I actually did (was going to travel internationally), and it wasn't as useful as you'd think. I would be talking to someone, and in the background someone else would be talking, and it would translate both people.
Only worked in a 1:1 in a quiet place. Still, can't complain for free.
As sister comments have pointed out, I find that it's very much dependent on the phone you're using. If it can't even show me a waveform then I don't think it's going to be able to tease out any sort of speech
Sorry I wasn't clear. It's about the hardware microphone. They're usually optimized to filter out far field sounds and only focus on near field sounds. So if you're trying to understand a conversation from even 10 ft away, it can be a problem. Not saying that's the situation that you encountered, but it does make sense that a $50 phone versus a $500 phone might have different capabilities when it comes to the microphone
It's not a model problem, but it is a SW problem. It should be able to distinguish nearby people from people farther away and give me options to set a threshold on who to include.
Not the same, but related: Gemini is great for querying text in another language, it provides really cogent, useful responses with just enough source language quotes to be able to reference the source text effectively.