Downsides to 48k sampling rate in Voice?

edited October 2019 in Photon Voice
Hi! Are there any downsides to using 48k sampling rate in Photon Voice?

Does it affect CPU of sending client? What about CPU of receiving client?
Does it cause bottlenecks if the room gets fuller? Say 8 people.
Most importantly, I reckon, does it cause issues of lagging connections and dropped audio on slower mobile connections?

I will be implenting a sampling rate option for testing, but would appreciate any pointers. Thanks!