Comment by numlock86 | Hacker News Reader

Comment by numlock86 | Hacker News Reader

Parent Full thread

I tried to do something similar not too long ago. Different application really, but I wanted to do something on the other end depending on how often it rang. Sadly I discovered that apparently in all-digital "modern" phone networks (SIP, VoIP etc.) the ringing tone you hear as a caller is simply ingested by your network operator and the tone you hear as the one getting called is basically just some audio file that plays over and over, triggered by a single "hey, someone is calling now!"-message sent by a modem unit, which to my surprise are still using AT commands to this day. Bummer. Turns out even the duration of how long it rings depends on too many factors, so what you hear on one end (ringing) might be off by multiple seconds on the other end.

At least it explains the "You only heard it ring once? I let it ring like five times before I hung up! Seriously!"-anecdotes I keep picking up. :)

> the ringing tone you hear as a caller is simply ingested by your network operator

In my experience this depends on the network in question: SIP supports "early media" and I've seen it in action a few times. You can tell by the fact that phone UIs usually don't count connection time in that state, yet you can hear a prerecorded announcement, music mixed with the dialing tone etc.

I remember some creative uses of this as well: If you controlled the signalling stack, it was possible to keep a connection open in this state indefinitely without either end getting charged.

Here's a relevant RFC: https://tools.ietf.org/html/rfc3960

The early media can still be generated at any point of the calling chain. It doesn't have to correspond to actual ringing, if ringing does happen on the POTS side.

(For a test, you can call someone in the UK and check if you hear Pink Floyd style beeps)

This is also true with silences. When we switched from analogue phone calls to digital phone calls in the 90's, SIP began to ingest white noise during silences due to discomfort perceived by the users thinking they got hung up or the network has been disconnected.

It's even worse on Zoom.

A simple experiment: let your counter party put a metronome next to their microphone, and enjoy the time distortion.

It is more fun in Skype calls. There is a 'gap' then you get a 10second audio clip 'played to you' in 5 seconds :)

> simply ingested by your network operator

That's an amusing typo - I suppose that should be 'injected', but Pac-Man pops to mind ...

Reply on news.ycombinator.com