Voices

A voice is what gets used when Popp generates narration or character dialogue. You have three ways to get a voice you want:

1. Pick a built-in voice from the ElevenLabs library.
2. Design a new voice from a text description.
3. Clone a real voice from an audio sample (yours, with permission).

Voices live in your library and can be assigned to any character, or used directly for narration in any movie.

Picking a built-in voice

When you create or edit a character, the right-hand panel has a Voice field. Click it to open the voice picker — you can browse, preview, and select from a library of high-quality voices in many accents and styles.

You can also tell Popp directly:

Use a deep, warm British male voice for this character.

Popp will pick a matching one from the library.

Designing a voice

If no built-in voice is right, you can design one from a text description. Popp will use ElevenLabs Voice Design to synthesize a brand-new voice from your description.

Design a voice: female, mid-30s, slight rasp, warm but with a slight skeptical edge. Think NPR host meets podcast comedian.

When you’re happy with the result, save it to your library and assign it to a character.

Cloning a voice

You can clone a real voice from an audio sample — yours, an actor’s, or anyone else’s (with permission).

You’ll need a clean audio sample, ideally 30 seconds or more of someone speaking with no background music or noise.

There are two underlying voice systems Popcorn uses:

• ElevenLabs voice clone — used for narration and TTS in most formats.
• Kling voice clone — used by the Kling video model when characters speak inside generated video clips (so the lipsync matches the cloned voice).

Popp will clone the voice into the appropriate system based on what you’re using it for. You can also explicitly ask it to clone into both.

Cloning a real person from a YouTube video

For UGC ads or reaction videos featuring a real creator, you can ask Popp to:

Cast a real person — find reference images and clone their voice from this YouTube link: <URL>.

Popp will:

1. Pull frames and audio from the YouTube video.
2. Generate a character based on the visual references.
3. Clone the voice from the audio.
4. Save both as a global character + voice you can reuse.

You’re responsible for ensuring you have the rights to use someone’s voice and likeness.

Where voices show up

• In any character — as the Voice property.
• In any narration — Popp picks a voice (or you can specify).
• In any HeyGen avatar talking-head — the voice drives the lipsync.
• In your sidebar under Elements → Voices — your voice library.

Tips

• Match the format. A grand cinematic narrator doesn’t suit a casual TikTok UGC ad and vice versa.
• Test before committing. When you design a voice, generate a 5-second test line first to see how it sounds in context.
• Reuse aggressively. Voices are how a recurring host sounds the same every time — even more important than how they look.