I am looking to setup realtime audio transcription to txt via Rayban Meta Smart Glasses to build training data. I don't think it would be able to identify voices .. yet, but even just straight text NLM's are able to disseminate between different people in a conversation.
Lets assume there is enough room for storage. That's always going to be.. is an issue.
Well, that's not a feature they have. You'll need to look for some extra apps that play nice with the glasses. Essentially, you need a way to get the audio from the glasses to an app that can listen to this audio and jot down what it hears, turning sounds into text as it goes. You might want to check out services that can do this magic trick of listening and writing in real time. Google's Cloud Speech-to-Text is a great place to start. Just make sure whichever service you pick can keep up with the conversation without needing breaks.
