FreeSWITCH Cookbook: Written by members of the FreeSWITCH team, this is the ultimate guide to getting the most out of the platform. Stuffed with over 40 recipes, just about every angle is covered, from call routing to enabling text-to-speech conversion.

As its name implies, FreeSWITCH will "switch" or "connect" various endpoints together. Part of that switching involves making semi-permanent connections to individual telephones or telephone service providers. Service providers are usually telephone companies (telcos) or ITSPs (Internet Telephony Service Providers). Read on to learn about the many ways that FreeSWITCH can connect your telephone to the world.

The recipes in this chapter will delve into the various ways to connect FreeSWITCH to telephones and service providers. FreeSWITCH can also utilize a locally installed sound card by means of the PortAudio library. The last recipe is for advanced users and discusses the subject of codec negotiation.

SIP phones or any SIP device with the ability to register, are essential in most FreeSWITCH installations for allowing users to communicate with each other. A registration is when a phone or other device informs FreeSWITCH that it is active and provides information (such as an IP address and port) on how to reach the phone across the network or Internet. FreeSWITCH stores this information for use later if someone wishes to contact the phone.

In this recipe, you will be registering a phone to FreeSWITCH. You will need to enter your credentials into your phone as well as into FreeSWITCH itself (both sides must match). We will cover only the FreeSWITCH server portion of registration in this book.

Many of us have a USB headset or a sound card in our laptops or desktop computers. In most cases, FreeSWITCH can utilize these audio devices. Usually this is done for basic troubleshooting or to feed an external audio source into FreeSWITCH. It is also handy when you wish to use FreeSWITCH as a softphone, which is demonstrated later in this chapter.

The previous recipe described the process of setting up FreeSWITCH to use a local audio device such as a USB headset or a sound card. This recipe builds on the previous one by showing you how to use a custom FreeSWITCH configuration that is tailored specifically for use as a softphone (be sure to have PortAudio working before trying this recipe). Lastly, we will direct your attention to a few open source softphone projects that utilize FreeSWITCH as the VoIP engine.

Configuring a SIP gateway allows you to connect with outside carriers or other SIP machines. You can connect with other FreeSWITCH or Asterisk boxes, or to upstream carrier SIP trunks.

SIP gateways have many, many options—too many to list here, so we'll review just a few.

Google Voice (also known as Google Talk) is accessed via the mod_dingaling module, which provides XMPP support. (If you would like to learn more about XMPP, visit http://xmpp.org.) Mod_dingaling can act as both a XMPP server and a XMPP client, though it is generally used for its client abilities. With Google Voice, XMPP simply serves to establish sessions (much like SIP) and the audio properties and stream are otherwise merely RTP, just as in any VoIP call, using codecs you already know, like PCMU, GSM, and so on.

When using Google Voice with FreeSWITCH, your FreeSWITCH system initiates connections to or from Google Voice and can convert the audio and signaling into any other form—including SIP. This allows calls from Google Voice to your SIP phone and vice versa.

Codec configuration is very versatile in FreeSWITCH. In IP telephony, there are several differing scenarios for negotiating and choosing codecs. To meet the varying demands, FreeSWITCH has several configurable modes of operation as well as real-time variables that can influence how codec negotiation takes place. Typically the goal should be to reduce transcoding or resampling as much as possible. Transcoding is the case where two sides of the call have different codecs and audio flowing in either direction has to be completely decoded and re-encoded to the opposite channel's format. Resampling is similar but it is required when each side of the call is running at a different sample rate and the audio has to be "resampled" to the correct rate. One or both of these can be necessary depending on where you direct your calls to and how you have your codec configuration set.