Audio8 TTS (offline, voice clone, 11 languages)

What it does

Before you install

Audio8 TTS asks more of your computer than the light dubbing engines:

• Disk: about 1 GB for the model, downloaded on the first line you dub.
• Memory: about 1.6 GB while it speaks, and up to 3 GB for a moment the first time it hears a new speaker.
• Time: each second of speech takes two to three seconds to make, on an Intel Core i7-13700T. A ten-minute video can take about half an hour.
• Processor: any 64-bit Intel or AMD processor, or Apple Silicon. No graphics card is used.

For quick dubs, Sopro or Chatterbox clone voices about real time or faster, in fewer languages.

Voices

In the Speakers tab, choose Audio8 TTS as a speaker's Dubbing engine, then pick a Voice. Turn on Fit dub to subtitle duration for the speaker to stretch or squeeze each line to its subtitle.

1 voice.

Audio8 TTS has one voice, Speaker's own voice, and no preset voices: every line clones a speaker. Subtitld cuts a sample of that speaker from the project's audio, from one of their subtitles of 3 to 12 seconds, or several short ones joined. If you have separated music from voice for this video, the sample comes from the voice alone. The Fallback reference clip in the add-on's settings is used only when no sample can be cut.

Subtitld sends the subtitle text with the sample, and Audio8 TTS uses it to come closer to the speaker. Without it, for example when the project's original language isn't set, the add-on clones from the sound of the sample alone, a little less closely. If your subtitles paraphrase the speech instead of transcribing it, turn on Don't send the subtitles of the reference clip.

The first line of each speaker takes a few seconds more, while the add-on listens to their sample. It remembers it, so the speaker's next lines start right away.

Models

There is one model, Audio8 TTS Preview 0.6B, in the 4-bit ONNX build its makers publish for processors. It is not inside the add-on: it downloads into Subtitld's shared model cache the first time you dub, from Hugging Face at a pinned revision, each file checked. An interrupted download resumes where it stopped. Updating or removing the add-on keeps the model.

The add-on runs that build without PyTorch. With the same voice, line and seed, it gives exactly what Audio8's own runtime gives.

Options

In Settings › Add-ons › Audio8 TTS (offline, voice clone, 11 languages) › Configure.

Languages

Chinese, Cantonese, Japanese, Korean, English, Dutch, French, German, Italian, Polish and Spanish. A speaker cloned from a clip in one of them speaks all the others. A project in any other language gets a message that says so: pick another dubbing engine for it.

Audio8 calls this model a preview: these are the languages it recommends. We checked one sentence in each of them, with Whisper and a voice-similarity score, not by ear in every language.

• zh
• yue
• nl
• en
• fr
• de
• it
• ja
• ko
• pl
• es

Platforms

Linux x86-64, macOS on Apple Silicon and Windows x86-64. Each release is built on all three, and on each one CI installs the add-on as Subtitld does, downloads the model and dubs with it.

Troubleshooting

• network_unavailableThe model couldn't be downloaded. Check your connection and try again: the download resumes.
• disk_fullNot enough disk space for the model, about 1 GB.
• unsupported_languageThe project's language is not one of the eleven, or not set. Set it in the project, or pick another dubbing engine.
• unsupported_voiceThe project names a voice this version doesn't have. Pick the voice again in the Speakers tab.
• bad_paramsThe line has no text, or the cloned voice got no sample: open the video, or set a Fallback reference clip. The message says which.
• out_of_memoryThe computer ran out of memory. Close other programs: Audio8 TTS needs about 3 GB free.
• internalSomething failed inside the add-on; the message has the details.

License

Fine for commercial work: the add-on is Apache-2.0, and so are Audio8 TTS's code and the Audio8 TTS Preview 0.6B weights. The dubs you make are yours. A cloned voice sounds like a real person: clone voices you have the right to use, and say that a dub is synthetic where it matters.