## Voice PE FLAC Playback Timeout - HA 2025.12 Regression

Hardware
1

Problem

Home Assistant Voice PE devices (Nabu Casa) cannot play TTS announcements after HA 2025.12 update.
Audio files are always generated as FLAC regardless of preferred_format settings, and ffmpeg_proxy
fails to deliver them to the speaker, resulting in timeout errors.

Error Messages

ESPHome Logs:
[E]: Pipeline 1: Failed to play URI: http://192.168.50.100:8123/api/esphome/ffmpeg_proxy/44e36dcaf2d8f03ab229846c98479b03/vkeoKCndXhkIBZzdm9mwIg.flac
[E]: Cannot play ‘http://…’ on pipeline 1: pipeline is busy
​

text

Environment

  • Home Assistant: 2025.12.2 (HA OS)
  • ESPHome: 2025.12.2 (Nabu Casa Voice PE)
  • TTS Service: tts.google_translate_say via tts.speak

Configuration

tts:

platform: google_translate
cache: true
service_name: google_say
language: “pl”

Service Call:
‘’’
“service”: “tts.speak”,
“data”: {
“message”: “Test”,
“media_player_entity_id”: “media_player.home_assistant_voice_097ee7_media_player”,
“options”: {
“preferred_format”: “wav”,
“preferred_sample_rate”: 44100
‘’’

Key Issues

  1. Format Ignored: preferred_format (wav/mp3/ogg) has NO effect — always FLAC
  2. Regression: Worked 2 weeks ago, broken in 2025.12
  3. Workaround: Google Home speakers work fine (no ffmpeg_proxy)
  4. ffmpeg: Available in HA OS, supports MP3/WAV

Hypothesis

Voice Pipeline in HA 2025.12 forces FLAC for ffmpeg_proxy, ignoring preferred_format.
FLAC playback times out on ESPHome speaker before audio can be delivered.

Status

Waiting for HA 2025.12.3+ patch.

2

¯\(ツ)/¯ Sounds like you are looking for…

3

I’m having a similar issue, just got a Voice PE for Christmas. I am using Piper running locally and the device fails to play anything. Well, if I power cycle the device it will respond on the first few prompts, but then stops. In the logs I see the same error as reported by the OP:

I know my pipeline is working because I can see in Debug that it processed, and I can also see responses in the Piper logs, and if I copy’n’paste that flac url into my browser it will play OK.

Any idea why the Voice PE is unable to play it?

4

Just want to add that I also have an Atom EchoS3R and that works perfectly fine with the same voice pipeline settings. This makes me think more that it’s an issue with the Voice PE device specifically.

5

I have this issue too (My voice preview stopped working as a TTS speaker (error: please call connect first))…

Please do as @Sir_Goodenough suggests and submit as an issue…
(if you don’t, I will :slight_smile: )