Skip to content

Seed Audio Added to External TTS Engines ​

Updated

Set engine to tts-sa inside ExtParam and Seed Audio runs on the existing TTS APIs. With this addition, MPS now offers seven external TTS engines.

  1. Select the engine: pass {"engine": "tts-sa"} in ExtParam (a JSON string). synExt.model currently supports tts-sa-v1 only.
  2. Select the generation mode: what you put into engineCustom.references decides the path — text only, text + image, or text + audio.
  3. Select subtitles: set engineCustom.audio_config.enable_subtitle to true and word-level timestamps come back in ExtInfo. The default is false.
  • It is the only engine that can generate speech from an image: pass a reference image in references (image_url or image_data, at most 1 and ≤ 10 MB) and the engine picks up style cues from it.

  • One request can cite up to 3 reference clips: address them as @audio1 / @audio2 inside text_prompt, where upload order maps strictly to the number. Getting the order wrong raises no error — it simply produces the wrong voice, so take care.

  • Passing references makes the outer VoiceId ineffective. This is a parameter explicitly exempted from the "outer level wins for duplicate keys" rule; together with the identically named parameter on tts-f there are only two.

  • Cloning is not yet supported: there is no CloneVoice path, so no reusable voice. Reference audio does work, though.

  • vol is a [0.5, 2.0] multiplier, not the (0, 10] scale other engines use. With ogg_opus, sampleRate supports 48000 only.

  • It is slow, so the async API is recommended. Four APIs are supported: TextToSpeech, TextToSpeechAsync, DescribeTextToSpeechAsyncTask and DescribeVoices.

  • TTS-Seedaudio API docs

  • MPS TTS Engines Compared

  • AI Dubbing docs

Built with VitePress · Deployed on Tencent EdgeOne Pages