Intro to VoIP Telephony (Part 2)

Welcome back! In the previous section we showed you how to register with the server.

Now it's time to make in and out calls. Here’s an example of a call flow.

An incoming or outgoing call is going to go the call server

and the call server is going to send out to the destination.

This is showing an actual call between Bob and Alice.

It is assumed that we are registered at this point.

The invite is initiating the call. And then you get a 180 Response that you're ringing.

Sometimes you'll see a 183 Session Progress.

And the difference is that the 183 also indicates that the server might be playing a ringtone

or sending us some kind of other media before the call is connected .

Then we get to the 200 OK and the acknowledgement.

After the acknowledgment, you get RTP communications both ways.

Then, depending on who ends the call, you will see a BYE message.

One big part of troubleshooting SIP

is when calls are not being completed we can usually tell who is disconnecting first.

If it turns out that the Q-SYS Core is closing the call for no reason, then Q-SYS is to blame.

If the SIP server is closing the call, well then SIP is to blame.

The Server almost always closes the call but sometimes we do see issues with the Core closing the call,

as a result of a mismatch setup issue.

This is showing the same call but with a network capture.

You can see the protocol here. You can see the start of the call flow with the INVITE, TRYING, and 200 OK.

That means that both parties can now send audio.

You can also see the protocol changes here in the capture and the codec being used, G.711 PCMU.

At this point, the BYE SIP packet takes over.

You can see who sent the BYE, and then then the 200 OK, and then the RTP traffic should stop and the call should be disconnected.

This shows another view of the call setup in Wireshark using the call flow view.

One thing to note is that you can capture this using Wireshark on a PC or laptop if you are connected to the same network

or if you have a capture utility on the Q-SYS Core that will capture the traffic.

These captures are referred to as PCAPs.

This will also capture the audio stream which you can play back if you are having audio quality issues.

This shows a call captured on an external device showing two different proxies or call servers involved.

A typical SIP call will go usually through one or more devices before it reaches its destination.

Next we have session description protocol, which is what it sounds like. It describes the session.

In there, you’ll find a session announcement invitations.

This is also where we negotiate various parameters of the call.

The softphone and the call server need to determine what capabilities (like audio codecs)

that they can support between the two of them.

Within the SDP you'll see the connection IP address,

the port's audio Codec, sample rates, and other media attributes.

Without them, you really can't set up the call correctly, so needless to say, the SDP is pretty important to the call.

Let’s look at this slide again to examine an early offer and a late offer (by the way Q-Sys supports both methods.)

Early offer includes SDP in the initial call setup. Each device is sharing their supported pieces, and chooses which one each supports.

Let’s talk about audio codecs for example. Let’s say Q-SYS supports six of them.

They'll pick one and tell us which one they've picked. This is called the early offer.

The late offer doesn’t send SDP.

It simply send a SIP message without session description protocol and then tells Q-SYS which codecs they support and then we pick from them.

Let’s talk media types.

M stands for the media name and the transport address. C is the connection. And A is the attribute.

And this is actually what it looks like if you open the network packet.

This is where we're trying to connect to this IP address, and the M is the media.

You'll also see the media attribute, G.722, and the sample rate. You can even see that BroadWorks is the SIP proxy.

Once that's all set up we're ready to send the audio with RTP.

As you recall, SIP is the signal element but RTP is the actual voice, the transfer of the media.

RTP doesn't just include audio but it can also be video and other realtime data.

If you do a network capture, a PCAP within the Q-SYS Core or a network capture on a PC, you can capture the audio and play it back on your PC.

Sometimes customers are going to complain about choppy audio or missing audio,

so you can use this capture to diagnosis the issue.

You can also encrypt the RTP traffic by turning on the SRTP.

Some providers require SRTP and TLS, which is used to encrypt SIP messages,

although this is not always the case.

Some providers will allow one or both.

Let’s talk audio codecs, which compress and decompress the signal

in an effort to reduce the bandwidth and transmit the digital voice data.

We’ll start with the high fidelity codecs and work our way down.

Q-SYS support G.722 which is not to be confused with the .1 and the .2 variants. Those are slightly different.

G.722 is a wide band codec for higher quality sound. It is a newer codec compared to G.711.

The bidirectional communication used is between 96 and 128 kilobytes per second.

G.711 is still a very popular codec that been around since 1972.

This was used on the Public Switch Telephone Network.

One great advantage is that it has low computational requirements to compress and decompress.

G.711 has two variants. There's a U law and an A law, which is primarily used outside the United States.

Then there's G.726, which requires about half the bandwidth of G.711 with around the same quality.

Q-SYS supports a 32 bit version of this.

G.729 is a very low bandwidth codec.

Previously this was a “paid-for codec” but recently the patent ran out and now it's royalty-free.

Remember, this is a very low bandwidth but you still get decent quality.

The thing to remember about low quality codecs is that they may not reliably send over DTMF.

The lower quality may corrupt the DTMF and it won’t reliably go through.

And we’ll get back to DTMF in the next video. For now, let's take a break and we’ll see you when you get back!