A guest digital device joins a playback session hosted by a host digital device and synchronizes playback of digital media content with playback on the host digital device. This enables multiple users to listen to or view media at the same time on their respective devices. A time-based synchronization process may be used to initially synchronize playback of the guest digital device to the host digital device. If synchronization is lost, the device may record and analyze audio during playback and a fingerprint-based synchronization approach may be applied that detects a time offset between fingerprints in the analyzed audio. If synchronization is still not achieved, a watermark-based synchronization approach may be applied in which watermarks are actively injected into the audio and the time offsets between the watermarks is detected.