From what I understand you are asking why we can't encrypt phone calls like we can with texts. This is my two cents, please correct me if I'm wrong.
Audio transferred over phone networks is heavily processed to carry as little data as possible to reduce network traffic. It's basically highly compressed sound only carrying the frequencies necessary for human voice. This means that whatever encoding scheme you are using is going to need a lot of error correction to make it to the other side, and it will have barely any bandwidth available.
Think about it this way. You are proposing that we encode voice data into digital packets, encrypt those packets, then find a way to encode those encrypted packets back into sound that can go over the network (getting heavily corrupted by compression along the way.) Then on the other side, decode the incoming audio back into packets, apply error correction, decrypt it and decode to audio. And we need to do this in real time with a tiny bandwidth.
I'm not saying it's impossible. After all, early computer modems sent data over phone lines. But to send audio in real time would result in a severe drop in sound quality due to the restrictions on data you can send, probably to the point where speech is no longer recognizable. If you were willing to forfeit real time calls, then you could send arbitrary amounts of data over the line and have better sound quality, but that defeats the whole point of a phone call vs a text.