#elevenlabs <break time=x.xs> tag misses next word to pronounce in tts. Is this a known issue?
1 messages · Page 1 of 1 (latest)
Short answer: I can’t find any official ElevenLabs note that <break time="x.xs" /> causes the next word to be skipped. The docs confirm break support (up to 3s) and mention that excessive use can cause instability (speed-ups, noise, artifacts), but they don’t list word-dropping as a known issue. (help.elevenlabs.io)
If you’re seeing this, ElevenLabs recommends:
- Use the standard syntax <break time="x.xs" /> and keep pauses ≤3s; avoid overusing breaks, as this can introduce artifacts while they work on improvements. (help.elevenlabs.io)
- Report reproducible cases to Support so they can investigate as a potential bug. (help.elevenlabs.io)
Background references:
- Announcement of pause support in the API/editor. (elevenlabs.io)
If you’d like, share a minimal prompt/model/voice example and I’ll point you to the best place to file it.
Sources:
Do pauses and SSML phoneme tags work with the API? – ElevenLabs
Do pauses and SSML phoneme tags work with the API? – ElevenLabs
How can I submit feedback, suggestions or a bug report? – ElevenLabs
ElevenLabs—Pause support via API and speech synthesis editor
by overusing, you mean twice or thrice?