Dimau Voice Studio — Voices, Dialogue & Music

Voice timeline

Browser voice audition

🎙 Record a real voice

With microphone permission you can record your own narration, download it, or send it directly to the Video Studio timeline on this device.

Timeline sec

MP3 / WAV production

Generate genuine speech audio with local open models or optional paid voices. Export a Windows render kit, run it on your PC, then import its result ZIP directly into Video Studio. Re-running in the same folder reuses unchanged audio; output captions use measured speech lengths.

Chatterbox / Kokoro models download their weights on first use; inference is local afterward. Browser previews are not recorded to audio files. Paid providers are never called by this web page.

🎹 Melody & singing workflow

Build a melodic guide with note pitch, beat position and lyric syllables. Download MIDI to use with your legitimately licensed singing synthesizer. Instrument preview is not a generated singing voice.

Tempo BPM

Note

Beat

Beats long

Syllable

Voice Studio ready. Edit a script cue or record your voice.

Selected dialogue

Character / actorScript and expressive directions

Start (sec)

Target length (sec)

Direction

Render engine

Engine voice ID

Language (BCP 47)

Pace (0.5–2)

ElevenLabs voice controls

Speech model

Stability (0–1)

Similarity (v2 / Flash)

Style strength (Multilingual v2) Speaker boost (Multilingual v2)

v3 uses expressive directions and stability 0, 0.5 or 1; keep pace at 1. Multilingual v2 and Flash support pace 0.7–1.2; expressive tags are removed. These settings apply to the downloaded render kit, not the browser audition.

Preview pitch (0.5–2)Browser audition voice

Voice reference & permission

Only use your own voice or recordings for which you have explicit rights. Chatterbox reference audio stays in your downloaded local render kit.

Reference recording (optional, ≤10 MB) I am the speaker or have the speaker's explicit consent for synthetic voice generation.

No voice reference selected.

Scope and provider limitations

ElevenLabs and Murf: paid-key local render integration. Hume Octave: API ending November 2026; not linked. PlayHT: unavailable. LOVO and Descript: inspiration for editing workflows; no undocumented endpoints used. Synthesizer V: MIDI export only; no proprietary synthesizer code copied. Fiji language/accent support must be evaluated per installed model.