Compatibility
Every connection is saved as a profile — a card you load onto a codec from a menu. The profile carries its transport, its target and its audio codec, so loading it is the whole job. What you pick depends on what is at the other end, and AirCodec covers the three cases a station actually meets.
The industry standard for dialling. Any Comrex, Tieline, AEQ, Intraplex or SIP phone can call AirCodec, and AirCodec can call it — by name, not by address.
What SIP is, and why it matters 02 — DirectAirCodec to AirCodec, point to point. Error corrected, retransmitted, password protected and encrypted — with your contact closures riding alongside the audio.
What Direct does 03 — Streams and the restPull an Icecast or Shoutcast stream in as a codec's audio, push to a Barix, point plain RTP at a fixed address, or loop one codec to another in the same box.
Streams, Barix and RTP01 — SIP
SIP is dialling. It is the system the entire telephone world moved to, and it does one job: it lets one device find another and agree to talk, without either of them knowing in advance where the other is.
It works the way a phone network works. Every device registers with a server and says here I am, and here is what I am called. When you dial a name, the server looks up where that device actually is right now and introduces the two ends to each other. Then it gets out of the way — the audio runs directly between the two codecs, not through the server.
This is why a remote can be on a hotel network one week and a phone hotspot the next, and still be reached by the same name.
SIP on its own is a telephone standard. In the 2000s the European Broadcasting Union sat down with the codec manufacturers — Tieline, Mayah, AETA, Orban among them — and wrote down how broadcast codecs should use it, so that different brands could call each other.
That document is EBU Tech 3326, published in 2007, also known as N/ACIP. It is the reason a Comrex can call a Tieline at all.
When the call is set up, each end offers a list of the audio codecs it can handle and they settle on one both understand. 3326 makes a handful mandatory so there is always common ground:
The honest caveat. Not every manufacturer implements all of 3326 — some leave out a mandatory algorithm, and two compliant boxes can still fail to agree if their codec lists do not overlap. That is why AirCodec carries the whole menu rather than a favourite few, and why the connection screen shows you what was actually negotiated instead of making you guess.
A codec that can only be called is half a codec. AirCodec dials out, takes calls in, and can answer them on its own — and in every case the thing you type is a name, not an address.
Load a profile and press Connect. AirCodec places the call the way a hardware codec does, so the far end does not need to know what it is talking to — a Comrex at the stadium answers an AirCodec exactly as it would answer another Comrex.
Each codec holds its own name on the account, so a remote dials the channel it wants rather than a shared number somebody then has to patch. Five codecs, five names — an incoming call lands where it was aimed.
Turn it on for the codec that has to pick up and an unattended site takes the call with nobody there. Optionally ring first for a set time before answering, so a studio can grab it by hand if somebody is in.
Why the name matters. An address changes when a remote moves hotel, when a site gets a new circuit, when a carrier rotates it overnight. A name does not. Both ends register with the server, the server keeps track of where each one actually is, and the call finds it.
It speaks the same standard as
Any EBU 3326 / SIP device can call AirCodec, and AirCodec can call it. That includes:
On a RØDECaster, SIP calling comes with the CallMe Pro tier — the free tier connects RØDECaster to RØDECaster and web guests only.
Broadcast codecs
SIP phones and softphones with G.722
Putting a hardware codec on a SIP network usually means somebody picks a username, picks a realm, picks a password, and gets one of them subtly wrong. AirCodec issues them instead. Name the unit, press Register device, and the server hands back the exact server, port, transport, username, password and realm — ready to type into the box once.
Settings → Devices. Type what the thing is called — press-box-comrex, remote-tieline — and press Register device.
They come straight back on screen: server, port, transport, username, password, realm. Nothing to invent and nothing to agree with anybody — they are generated against your account.
Enter them in the Comrex or Tieline's own SIP settings and it registers. That is the only time anybody types a credential — AirCodec's own five codecs are provisioned and never need it at all.
The unit joins the endpoint list for everyone on the account and is dialled by name from a profile. Move it to another building and the name still finds it.
Open full size
Standards, not a walled garden. SIP is how the hardware already works, so nothing at the far end needs an AirCodec licence, a plug-in or a firmware level. If a unit can make an EBU 3326 call today, it can call AirCodec today.
02 — Direct
SIP buys you interoperability. Direct buys you the studio-to-transmitter link on a bad night. It is AirCodec to AirCodec, point to point over RIST, with no registrar in the middle and nothing to negotiate — because both ends are the same software and already agree.
That is what lets it do things a standards call cannot: correct errors in both directions independently, ask again for a packet that went missing, carry your contact closures in step with the audio, refuse a caller who cannot produce a password, and encrypt the whole thing.
One link carries the programme, the automation and the station data — behind an address list and a password, on ports derived per codec rather than a page of firewall rules.
A standards call has one answer to packet loss: live with it. Direct has two, and uses whichever fits the moment.
An incoming Direct link either carries the connection password or it does not. Without it there is no session and no audio, and the refusal is logged with the address it came from.
Before the password is reached, the caller has to be an address you allow. Accept Direct callers from takes as many entries as you need — and anything not on the list never gets as far as a password prompt.
Per codec, you choose which kinds of call it will take: incoming Direct, incoming SIP, either, or neither. Anything not ticked is refused outright — a codec that only ever dials out need not listen for anything.
Set that password and the link is encrypted with AES-256 — the programme, the contact closures and the RDS data together, as one protected stream across the public internet.
A Direct link allows a separate receive-side codec. Run a heavy codec down to the transmitter and something lean on the way back, or more correction on the leg that needs it and none on the leg that does not.
Open full size
Use both. Nothing makes you choose. A connection list can hold a Direct link as the first slot and a SIP call as the backup, so the link that has to hold is the one that normally carries you — and when it goes, the standards call picks it up.
03 — Streams and the rest
Sometimes the thing you need is already running and has no interest in being dialled: a programme stream on a server, a Barix box screwed to a wall, a legacy unit that only knows one address. Each of those is a transport you pick from the same menu, and once it is loaded the codec treats it like any other connection.
Point a codec at an HTTP, Icecast or Shoutcast stream and it arrives as that codec's audio — MP3 and the rest. No call, no far end to configure, nothing to answer.
Send the line input to a Barix box as a plain RTP push, or over Barix's own subscribe-based BRTP; receive from one; or run both directions at once.
Plain RTP where the far end is fixed and does not dial, with a Comrex preset that fills in what a Comrex expects so you are not guessing payload types and ports.
A loopback is a connection like any other: two instances on the same machine linked over Direct. For proving a path before the show, and for routing audio around the building.
Mix them in one list. The connection list takes three slots in order, and they do not have to be the same kind. A Direct link first, a SIP call second, a stream third — so when the good path goes, the codec works its way down whatever you have, and comes back up the moment the first one proves stable again.
A remote is usually on somebody else's network. AirCodec handles that with the same tricks the wall-mounted units use, so a studio engineer is not on the phone to a hotel's IT department before a game.
Direct and RTP ports are derived per codec · no five-page port worksheet
AirCodec
A Comrex at the stadium, a Barix at the transmitter, a laptop in a press box, or all three on the same account. Say what you are trying to link and we will tell you straight whether AirCodec does it.
More on AirCodec
