🔥 Voice interview Michael Hansen | HA | Raspberry | Piper | Rhasspy

Interview with Michael Hansen from ‪@Nabu-Casa‬ . We talked about opensource voice technology, ‪@home_assistant‬ , Piper TTS, Mycroft AI, Coqui AI, locally running and privacy aware voice technology, simple Wyoming protocol and much more.
It's been a pleasure having you as guest, dear Synesthesiam 😊.
00:00 Introduction Mike
03:10 Voice data support for underrepresented languages
07:35 Talking on Piper TTS
10:30 Raspberry Pi Zero (Piper TTS)
11:50 Streaming with Piper TTS
14:45 Piper TTS server mode / STDIN / JSON
16:50 Wyoming protocol
20:10 Piper TTS Python integration
22:00 SSML Support for Piper TTS
24:30 Microsoft SAPI and Piper TTS
26:45 UI for Piper TTS and longer text inputs
28:55 Synesthesiam - wtf?!
32:00 Subscribe Thorsten-Voice TH-cam channel
33:45 Shutting down of Coqui AI
38:50 Best hardware setup for a private voice assistant
49:55 Home Assistant year of the voice
1:00:00 Rhasspy Roadmap
1:03:48 Outro
Do you want to support me? Please subscribe to my channel. If you like you can donate using ko-fi:
ko-fi.com/thorstenvoice
My Piper TTS playlist: • Piper TTS
Links:
===
* Nabu Casa: www.nabucasa.com/
* Home Assistant: www.home-assistant.io/
* Year of voice: www.home-assistant.io/blog/20...
* Voice assistant contest: www.home-assistant.io/blog/20...
* github.com/rhasspy/piper
* github.com/synesthesiam
* github.com/rhasspy/rhasspy
* github.com/rhasspy/wyoming
HW recommendations:
===
Please subscribe to my channel 😊.
th-cam.com/users/ThorstenMue...
---
- www.Thorsten-Voice.de
- github.com/thorstenMueller/Th...

[@spiritual_audiobooks]

My biggest thanks to Michael Hansen, who made my TH-cam channel possible through Piper TTS and thus made an incredible number of book treasures available as audiobooks for many grateful listeners. ❤❤❤

[@zugaldia]

Great interview. Keep them coming!

[@ThorstenMueller]

Thanks a lot for your kind feedback 😊. Do you have any special person in mind you would like to see an interview with?

[@bephrem]

very exciting

[@beneadie3202]

Love your Channel Thorsten. Could you do a video on how to run NVIDIA's Riva FastPitch? The demos on their site look amazing but I am such a noob that I can't even understand how they can run. Thank you in advance 😊

[@ThorstenMueller]

Thanks for your kind feedback, i really appreciate it 😊. I've added NVIDIA Riva FastPitch on my TODO list but can't tell you when i'll find the time for creating a video on it.

[@biskero]

running on RPI 5 and it's working fine. Just want to clone my voice and cannot find the way to do it. I have bunch of wav files but how do I do the training model .onnx?

[@ThorstenMueller]

Happy it's working for you. I made a step by step tutorial on how to clone your voice using Piper TTS - have you already seen this? th-cam.com/video/b_we_jma220/w-d-xo.htmlsi=yjdXJIJQ1p693jRy

[@biskero]

@@ThorstenMueller I saw the video and looks a tedious process. I was hoping to a more streamline process like providing some scripts that take a set of wave files, run a script and training is done. In any case will take a look again and see how it goes. Thx a lot for the videos are very helpful !

[@__________________________6910]

I love Piper It is fast, offline, and lightweight but the voice quality needs to improve like it is more realistic (add emotions)

[@ThorstenMueller]

Maybe the announced future work on SSML could bring more benefits in context of realistic to it. For real emotions you need emotional voice datasets to be trained on. Btw. for my german Thorsten-Voice models there are emotions available 😉.

[@__________________________6910]

@@ThorstenMueller 😊