Depends on which exact voice you are using. Most VoIP uses standard 729 or 711 codec to carry the voice. It is not encrypted so all you have to do is somehow capture it and you can play it back. Wireshark can playback 711 based calls as a standard feature.
Capturing the data is a different story likely the only way would be over a open wireless connection. Anything else would require some form of physical access to the network.