When a generated part fights a lead vocal, turning it down may help but does not explain the conflict. The issue can be shared register, similar attack shape, a busy lyric-adjacent rhythm, or an arrangement that never changes density. Start by deciding what the vocal is doing at each moment: carrying words, holding a note, answering itself, or leaving a gap. Then give the accompaniment a distinct role.
Map the vocal’s active zones
Listen once without looking at the mixer. Mark where intelligibility matters most: phrase starts, consonant-heavy lines, names, and harmonic turns. Mark where the singer leaves air. This map is more actionable than a generic instruction to “make room.” It tells you when an accompaniment can rise, not only when it must disappear.
Separate roles before EQ
Try an octave change, a different inversion, or a later rhythm before reaching for a narrow EQ cut. A pad that speaks between words may need fewer frequency changes than one that plays through every syllable. If a generated arpeggio mirrors the vocal’s rhythm and pitch range, simplify it to chord pulses or use it as a response at phrase endings.
Automate density, not just level
Mute notes, shorten releases, pull a send down, or substitute a single sustained tone under the lyric. These are arrangement moves. Use level automation after the part has a defensible role. Bus routing matters here: aux tracks route other tracks and can carry shared processing, so a change to a shared effect can alter vocal space as well as the instrumental.
Worked example: chorus with a synthetic choir
The chorus vocal leaps upward on the title phrase while a generated choir holds close harmony in the same range. Keep the choir in the first half of each bar, but remove its upper note for the title phrase. Raise a lower octave response after the phrase ends and let the vocal’s own delay fill the gap. At equal overall loudness, compare this against simply ducking the entire choir. The edited version should make the lyric clearer without making the chorus feel empty.
Limitations
Space is contextual: a dense vocal arrangement may intentionally blur words, and a sparse song may need a more exposed lead. Do not assume an EQ move is reversible across playback systems. Keep a dry vocal safety track and preserve permission/provenance records for any generated voice-like material. Continue with level-matched referencing and effect printing with options when handing the session to someone else.
Sources & working limits
- Apple: Logic project basics
Aux and output tracks route other tracks. - Apple: route audio via send effects
Sends split signal into parallel aux paths and return to output. - Apple: export takes and comps
Comps can be exported while original take content remains.
Sources checked 2026-09-19. Product instructions describe the documented version; controls and availability may change. Session examples are illustrative, not measurements or listening-test results.