Guides
Voice, sound effects and music for a reel: three calls, 0 credits
A twenty-second reel's soundtrack built from three ElevenLabs calls on aikizi. What voice, SFX and music normally cost, why ours cost nothing, and how to write each prompt.
Part 9 of 11 · ← previous · all parts · next → The gallery as a reference library
A twenty-second reel needs three sounds: a line of voice, one effect, and a bed of music. On aikizi those are 8, 12 and 24 credits — except that this season the first ten minutes of audio every day are free on every account, so the soundtrack below cost nothing. We built it in three calls. The honest part first: aikizi does not generate video, so this is sound for a reel you cut elsewhere.
What the three calls cost
| Call | Normal price | What we paid | What came back |
|---|---|---|---|
| Voice · one spoken line | 8 credits | 0 | An mp3 of the line, read once |
| Sound effect · 6 seconds | 12 credits | 0 | A 6-second mp3, exactly as asked |
| Music · 20 seconds, instrumental | 24 credits | 0 | A 20-second instrumental mp3 |
| Total | 44 credits | 0 | Inside the free daily ten minutes |
The free window is ten minutes a day per account, shared across voice, sound effects and music, resetting at midnight UTC. Ten minutes is a lot of audio — our whole reel used well under one. Past it, the prices above apply.
The picture it was built for
The three calls, as written
The line
Voice takes text and reads it. The whole craft is in the writing, because you cannot direct the delivery much: short sentences, concrete nouns, and a full stop where you want a breath. Ours was two sentences: "The lamp is lit before the street wakes. One flame, one step, and the stone still wet from the night." Twenty-one words, which lands at about eight seconds — roughly two and a half words a second is a useful figure when you are writing to a duration.
The effect
Sound effects take a description and a duration between half a second and twenty-two seconds. Describe the sequence, not the object: "A small oil lamp wick catching light, then a soft steady flame, with faint early-morning street ambience and a distant temple bell." Naming the order — catching, then steady — is what stops it returning six seconds of undifferentiated crackle.
The bed
Music takes a style description, a duration up to 180 seconds, and an instrumental flag. Set instrumental true for anything with a voiceover over it, or you will be fighting two voices. Ours: "Slow devotional instrumental, solo bansuri flute over a low tanpura drone, a single soft tabla pulse entering late, dawn raga feeling, unhurried, no vocals." Naming the instruments does more than naming the genre.
What went wrong, and it is worth knowing
We cannot play these files to you on this page. Generated audio lands in your own library behind your account, and the media address needs a signed-in session — an anonymous request for one returns a 404. So the three files are real, they cost what the table says, and they are in the library they were made in; this page can tell you about them and cannot embed them. If you want to hear the technique, the cheapest route is to make the same three calls yourself, which today costs nothing.
Cutting the three together
- Write the line first and time it at about two and a half words a second. The line decides the length of everything else.
- Generate the music at the full length of the reel, instrumental, so you are not looping a short bed.
- Generate the effect for the exact seconds it occupies, not for the whole reel.
- In your editor, lay music under everything, drop the effect on the action, and put the voice on top; pull the music down about 10 dB under the voice.
- Leave the last two seconds of music running past the voice. A reel that ends on a spoken word feels cut off.
The limit, said plainly
There is no video generation on aikizi. It was retired, and no page in this series will tell you otherwise. What this site gives a reel is the stills, the sound and the decode of a reference clip — video and audio decode both still work, at 10 or 20 credits, on the Pro plan. The cutting happens in your editor.
FAQ
How much do voice, sound effects and music cost on aikizi?
8 credits for voice, 12 for a sound effect and 24 for music. This season the first ten minutes of audio each day are free on every account, shared across all three and resetting at midnight UTC, so a short reel's soundtrack usually costs nothing.
Can I generate video on aikizi?
No. Video generation was retired. You can generate the stills and the sound, and you can decode a video or a piece of audio at 10 or 20 credits on the Pro plan, but the cutting happens in your own editor.
How long should the voice line be for a twenty-second reel?
Write to about two and a half words a second. Our twenty-one-word line ran about eight seconds, which left room for the effect and a music tail.
How do I write a sound effect prompt?
Describe the sequence rather than the object — "a wick catching, then a steady flame, with faint street ambience" — because naming the order is what stops it returning undifferentiated noise. Durations run from half a second to twenty-two.
Should generated music be instrumental?
Yes, for anything with a voiceover over it. Two voices compete and you will lose the one that matters. Name the instruments rather than the genre; it does more work.
The free ten minutes reset every day, so the cheapest experiment available here is three audio calls in the Studio. The stills that go over them are part 1, and the whole run-through is part 11.
Part 9 of 11 · ← previous · all parts · next → The gallery as a reference library
