Voice for your character

Generate a voice sample from your character image, then reuse it so your character sounds the same in every clip. Listen to the sample before you move on.

Prima di iniziare

Have your character reference image on the canvas. A three-view image works well. Write one or two sentences describing the voice: pitch, texture, pace, accent. Decide whether you will use Seedance 2.0 or Seedance 2.5, since this sets how long your voice sample can be. Each voice and video generation costs credits.

Choose your character's voice

Click your character reference image.
Select Audio.
Select Generate Voice.
Select Seed Audio 1.0.
In the prompt, @ your character, briefly describe the voice you want, and add a line for the character to say.
Generate.

Flick will generate a voice based on your character and description.

The line can be anything. It does not need to match the dialogue you use later.

How long should the voice be?

Seedance 2.0: keep it under 7 seconds. Seedance 2.0 accepts at most 15 seconds of audio, so if a second character joins the clip, each character only gets about 7 seconds.
Seedance 2.5: the voice can be longer than 15 seconds.

Creative direction: why this voice

Sella's voice comes from her backstory. She is a noble living under a false name, quietly investigating her own family. She commands powerful magic. She is young, but she has seen a great deal.

Her final voice description:

early 30s, slight British accent, low alto voice, smoky and velvety with a slight rasp, rich chest resonance, slow controlled pace, cool and commanding with a trace of dry amusement

Each part of it comes from who she is.

Raised as a noble: slight British accent, velvety tone, slow controlled pace. She grew up expecting to be listened to, so she never rushes.

Living under a false name: cool. She keeps her voice level and gives little away. A bright, sweet voice invites people closer. Sella needs to keep them at a distance.

Powerful magic: low alto, rich chest resonance, commanding. Her authority does not come from volume. The depth suggests she holds more power than she shows.

Young, but worldly: early 30s, smoky with a slight rasp, a trace of dry amusement. She is young, but her voice sounds a little older than she looks. The rasp and the dry humor tell the audience she has seen more than her years suggest.

A bright, sweet voice would make Sella sound innocent and easy to read, the opposite of who she is. When you choose a voice, start from the character's story: who she was raised to be, what she hides, and what she can do.

Genera il tuo video

Click your character image, scene image or voice clip.
Select Video.
Select Omni Reference.
Select Seedance. Other models also support this.
Type in your prompt.
In the prompt, write "reference voice and accent from" and @ your voice clip.
Generate.

Flick will generate a video using your character's voice.

Same voice, four scenes

Create a 5-second continuous live-action shot. Use @Image2 for the location, firelight and palette. Sella's identity matches @Image1. Use @Audio1 for Sella's vocal identity and accent; the spoken words for this shot are the line below.

Sella sits against the stone wall in the empty spot beside the dying fire, turning one coin from her necklace between her fingers, eyes on the coin and already wet. The fire pops once; she stops turning the coin. Sella (English, low and quiet in her normal voice for the first sentence, holding herself together; a short breath catch before the second sentence; her voice starts to tremble on "I'm still" and cracks on "waiting"): "You promised me a drink. I'm still waiting." After the last word she presses her lips together and closes her fist around the coin; one tear runs down her cheek and she lets out one shaky breath through her nose.

Slow push-in from medium close-up toward close-up that stops when the fire pops; then the camera holds still. The firelight falls on one side of her face; the other side stays in darkness. Sound consists of Sella's voice, one breath catch, one shaky exhale, soft fire crackle and one pop. No music.

Create a 5-second continuous live-action shot. Use @Image2 as the exact opening frame, composition and lighting; the frame opens empty. Sella's identity matches @Image1. Use @Audio1 for Sella's vocal identity and accent; the spoken words for this shot are the line below.

Sella walks into the frame from the bottom edge, her back to the camera, and climbs the dark near slope toward its crest at a steady, unhurried pace. Halfway up, without turning, she speaks. Sella (English, low and level, quiet certainty, no rise in volume): "Let them come. I'm done running." On the last word she stops on the crest, directly in front of the sun's glow, in full silhouette; the blowing dust streams past her and she stays still to the end.

The camera is locked and does not move; the composition of @Image2 stays unchanged, and Sella grows smaller as she walks away from the camera. Sound consists of Sella's voice, close and clear despite the distance, her footsteps on loose rock, steady wind and cloth flapping. No music.

Create a 5-second continuous live-action shot. Use @Image2 for the location, moonlight, mist and palette. Sella's identity matches @Image1. Use @Audio1 for Sella's vocal identity and accent; the spoken words for this shot are the dialogue below.

Sound rules for the whole shot: Sound consists of Sella's whispered dialogue, two footsteps, one soft rustle of fern fronds, and quiet night forest ambience of faint insects at one constant low level, present from the first frame through the last frame. Nothing else is heard. No background music; only the dialogue, the ambience and the action sounds remain. No score, no soundtrack. No drum hit, no low boom, no bass hit, no impact hit, no stinger, no riser, no swell, no whoosh, no fade-in, no fade-out, no transition effect. No subtitles.

0-1s: Sella lies hidden among the ferns; only her eyes show between the fronds, the lower half of her face hidden. She is completely still. Only the steady night ambience is heard. 1-2s: <One heavy footstep lands on soft forest ground somewhere in the dark, dull and muffled.> <The fern fronds rustle softly as they tremble.> Her eyes lift toward the source of the footstep and lock there. The ambience stays at the same level; no drum, boom, bass hit or stinger marks this moment. 2-4s: Dialogue language: English. Sella whispers to herself in a half-voiced whisper with her own voice still audible, slow and shaky, steadying herself: {It's okay, Sella.} 4-5s: <A second footstep lands, slightly closer, equally dull and muffled.> She stays motionless and unblinking, eyes fixed on the source of the sound. The ambience holds at the same level to the last frame, then the sound stops cleanly.

The camera stays locked on an extreme close-up of her eyes, seen through fern fronds in the foreground, for the whole shot. Cold moonlight from above catches her eyes; the rest of her face stays in darkness.

Create a 5-second continuous live-action shot. Use @Image2 for the location, light and palette of the background: the burned village, cold overcast light, drifting ash. Sella's identity matches @Image1. Use @Audio1 for Sella's vocal identity and accent; the spoken words for this shot are the line below.

Sella's hand hangs at her side, fingers trembling. A flake of ash lands on her knuckles; her fingers slowly curl into a tight fist. Sella (English, low and shaking, forced through clenched teeth, rising on the final words): "I'll burn every last one of you." On the last word she opens her hand; a blue flame gathers in her palm and holds there, lighting the falling ash.

Locked extreme close-up on her hand at waist height, with the ruins of @Image2 heavily out of focus behind it; the camera does not move. The blue flame is the only color in the frame. Sound consists of Sella's voice and breath, near silence with faint wind, then the low hiss of the flame igniting and holding. No music.