and invites users to turn up the unison, while other synths show heavier load per voice.^[6] If a project stutters, freeze or bounce unison-heavy tracks before cutting voices you actually need.