Local loopback

Before you try contact other people or test over the network by contacting an echo server, it's a good idea to make sure your machine's audio hardware is set up properly. An easy way to verify this is Speak Freely's local loopback facility, which allows you to open a connection to your own machine that does not go over the network. Sound you transmit is stored in memory, then replayed shortly after you end the transmission. You can evaluate different compression modes and other options, and set your audio input and output levels optimally for the modes you're using.

To establish a local loopback connection, use the Help/Local Loopback menu item. Once the connection window appears, try transmitting short sequences of sound (like "Testing: one, two, three, four."). If all is well with your audio hardware, about a second after the end of each transmission you'll hear your voice replayed. If you don't hear anything, make sure your speaker is plugged into the right jack on the sound card and that the speaker volume is turned up. If you hear only a quiet hiss or hum, see if your microphone is plugged into the correct jack (a common error is to plug the microphone into a "Line in" jack designed for higher-level signals than the microphone generates). If you've checked these things and still can't hear anything in local loopback, try the suggestions in the list of frequently asked questions on setting up your audio hardware.

Since local loopback stores audio in memory rather than sending it over the network, the length of transmission it can store is limited. For short test messages, this isn't usually a problem unless your machine has extremely little free memory. If you're using voice activation, note that local loopback does not begin replay when silence is detected, but only when you end the transmission. The reason for this is that even though nothing is being sent, voice activation must still "listen" in order to resume transmission as soon as you resume speaking. With half-duplex sound hardware, this would prevent the looped-back sound from being played. Even with full-duplex hardware, the combination of voice activation and loopback would lead to an endless series of echoes if sound from the speaker triggered the microphone.