could perhaps be really simple:
item:
audio: (file, base64, url?)
level: (full text, paragraph, sentence, phrase, word, syllable, phoneme, vad, etc..)
info: string (what kind of markup?)
source: string
comment: string
text: string
start: int ms
end: int ms
items: list
all attributes optional? defaults for what to do about annotation /validation for each level
could perhaps be really simple:
item:
audio: (file, base64, url?)
level: (full text, paragraph, sentence, phrase, word, syllable, phoneme, vad, etc..)
info: string (what kind of markup?)
source: string
comment: string
text: string
start: int ms
end: int ms
items: list
all attributes optional? defaults for what to do about annotation /validation for each level