It appears that the media_player component might not be working properly as the state of mine do not (always) update their state when playing.
I have been having some really irritating issues with TTS for some time. Often the message is cut significantly short and recently it usually doesn’t start until up to about a second into the message which at best is mildly annoying but at worst makes the message difficult to understand if it is short.
I think I have tracked down the reason that sometimes the message is being cut short. If I watch the state of the media_player it often doesn’t change from ‘paused’ to 'playing'.
Has anyone else noticed this?
Does anyone have any ideas why this is happening?
Does anyone know internally what causes a state change? I’d like to explore if it could be anything at my end, although I can’t think of anything. (I’m using Sonos. And hassio if that is relevant).
Here is my config which is very standard for TTS.
announce:
sequence:
# Take a snapshot
- service: media_player.sonos_snapshot
data_template:
entity_id: "{{ 'media_player.' ~ room }}"
# Unjoin
- service: media_player.sonos_unjoin
data_template:
entity_id: "{{ 'media_player.' ~ room }}"
# Set volume
- service: media_player.volume_set
data_template:
entity_id: "{{ 'media_player.' ~ room }}"
volume_level: "{{ volume }}"
# Say the message
- service: tts.google_say
data_template:
entity_id: "{{ 'media_player.' ~ room }}"
message: "{{ message | replace('\n',' ') | replace(' ',' ') | replace(' ',' ') }}"
# Wait until Sonos starts playing
- wait_template: >-
{{ is_state('media_player.' ~ room , 'playing') }}
timeout: '00:00:30'
# Now wait until it has finished playing
- wait_template: >-
{{ is_state('media_player.' ~ room , 'paused') }}
timeout: '00:02:00'
# Restore from snapshot
- service_template: >
{% if last_message | default(True) %}
media_player.sonos_restore
{% else %}
script.no_op
{% endif %}
data_template:
entity_id: "{{ 'media_player.' ~ room }}"