Blog · AIKIZI Sound · Decode to voice
A voice for your decoded character
From a decoded picture to a spoken line, a sound bed and a minute of music: read MOOD and WORLD as a casting note, pick an ElevenLabs voice, write three lines, build the room. About two of your ten free minutes.
Take the MOOD and WORLD tabs of a decode and turn them into a casting note: age, warmth, pace, accent. Pick the nearest voice in Studio, write three lines that could only be this character, tag one of them, and press Create. Then make the room they stand in, and the music under it.
A picture of a character is a promise. The voice is where the promise gets kept or broken. Most people never get to hear their characters because casting a voice used to mean money and a week. With Eleven v4 in AIKIZI Studio it means a casting note, a voice picker, and about a minute of the day's free ten.
Read the decode for the voice
Open a decode of the character, yours or one from Explore. Each tab carries a clue for the ear.
Age and presence
A child, a soldier, an elder. This sets the register before anything else.
Temperature and pace
"Serene", "coiled", "wry". Serene is slow and low. Coiled is quick and close to the microphone.
Where the voice grew up
A port city, a monastery, a space station. The world suggests an accent, or the lack of one.
Texture
Brass and leather ask for a dry, close voice. Silk and mist ask for breath and air.
Woman, late twenties. Serene but watchful. Low register, unhurried, close to the mic. Grew up bilingual in a port city; a faint accent, never played up. Dry, warm, no breathiness.
Pick the voice
In Studio, choose ElevenLabs Voice (V4) and open the voice picker. The voices are a curated set from ElevenLabs' library, chosen so that most casting notes have a near neighbour. Play three previews against your note. Choose the one that is right at rest; tags can add a laugh or a whisper, but they cannot change who the voice is.
Change one word of the note and look again. "Serene" and "tired" sit in different voices. Casting is finding the voice that does not need direction for its first line.
Write three lines
Not a speech. Three lines that could only be this character, in three different states, so you can hear the range.
At rest
Something ordinary said their way. "The tea is cold. Sit anyway."
Under pressure
Short words, a tag if it helps. "[quietly] Not here. Not now."
Alone
The line they would never say to anyone. "[sighs] I could have gone with him."
Paste the three lines as one take with a blank line between them. Press Create. Listen for one thing: does the voice at rest sound like the picture? If yes, the character exists now. Read Eleven v4: voices that act for the tags and the writing rules.
Give the world a sound
The WORLD tab is a shot list for sound. A port city at dusk is gulls, distant cranes, water against a hull, a bed you can loop. Switch Studio to SFX, describe the bed in one sentence with "no voices, no music", set it to twenty seconds, flip Loop. Then switch to Music and write a sixty-second instrumental brief that takes its mood from the MOOD tab: "warm and watchful, sparse piano and a low cello, 72 BPM in D minor". The sound and music guides go deeper on each.
The ten-minute budget
| Take | Model | Uses |
|---|---|---|
| Three lines, one take | Eleven v4 | about 0:20 |
| Port at dusk, loop | ElevenLabs SFX | 0:20 |
| Sixty-second bed | Eleven Music | 1:00 |
| Two more casting tries | Eleven v4 | about 0:40 |
| Total | about 2:20 of 10:00 |
That leaves most of the day's free minutes for the next character. Everything lands in your Library with a download, ready for the edit.
Hear your character
Decode the picture, cast the voice, make the room and the music. Ten minutes of ElevenLabs audio a day, free this season.
Open StudioHow the free minutes work
