WAM Bridge is a Windows-first Python bridge that lets you stream audio from your PC to Samsung Wireless Audio Multiroom speakers over Wi-Fi. It supports two paths: a command-line tool that serves a tokenized local stream via Samsung’s SetUrlPlayback API, and a foobar2000 output component that routes whatever foobar is playing (including internet radio) directly to your speaker.
Everything here was measured against a physical SPK-WAM550 running firmware WAM550WWB-3117.1. Other models in the Samsung WAM family are untested.
WAM Bridge is alpha software built by one person against one speaker. It works, but it is not polished. Read the latency note below before you install.
What works#
Stream local files and internet radio. SSDP speaker discovery, saved device aliases, custom radio stations, TuneIn preset browsing.
f32le PCM from foobar encoded to FLAC, served over local HTTP to the speaker. Validated on a physical M5 on 2026-08-02.
Volume, mute, pause, play, stop, standby from the CLI. Hardware volume routing available for lower-latency control.
Automatic speaker discovery, UPnP renderer, Quick Settings tile, widgets, and native TuneIn browsing with artwork.
The latency note#
The last measured end-to-end FLAC latency was 6.7 seconds (historical figure, measured before the host buffer was reduced). With buffer_extra=0, the host queue is approximately 1.9 seconds. A fresh ear-to-clock measurement has not been taken yet, so the 6.7 s figure is outdated rather than current.
Playback runs at wall-clock speed and the seekbar is honest. Pause, stop, and skip still carry PCM pipeline latency because they act through the PCM path. Setting hardware_volume=1 routes the volume slider directly to the speaker for lower-latency volume control.
Requirements#
CLI:
- Python 3.14+
- FFmpeg available in
PATH - Computer and speaker on the same LAN
- Windows Firewall access for the bridge on private networks
Foobar2000 component:
- Windows 10/11 and foobar2000 2.x x64
- Computer and speaker on the same LAN
- A saved WAM device profile
- No separate Python or FFmpeg installation after the component is installed; its helper bundle includes what it needs