Mix Bus
Architecture

Set your routing before you touch a fader

Build your bus architecture first. Empty buses cost nothing; re-routing after you've already set levels costs hours. The core principle: group things that belong together, process groups before the master, keep parallel processing separate from the main signal path.

The three problems architecture solves

The signal hierarchy: tracks → groups → master

Individual tracks route to group buses, not directly to the master. Group buses route to the master. Send/return buses receive post-fader sends and return to the master alongside the groups. Click any node below for detail.

Group buses

The four standard groups

What belongs on a group bus

Fix problems at the track level before the signal hits the bus. The bus is for shared processing that benefits the whole group: glue compression, low-end management, saturation.

Drum Bus processing chain:
  1. Glue Compressor (SSL G-Bus emulation)
     – Ratio:     2:1
     – Threshold: −18 to −22 dBFS
     – Attack:    30 ms  (let the transients breathe)
     – Release:   auto
     – Knee:      soft
     – GR target: 2–4 dB
     – Makeup:    +2 dB

  2. Optional: Drum Buss (Ableton)
     – Crunch:    very subtle (adds harmonic density)
     – Boom:      off or minimal
     – Transients: slight enhance

  3. Post-fader send to parallel comp return
     – Send level: start at −inf, raise slowly

  Bus fader: −3 to −6 dB below unity

Bass Bus processing chain:
  1. Compressor
     – Ratio:     3:1
     – Threshold: −14 to −18 dBFS
     – Attack:    10–20 ms
     – Release:   auto
     – GR:        3–5 dB

  2. EQ Eight (M/S mode)
     – Mid channel: high-pass at 30 Hz (remove DC)
     – Side channel: low-pass at 120 Hz (mono enforce)
     – This collapses bass to mono below 120 Hz
       without affecting the stereo mid-high content

  3. Subtle saturation (optional)
     – Adds 2nd/4th harmonics at 80–160 Hz
     – Makes sub audible on small speakers
     – Keep very subtle: 10–20% drive max

  Bus fader: −3 to −6 dB below unity

Parallel processing buses

Parallel routing keeps the original signal untouched while a heavily processed copy blends alongside it. Unlike a plugin's wet/dry knob, you get an independent fader on the processed return and can add more processing to it separately.

Parallel compression (New York compression)

Send the drum bus to a heavily compressed return: 8:1 or higher, threshold so low the compressor never lets up. Soloed, it sounds crushed and lifeless. Blended under the dry at −12 to −20 dBFS, it fills the gaps between transients with dense sustained energy. Punch stays in the dry. Body comes from the return.

Parallel Compression — Return Track Setup:

  Return track: "PARALLEL COMP" (or "NY COMP")
  Receiving: post-fader send from Drum Bus (and optionally Bass Bus)

  Compressor settings (on the return track):
    Ratio:      8:1 – 20:1   (crush it — this should sound bad)
    Threshold:  −24 to −30 dBFS  (compressor fires constantly)
    Attack:     5–15 ms
    Release:    50–100 ms
    Makeup gain: +10 to +15 dB  (bring the crushed signal back up)

  Return fader: start at −∞, raise slowly until you feel density
  Target blend: return sits about 12–20 dB below the dry drum bus peak

  Test: solo the return. It should sound terrible — pumping,
  distorted, no transients. If it sounds okay, compress harder.
  The effect only works in the blend.

Parallel saturation

Same routing, saturator instead of compressor. A heavily saturated or bitcrushed parallel signal under an 808 adds upper harmonics that carry through small speakers, which is why this technique is everywhere in trap and hip-hop.

Parallel Saturation — Return Track Setup:

  Return track: "SAT RETURN"
  Plugin: Saturator / Overdrive / Redux (bitcrusher)

  Settings vary, but the principle:
    – Drive the saturator hard (50–80% input gain)
    – Target 2nd and 4th harmonic generation (tube/tape modes)
    – EQ the return: HPF at 200 Hz (remove low-end mud from saturation)
    – LPF at 6–8 kHz (remove harsh high-frequency artifacts)
    – Blend at −15 to −25 dB below dry signal

  On 808/sub bass: saturation return makes the sub audible on
  small speakers that cannot reproduce the fundamental frequency.

Send/return effects buses

Reverb and delay go on send/return buses, not inserts. Insert reverb gives each track its own room; shared send reverb puts everything in the same space. That shared decay is what makes a mix sound like a performance instead of isolated tracks stacked together.

Always post-fader for reverb and delay. Post-fader means the reverb tail stays proportional to the source as you move the fader. Pre-fader leaves the reverb running after you pull the source down, so you hear the room with nothing in it.

Setting up the shared reverb return

Shared Room Reverb — Return Track Setup:

  Return track: "REVERB" (or "ROOM")
  Reverb plugin set to 100% wet ONLY (never add dry signal to the return)

  Recommended starting settings for electronic music:
    Pre-delay:  18–22 ms   (separates source from reverb onset)
    Decay time: 1.0–1.5 s  (versatile for 100–140 BPM)
    Early reflections: medium density, medium size
    Diffusion:  medium–high

  EQ on the return (mandatory):
    HPF at 200–300 Hz  (prevent reverb muddying the low end)
    LPF at 8–10 kHz    (soften harsh reverb high-frequency content)
    Optional: gentle low-mid cut at 400 Hz (reduce boxiness)

  Subtle compression on return (optional):
    2:1, slow attack, fast release
    Prevents reverb tail from swelling louder than the source

  Send levels from each track:
    Drums (snare, hats): medium send (−12 to −18 dB)
    Kick: very low or no send (reverb on kick muddies low end)
    Bass: no send (reverb on bass destroys low-end clarity)
    Lead: medium send (−12 to −15 dB)
    Pads: heavy send (−6 to −12 dB)
    Vocals: medium-heavy send (−10 to −14 dB)

Using multiple reverb buses

One reverb creates cohesion. Two gives you separation: a tight room for drums, a large hall for pads. Use the table below to pick the right type.

The master bus

Everything sums here, so every move you make affects all elements at once. The master bus works well for processing that genuinely benefits everything equally, but it's not where you fix upstream problems. If the kick is too loud, turn down the kick track, not the low end on the master EQ.

The master bus processing chain

Master Bus Chain (in order):

  1. Spectrum analyzer (not processing — always running)
     – SPAN or similar, to catch low-end buildup and frequency imbalances

  2. Gentle EQ (optional — only if needed)
     – Broad, low-Q moves only (Q ≤ 0.5)
     – Example: gentle high shelf +1 dB at 12 kHz for air
     – Example: very gentle low shelf −1 dB at 120 Hz if muddy
     – Nothing sharp, nothing surgical — those fixes belong upstream

  3. Glue compressor (SSL G-Bus / Glue Compressor)
     – Ratio:      2:1
     – Threshold:  −3 to −5 dBFS  (catches only the very tops)
     – Attack:     10–30 ms
     – Release:    auto
     – Knee:       soft
     – GR target:  1–2 dB MAXIMUM
     – This should barely move the GR meter — it is glue, not leveling

  4. Saturation (optional, very subtle)
     – Harmonic exciter at 2nd/4th harmonics
     – 5–15% drive maximum
     – Adds very subtle warmth and presence
     – Use with extreme restraint — it changes the entire mix character

  5. Limiter (brickwall, final stage)
     – Ceiling: −0.1 to −0.3 dBTP (true peak, not dBFS)
     – Input gain: set so loudest passages cause 1–3 dB limiting max
     – ISP prevention: ON
     – Leave headroom for mastering: pre-limiter peak should be −3 to −6 dBFS

What not to fix on the master bus

Gain structure across the bus hierarchy

Summing raises the level. 16 drum tracks at −18 dBFS don't sum to −18 dBFS at the bus, since they sum louder. Pull the bus fader down to compensate; don't chase it by turning down individual tracks.

Stem exports

A stem is a group bus bounced to a stereo file (internal processing printed, master bus excluded). Sum all stems together and you get the original mix back. Used for stem mastering, remix packages, live performance, and archiving.

Group buses must be self-contained

A stem only captures what passes through that group bus. If any processing returns outside the bus (parallel chain directly to master, etc.), the stem is missing that element. Everything contributing to a group needs to pass through the bus before the master. Reverb returns are the exception: stems are typically bounced dry, with reverb exported separately or re-applied by the mastering engineer.

Stem export checklist (Ableton):

  For each group:
    1. Solo the group bus
    2. Mute all return tracks (reverb, parallel comp) unless
       specifically including them in a "wet" stem
    3. Export: File > Export Audio/Video
       Format:    WAV, 24-bit, 44.1 kHz (or session sample rate)
       Normalize: OFF
       Master:    OFF (export pre-master-bus)
       Loop:      match your arrangement, with tail padding (4–8 bars extra)

  Naming convention:
    ProjectName_DrumStem_v1.wav
    ProjectName_BassStem_v1.wav
    ProjectName_SynthStem_v1.wav
    ProjectName_VocalStem_v1.wav
    (Optional) ProjectName_ReverbReturn_v1.wav
    (Optional) ProjectName_ParallelComp_v1.wav

  Verification:
    Import all stems into a new empty session
    Sum them together — the result should match your original mix
    (minus master bus processing).
    If they don't match, something was routed incorrectly.

DAW-specific setup

Ableton Live

Ableton Bus Architecture:

  Group Tracks → act as group buses
    Right-click any track → Group Tracks (Cmd+G / Ctrl+G)
    The group track IS the bus — it receives the summed output
    of all tracks inside it
    Add processing on the group track's device chain

  Return Tracks → parallel / send buses
    Created via Create > Insert Return Track
    Sends appear on every channel's send section (A, B, C...)
    Set to Post-Fader for reverb/delay
    Set to Pre-Fader for headphone monitoring sends
    Return tracks route to the master by default

  Parallel compression routing:
    Method 1: Use a Return track as described above
    Method 2: Use an audio track with "Audio From" set to
              the group track — set monitor to "In" for live routing
              This method allows processing the parallel signal before
              it reaches the master and gives a dedicated channel fader

  Master Track → the master bus
    Add processing here: Glue Compressor, EQ, Limiter
    Limiter always last in the chain

Logic Pro

Logic Pro Bus Architecture:

  Summing Stacks → group buses (Logic 10.3+)
    Create Folder Stack or Summing Stack from selected tracks
    A Summing Stack creates an aux channel that acts as the group bus
    Processing on the stack master channel = group bus processing

  Aux Channels → parallel / send buses
    Already present in Logic's Mixer
    Route sends from tracks to Bus 1, Bus 2, etc.
    Create Aux channels in the Mixer to receive those buses
    100% wet on the Aux for reverb/delay returns

  Manual bus routing:
    Track output → Bus 1
    Aux channel input → Bus 1
    This is the fundamental Logic parallel processing setup

  Master Output → master bus
    Fader channel labeled "Stereo Out" in the Mixer
    Add processing here for master bus chain

FL Studio

FL Studio Mixer Architecture:

  Mixer channels act as tracks AND buses
    Each channel can receive audio from instruments or other channels
    Route any channel into any other via the routing section

  Creating group buses:
    Create a mixer channel labeled "DRUM BUS"
    Route drum instrument channels into it:
      Click source channel → enable "Route to" for DRUM BUS channel

  Send channels → parallel / return buses
    Create a new mixer channel (e.g., "PARALLEL COMP")
    Enable it as a send from your drum bus channel
    Set the send to 100% wet — add processing on the send channel
    The send returns automatically to the master if routed there

  Master channel (channel 1 in Mixer) → master bus
    Always at position 1
    Processing added here applies to the full mix output

Routing hierarchy and bus processing cheat sheet

Signal routing hierarchy

Group bus processing guide

Glue Comp: 2:1, −20dBFS threshold
Attack: 30ms  Release: auto
GR: 2–4 dB   Knee: soft
Saturation: optional, subtle
Fader: −3 to −6 dB

Compressor: 3:1, −16dBFS threshold
Attack: 10–20ms  Release: auto
M/S EQ: low-pass S channel at 120Hz
(enforces mono below 120Hz)
Saturation: subtle for small speakers

Light comp: 2:1, high threshold
HPF group at 120–200 Hz
(removes layered synth mud)
Stereo check: verify mono compat
Fader: −3 to −6 dB

Light comp: 2:1, gentle
EQ: cut competing frequencies
Level automation across sections
Typically lower fader than other groups
(−6 to −12 dB in electronic music)

Ratio: 8:1–20:1 (crush it)
Threshold: −24 to −30 dBFS
Attack: 5–15ms  Release: 50–100ms
Makeup: +10 to +15 dB
Blend: return at −12 to −20 dB under dry

100% wet (never dry on the return)
Pre-delay: 18–22ms
Decay: 1.0–1.5 s
HPF return at 200–300 Hz
LPF return at 8–10 kHz
No send from kick or bass

EQ: broad strokes only (Q ≤ 0.5)
Glue comp: 2:1, −3 to −5 dBFS
1–2 dB GR maximum
Saturation: optional, very subtle
Pre-limiter peak: −6 to −3 dBFS

Ceiling: −0.1 to −0.3 dBTP
ISP prevention: ON
Soft clip before brickwall: optional
GR during mixing: 1–3 dB max
Loudness decisions: leave to mastering

Common routing mistakes