I frequently have a hard time hearing certain things in the sim that aren’t a problem with real controllers. I think part of it is the way ATC transitions tempo. Especially from numbers to words. There is a pause with the numbers but the tail end of a number the word will run together with the last number. In my mind it’s changing the callout that I’m mentally processing.
As an example, I’ll hear 1 … 3 … for at 7 instead of 1 … 3 … 4 … at 7.
I could be imaging this but it doesn’t happen when I listen to actual radio communications. Even when spoken fast, a real person or speech synthesized METAR usually clearly defines the words and phrases.
It’s like Azure Voice can do natural language but gets tripped up a little on aviation speak.