From what I know, neuro "listens" by turning what people say into text first. How does she know who said what? Can she actually hear the difference between their voices or does she just detect other people's discord output or something similar to that?
#How does neuro differentiate between different people?
next girder
pale estuary
From what I’ve heard it’s the latter option
uneven cedar
As far as I know, she just knows which discord user the audio is coming from
smoky cradle
Most likely. She can distinguish not just the person, but the source as well. So the messages she is fed are probably prefixed with something like "vedal987@discord_dm", " evil@voice_call", "random_person@twitch_chat" etc.
burnt mesa
I randomly decided to check discussions and its something I can answer. With discord, you can get seperate audio data with associated discord ID, so you can process each person's audio data separately. But there is also AI capable of separating multiple voices assuming they sound different enough. But then you have to manually assign which voice is which for the AI output.
I have code for the former solution. https://github.com/pladisdev/Discord-AI-With-STT
pastel jewel
like everyone else said, she can just tell what discord account is speaking, which is why she will often refer to people by their discord name, e.g. calling ellie 'dumb lil robot'