So I just bought the home assistant voice preview edition, it works great (as in I can wake it up with the wake word and control things) however it doesn't talk? I have piper set up but it doesn't talk back like say "How can I assist", or "turned off the light" etc.
#HAVPE works but no voice back
tribal holly
atomic ivy
That's a normal feature, when you do an obvious command (turning on a light, etc) it will make a noise and do it.
tribal holly
How can I get it to speak back though?
atomic ivy
I don't know, honestly. I use an llm and it always speaks back, even if I don't want it to, so maybe try that.
tribal holly
Do you happen to know if speech to phrase ever speaks back? I'm trying to avoid using llms / ai especially non local ones
lone fossil
Ask it what is the weather
tribal holly
Do I first need to expose my weather to it?
lone fossil
No. It should say "sorry, blah-blah..."
THen you will know that Piper works
Because it moght be broken.
tribal holly
Yeah it doesn't respond even "sorry I can't find that command" or w/e
The only thing in pipers log is this also not helpful
DEBUG:wyoming_piper.handler:Sent info
lone fossil
Check the pipeline, and also restart the Piper. If you have Cloud subscription, try that as TTS. To rule out the VPE
tribal holly
I restarted piper and the log seems to indicate its trying to respond but I don't hear it
DEBUG:wyoming_piper.handler:Sent info
DEBUG:wyoming_piper.handler:Sent info
DEBUG:wyoming_piper.handler:Text stream started: voice=SynthesizeVoice(name='en_US-hfc_male-medium', language=None, speaker=None)
DEBUG:wyoming_piper.handler:Synthesize(text='0.0 °C and clear', voice=SynthesizeVoice(name='en_US-hfc_male-medium', language=None, speaker=None), context=None)
DEBUG:wyoming_piper.handler:synthesize: raw_text=0.0 °C and clear, text='0.0 °C and clear.'
DEBUG:wyoming_piper.handler:Loading voice: None
DEBUG:wyoming_piper.download:Checking /data/en_US-hfc_male-medium.onnx
DEBUG:wyoming_piper.download:Checking /data/en_US-hfc_male-medium.onnx.json
DEBUG:wyoming_piper.handler:Text stream stopped
DEBUG:wyoming_piper.handler:Sent info
I don't have cloud sub either btw
When setting up the assistant selecting the piper voice and it says "try voice" that works fine too
I've just realized it doesn't speak even on the companion app
In the pipe line debug thing I just noticed this not sure if it helps but...
tts_output:
token: noqiNPY-HAUJJHSsMHRLUw.mp3
url: /api/tts_proxy/noqiNPY-HAUJJHSsMHRLUw.mp3
mime_type: audio/mpeg
stream_response: false
Why is stream_response: false false
lone fossil
Stream means it will, well, stream sound instead of waiting for full generation and providing full file.
AFAIK Piper doesn't support streaming.
There probably is a problem with voice setup in pipeline. I remember I had that.
tribal holly
I also have tried to sent tts with piper to the voice pa and it doesn't work. However sending a media file does
lone fossil
Try in pipeline debug if you can actually listen for result audio.
lone fossil
I also have tried to sent tts with piper to the voice pa and it doesn't work. Ho...
Well that's Piper trouble.
tribal holly
I don't hear anything on the pipeline debug
I also tried to reinstall it all still not working
tribal holly
Can anyone help me diagnose this?
dire spade
You'd have to investigate the piper logs and see what is going on there it seems
Could try to clear tts cache using the tts.clear_cache action in Dev tools
tribal holly
I just reset the cache the piper log tells nothing 😭. Its not just the voice preview its my android phone which can't hear it either.
When using the companion app triggering assist to try get it to say "how can I assist" all that's printed on the piper log is
DEBUG:wyoming_piper.handler:Sent info
Which seems unrelated
The tts folder is empty 🤷♂️
tribal holly
Using developer mode action tts.speak I actually get some errors.
DEBUG:wyoming_piper.handler:Text stream started: voice=None
DEBUG:wyoming_piper.handler:Synthesize(text='Test', voice=None, context=None)
DEBUG:wyoming_piper.handler:synthesize: raw_text=Test, text='Test.'
DEBUG:wyoming_piper.handler:Loading voice: None
DEBUG:wyoming_piper.download:Checking /data/en_US-hfc_male-medium.onnx
DEBUG:wyoming_piper.download:Checking /data/en_US-hfc_male-medium.onnx.json
DEBUG:wyoming_piper.handler:Text stream stopped
ERROR:asyncio:Task exception was never retrieved
future: <Task finished name='wyoming event handler' coro=<AsyncEventHandler.run() done, defined at /usr/local/lib/python3.11/dist-packages/wyoming/server.py:31> exception=ConnectionResetError('Connection lost')>
Traceback (most recent call last):
File "/usr/local/lib/python3.11/dist-packages/wyoming/server.py", line 41, in run
if not (await self.handle_event(event)):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/dist-packages/wyoming_piper/handler.py", line 56, in handle_event
await self.write_event(self.wyoming_info_event)
File "/usr/local/lib/python3.11/dist-packages/wyoming/server.py", line 29, in write_event
await async_write_event(event, self.writer)
File "/usr/local/lib/python3.11/dist-packages/wyoming/event.py", line 136, in async_write_event
await writer.drain()
File "/usr/lib/python3.11/asyncio/streams.py", line 378, in drain
await self._protocol._drain_helper()
File "/usr/lib/python3.11/asyncio/streams.py", line 167, in _drain_helper
raise ConnectionResetError('Connection lost')
ConnectionResetError: Connection lost