Amazon Transcribe

Testing Amazon's automatic speech recognition (ASR) service with my son

Published: Tuesday, Dec 1, 2020 Last modified: Wednesday, Apr 14, 2021

Trying Amazon Transcribe two speaker option:

Transcript

Speaker 0: Charlie.
Speaker 0: Charlie.
Speaker 0: What? Can you just pause that? Why? Because I want to record a conversation with you.
Speaker 0: I wanna ask you a couple of questions, Charlie. How old are you? Fine.
Speaker 0: Where do you live?
Speaker 0: State of par. What color is your hair?
Speaker 0: I don't know. Brown. Okay.
Speaker 0: And who's your mother?
Speaker 0: Mommy. What's her name?
Speaker 0: Blue.  
Speaker 0: And your daddy?
Speaker 0: Carl.
Speaker 0: That's great.
Speaker 0: What is Kaka's name?
Speaker 0: Christie.
Speaker 0: What is your sister's name? What's Mei Mei's name? David.
Speaker 0: Perfect. How old is ever
Speaker 0: to?
Speaker 0: And
Speaker 0: Okay, that's it. Can you say goodbye? Please, Charlie. He does not.
Speaker 0: Goodbye.

Bizaarely, IIUC, it’s just one speaker according to AWS?!

https://s.natalian.org/2020-12-01/conversation-charlie.json