The cloned-voice choir happens because Suno generates one vocal timbre and stacks it. A real choir is dozens of different voices — you have to fake that variety.
1. Name the choir precisely: 40-voice gospel choir, mixed male and female, rich blend beats 'choir.' Size words ('40-voice,' 'full choir') and blend words ('mixed,' 'rich blend') matter.
2. Write divisi parts. Don't give the choir one line — write parts:
[Choir: sopranos carry the melody, altos and tenors harmonize below]
Multiple written parts force multiple rendered voices.
3. The layering trick (best results): generate the choir section 2-3 times with slightly different style wording (gospel choir, church choir, warm, soulful choir, deep), then layer them in a DAW, panned left/center/right. Three different 'choirs' = instant depth and realism.
4. Keep the choir in its lane: [Choir enters only on the chorus, powerful] — choirs that sing everything sound fake; choirs that ARRIVE sound epic.
5. Exclude: solo voice, thin choir, single singer.
Good luck! Hank founder of SonorixPG
