If sound arrives while you're sending live audio and your audio hardware is half-duplex, Speak Freely, by default, simply discards the incoming sound and increments the number of lost input packets shown in the Extended Status dialogue. If you check the Options/Break Input menu item, sound that arrives while you're sending will, instead, interrupt your transmission, letting you know that the other person wants to say something.
Network traffic congestion and the fact that packets can travel on a variety of routes between two sites can lead to random pauses (jitter) in the sound you receive. To reduce the severity of the pauses, Speak Freely usually delays playback of the first in a sequence of sound packets to provide some margin for subsequent packets to arrive, even if slightly delayed. This improves the quality of the sound, but at the cost of introducing an additional delay before you start to hear a transmission from another user. The Options/Jitter Compensation menu allows you to select a variety of anti-jitter delays ranging from none at all to three seconds. If you're communicating across a local network, "None" is the best setting. The default, 1 second, generally gives much better results across Internet connections than no delay. If you have severe delay problems, you might want to try a higher setting. Lower jitter compensation times are usable when communicating between sites with high-bandwidth connectivity to the Internet.
If sound arrives simultaneously from more than one host, the packets are interleaved. This makes it difficult to understand, but it does permit interrupting a long winded speaker in a conference call.