Phonemes and IPA (International Phonetic Alphabet) Symbol Input for Accurate Pronunciation | Voters

Phonemes and IPA (International Phonetic Alphabet) Symbol Input for Accurate Pronunciation

in progress

Rayan

I would like to suggest a feature that allows users to input IPA (International Phonetic Alphabet) symbols to generate a specific pronunciation for certain words or names. This feature would be extremely useful for language learners, linguists, and anyone needing to generate accurate and precise pronunciations. With this feature, users could ensure that their desired pronunciation is captured, even if the word or name is difficult to pronounce or has multiple possible pronunciations. This would greatly lessen how many times you need to re-generated and improve the accuracy of the generated voices.

April 23, 2023

moshe

We need this in the multilingual model, ASAP. The whole point of IPA is to render proper pronunciation in languages other than English. Having IPA in the English-only model, with its American English phonemes, is completely useless. If we cannot specify the language for select words and paragraphs, which is what we really need, then at least give us IPA.

Marcio Puga

Would SSML work with the current Professional Voice Cloning (PVC)?

marsbars

This is great!!

Jonathan

Merged in a post:

PHONEME and SSML

audioguide

any plan to generate voice based on SSML and/or IPA phonems ?
Th issue with synthetic voice is that there are always mispronunciation !!!
When you generate a lot of text, it becomes a nightmare

September 26, 2023

Vex

marked this post as

in progress

Currently in progress for english v1

Languageproducers

Vex: This is great, what is the expected timeline to finish?

Sebastian Plasschaert

Especially the option to force a language for a word or phrase with the use of SSML.

Mike Freeman

This needs to be a high priority. This would be such an enormous game changer for ElevenLabs.

Albert

This would be great!

Jody Skinner

I was working on a manuscript with Welsh and Saxon vocabulary terms, some of which were impossible to render in English spelling tweaks. IPA would be the best solution to this problem.

Mark Rittman

I frequently encounter mispronunciations when it comes to the homograph "read," technical terminology, and people's names. To ensure correct pronunciation, I often need to make multiple adjustments to the text input, sometimes resorting to alternative spellings or substitutions. Introducing a feature that allows for phonetic text input would grant us more precise control over the pronunciation, resulting in higher-quality recordings and a smoother workflow. For many individuals, this addition is indispensable in their work.

→