Hume Octave 2 description field vs Sume TTS emotion controls

Hume's overview says Octave 2 (preview) supports a description field for voice customization and lists "Acting instructions" as coming soon. Sume TTS 1.0 already has a small set of controls under generation_config: an emotion string of up to 64 characters, speed from 0.6 to 1.5 and volume from 0.5 to 2.

Hume's page is the TTS overview, read 2026-10-01; Sume's controls are from the OpenAPI schema behind the API reference.

How do the controls differ?

One is free text about the voice; the other is a short guide plus two numbers.

What does the Sume emotion field do?

The schema calls it an "optional emotion guide for generation". It is one short phrase, not a script of stage directions, and the docs do not list an allowed vocabulary. Try a few phrases on a sample and keep the ones you like.

What does a request with all three look like?

Controls sit inside generation_config.

curl -X POST https://api.sume.com/v1/tts-1.0/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "transcript": "We did it. We actually did it.",
    "avatar_handle": "@narrator",
    "generation_config": { "emotion": "relieved and happy", "speed": 0.95, "volume": 1.2 }
  }'

Should I wait for Hume's acting instructions?

Only if the performance control is what you need and you can use Hume. The page gives no date for the feature, so do not plan around it. For what Sume ships today, see TTS with emotion, speed and volume.