Problem with tts and mp3

Hello All,

I want to create a spoken manual about what you can do with a wallpanel in a room.
There is an option to choose music depending on the mood.Like classic etc.
There is a button that people can push and then via TTS i tell the possibilities.
What i want is that TSS say: you can choose the music fits to your mood for example classic: Play classic mp3 and then tts again: or rock and play rock.mp3

The first TTS goes well but it never plays the mp3.
what do i do wrong, can someone point me in right direction

alias: New script mp3
sequence:
  - service: tts.cloud_say
    data:
      entity_id: media_player.woonkamer_speaker
      message: >-
        hello this is a test play classic music
      language: en-EN
  - service: media_player.turn_on
    target:
      entity_id: media_player.woonkamer_speaker
  - service: media_player.play_media
    data:
      media_content_id: /local/klassiek.mp3
      media_content_type: audio/mp3
    target:
      entity_id: media_player.woonkamer_speaker
mode: single

I got locla MP3 files playing via automation using the following:

- id: '1615833464125'
  alias: Test Audio File
  description: ''
  trigger: []
  condition: []
  action:
  - service: media_player.play_media
    data:
      media_content_type: music
      media_content_id: /share/MagicChime.mp3
    target:
      entity_id: media_player.local_md_12
  mode: single

I do also have in my configuration.yaml
media_source:
I am not sure if this is required but playing local MP3’s started working when I had the above automation and this added.

@mandelbug: If is use them separately it works so mp3 plays

nobody any idea?

Is the MP3 playing before the speaker is on? i.e. can you put a “sleep” in between turning the speaker on and playing the MP3