Audio

Every pronunciation row carries playable sound. Where it comes from and how to use it.

The sounds array

A pronunciation, and for Japanese each inflection's pronunciation, carries sounds: one object per voice with a playable mp3_url, the voice's name and gender, and the alphabet the voice was driven from, IPA for English and phonetic for Japanese. English entries carry two voices, Emma and Michael. Japanese entries carry Himari today.

{
  "id": "a4ad93cf-753a-49cd-b85c-71b6876a9497",
  "key": "sounds/01M45PR236JEQ21RS26407FZ32.mp3",
  "voice": "Himari",
  "mp3_url": "https://storage.immit.co/sounds/01M45PR236JEQ21RS26407FZ32.mp3",
  "gender": "female",
  "alphabet": "phonetic"
}

Playing it

The URL is a public object on storage.immit.co and plays as is in an audio element or a native player. Play it directly or cache it on your side, at any scale; a sound fetch is not a request to /v1 and never counts against the quota.

Coverage

Not every row has sound yet; an empty sounds is normal on a rare inflection. A client shows a speaker where the array is non-empty and nothing where it is empty.