FaceTime's audio quality is both higher data rate and better audio codecs. FaceTime uses AAC audio and samples at close to the full audible audio band. It also uses a relatively high data rate compared to typical voice calls.

LTE and 5G has the option of higher quality voice codecs (EVS) than older cellular networks when using VoLTE or VoNR. They support wider band sampling and higher data rates for better quality audio. Why most voice calls sound bad is inter-carrier compatibility still sucks. Typically calls will downgrade to AMR-WB or AMR-NB (8KHz and 4kHz sampling respectively).

FaceTime calls, or any other pure IP voice services, don't need to downgrade codecs to interoperate. They might have to change encoded bitrates based on network conditions but never go with some lowest common denominator codec because another user is on a different network.