This project provides a media player (custom component) for Home Assistant that plays TTS (text-to-speech) via a Bluetooth speaker.
If you're using HA's Bluetooth device tracker (for presence detection), this project also provides a replacement Bluetooth tracker that allows both components to play nicely together.
Since the Bluetooth tracker constantly scans for devices, playback of audio on the Bluetooth speaker may be disrupted / become choppy while scanning. These custom components work together to ensure only one of them is accessing Bluetooth at any given time.
The flow is something like this:
sudo apt-get install pulseaudio pulseaudio-module-bluetooth bluez mplayer sox libsox-fmt-mp3
sudo adduser pi pulse-access
sudo adduser homeassistant pulse-access
In /etc/pulse/system.pa
, add the following to the bottom of the file:
### Bluetooth Support
.ifexists module-bluetooth-discover.so
load-module module-bluetooth-discover
.endif
#set-card-profile bluez_card.00_2F_AD_12_0D_42 a2dp_sink
The last part is to persist the setting for a2dp, in case your bluetooth seems to default to a different profile. I have commented it out because it seems to be flakey.
You may want to uncomment this line if your audio is getting cut off:
### Automatically suspend sinks/sources that become idle for too long
#load-module module-suspend-on-idle
Create the file /etc/systemd/system/pulseaudio.service
and add the following to it:
[Unit]
Description=Pulse Audio
[Service]
Type=simple
Environment=DBUS_SESSION_BUS_ADDRESS=unix:path=/run/dbus/system_bus_socket
ExecStart=/usr/bin/pulseaudio --system --disallow-exit --disable-shm --exit-idle-time=-1
[Install]
WantedBy=multi-user.target
Enable the service to start at boot time.
sudo systemctl daemon-reload
sudo systemctl enable pulseaudio.service
Give pulse user access to bluetooth interfaces
edit /etc/dbus-1/system.d/bluetooth.conf
add the following lines:
<policy user="pulse">
<allow send_destination="org.bluez"/>
<allow send_interface="org.bluez.MediaEndpoint1"/>
</policy>
sudo bluetoothctl
scan on
pair 00:2F:AD:12:0D:42
trust 00:2F:AD:12:0D:42
connect 00:2F:AD:12:0D:42
quit
Create the file [PATH_TO_YOUR_HOME_ASSSISTANT]/scripts/pair_bluetooth.sh
and add the following to it. Make sure to replace the Bluetooth address with that of your Bluetooth speaker.
#!/bin/bash
bluetoothctl << EOF
connect 00:2F:AD:12:0D:42
EOF
Make sure to grant execute permissions for the script.
sudo chmod a+x [PATH_TO_YOUR_HOME_ASSSISTANT]/scripts/pair_bluetooth.sh
In /etc/rc.local
, add the following to the end of the file to run the script at startup:
# Pair Bluetooth devices
[PATH_TO_YOUR_HOME_ASSSISTANT]/scripts/pair_bluetooth.sh
exit 0
Copy the TTS Bluetooth Speaker component (from this GitHub repo) and save it to your Home Assistant config directory.
custom_components/tts_bluetooth_speaker/media_player.py
This step only applies if you're using the Bluetooth tracker.
Copy the Bluetooth Tracker component and save it to your Home Assistant config directory.
custom_components/bluetooth_tracker/device_tracker.py
pactl list sinks
You should see something like:
Sink #1
State: SUSPENDED
Name: bluez_sink.00_2F_AD_12_0D_42.a2dp_sink
If it instead says headset_head_unit, you can switch to a2dp profile as follows:
pactl set-card-profile bluez_card.00_2F_AD_12_0D_42 a2dp_sink
Check again and validate it is using a2dp.
Test using command line if mplayer can stream to a2dp
mplayer -ao pulse::bluez_sink.00_2F_AD_12_0D_42.a2dp_sink -channels 2 -volume 100 /some/mp3file.mp3
By this stage (after a reboot), you should be able to start using the TTS Bluetooth speaker in HA.
Below is an example of how the component is configured. You need to specify the Bluetooth address of your speaker, and optionally set the volume
level (must be between 0 and 1). If you find your speaker is not playing the first part of the audio (i.e. first second is missing when played back), then you can optionally add some silence before and/or after the original TTS audio hsing the pre_silence_duration
and post_silence_duration
options (must be between 0 and 60 seconds). If you've change your TTS cache directory (in your TTS config), then you should set the cache_dir
here to match.
media_player:
- platform: tts_bluetooth_speaker
address: [BLUETOOTH_ADDRESS] # Required - for example, 00:2F:AD:12:0D:42
volume: 0.45 # Optional - default is 0.5
# pre_silence_duration: 1 # Optional - No. of seconds silence before the TTS (default is 0)
# post_silence_duration: 0.5 # Optional - No. of seconds silence after the TTS (default is 0)
# cache_dir: /tmp/tts # Optional - make sure it matches the same setting in TTS config
If you're using the Bluetooth tracker, you probably already have this in your config:
device_tracker:
- platform: bluetooth_tracker
To test that it's all working, you can use Developer Tools > Services in the HA frontend to play a TTS message through your Bluetooth speaker:
{ "entity_id": "media_player.tts_bluetooth_speaker", "message": "Hello" }
Another way to test it is to add an automation that plays a TTS message whenever HA is started:
automation:
- alias: Home Assistant Start
trigger:
platform: homeassistant
event: start
action:
- delay: '00:00:10'
- service: tts.google_translate_say
data:
entity_id: media_player.tts_bluetooth_speaker
message: 'Home Assistant has started'