You are a female narrator voicing male characters, other women, monsters, gods, and whatever your author dreamed up at 2 a.m. This page is your cheat sheet. Print it, tattoo it, just use it.
~80% performance. ~20% processing. Your listeners already know you're a woman — that's not a secret, it's not a problem. Your job isn't to fool anyone. It's to suggest a distinct, consistent person so convincingly that their imagination does the rest of the heavy lifting.
Record your takes clean and dry. Shape everything in post. Never live-process your real performance take — you'll chase the effect instead of the character and end up with both sounding bad. A real-time modulator is an audition toy, not a production tool.
Forcing your pitch as low as physically possible is the #1 beginner mistake. It's unsustainable across a 9-hour audiobook, it strains your voice, and it reads as caricature — not character.
A suggested tenor or baritone will always land better than a tortured attempt at bass. And remember: female → other female character work is dramatically easier. Lean into that whenever the cast allows.
Auburn Sounds Graillon 3 — free version includes formant shifting. This is the legitimate free twin of the $99 Soundtoys Little AlterBoy. You do not need to spend money to sound convincing.
Cleanup: your DAW's built-in noise reduction + EQ is enough to start. Stop waiting for perfect gear.
Pitch & formant plugin on your own voice = post-production processing. ACX-allowed. You performed it, you processed it, it's yours.
AI voice conversion into a meaningfully different voice = AI synthesis. Banned on ACX — except the invite-only "Voice Replica" beta (your voice, your approval, labeled).
The test is simple: did a real human perform it? If yes and you massaged it with DSP, you're good. If no, or if an AI generated the voice, that's synthesis — read the platform rules carefully.