synesthesiam

On my list to check out tomorrow :D

As the sibling comment mentions, the next version of Piper will no longer use espeak-ng to avoid potential GPL licensing issues.

Yes, it is possible by installing the Wyoming Satellite software on the Mark II. All the pieces are there, it's just missing some easy-to-install firmware. Source: I worked at Mycroft on the Mark II and am now the voice guy for Nabu Casa (Home Assistant).

Look into training a Dreambooth model. Huggingface has a guide for this.

You may want to try Piper for this case (RPi 4): https://github.com/rhasspy/piper

Thank you for the kind words, @follower!

I'm the author of Piper; it is a successor to Larynx (originally named Larynx 2). Piper uses the same underlying model as Mimic 3, which I developed before joining Mycroft. However, Piper uses a different library to get word pronunciations, so the voices aren't compatible between the two projects.

It's been an awesome year so far with Nabu Casa, and I'm very fortunate to be able to work on something I love. I hope to contribute to the open source voice space for many years to come :)

Neon AI and OVOS are taking over development of Mycroft Core and the Mark II: https://www.reddit.com/r/Mycroftai/comments/1212h87/neon_ai_...

Coauthor of the blog post here. You're right, I said it on the live stream we had today but forgot to mention it in the blog post: the i5 is from a Lenovo ThinkCentre M72e. They're available refurbished for less than the cost of a Pi 4 these days, so it seemed to be a good comparison!

I used to work for Mycroft, so I'm hoping to eventually create an image that's compatible with Home Assistant pipelines.

For now, though, you may want to check out OVOS: https://openvoiceos.com/

Home Assistant and Rhasspy 3 will be able to share voice services thanks to the shared Wyoming protocol. Rhasspy 3 will have more options, including lots of experimental services.

Mark I or Mark II?

I'm working right now on adding voice control to Home Assistant: https://www.home-assistant.io/blog/2022/12/20/year-of-voice/

Our goal is local voice control of your devices in your native language (no Internet required). Hardware is still an open question, but I believe we will succeed due to our focus on a limited domain (IoT) and large open source community.

Have you tried whisper-cpp?

https://github.com/ggerganov/whisper.cpp/tree/master/example...

This has always been a struggle. Rhasspy can gather lists of songs, artists, etc. but it will have to guess many of their pronunciations. And it seems artist/band names often purposely thwart conventional pronunciation rules :P

They share a lot of the same pieces, but voice2json is meant to work in Unix-style pipelines. Rhasspy has MQTT/HTTP/Websocket APIs instead.

Sorry about that. I'm going to be working more closely with native German speakers to get it right!

This was the reason I designed voice2json [1] :)

[1] https://voice2json.org/

You're welcome! What sort of hardware did you settle on for the satellites?

Rhasspy author here, thanks for posting! Just wanted to mention that I've joined Nabu Casa (creators of Home Assistant) this month, so Rhasspy will be receiving updates again and be a major part of Home Assistant's "Year of Voice" in 2023 :)

Is there a good guide for writing Rust-like C++, e.g. using stuff in the recent standards to avoid the many footguns?

The amount of data depends on if there's a voice for the language already. If so, about 2 hours of data is usually good enough. Otherwise, 10-20 hours usually does it.

The voices are under a CC-BY-SA license, so you can generate all the audio you want (offline), even for commercial usage.

Hi all, author here. Besides the tech of Mimic 3 itself, I'm interested in training voices in as many (human) languages as possible. All it takes is one person willing to donate a dataset for everyone to benefit!

...well, that and a bunch of stuff with phonemes. But I'll do that part :)

I believe I fixed the bean bug ;)

Python is only really the glue here. The models are trained in PyTorch and exported to Microsoft's Onnx runtime (C++). So the bulk of the inference CPU cycles are outside Python.

Author of Mimic 3 here. Thanks! Hoping to release the first version this month :)

Sure. I joined the company last year, and Joshua has mainly been busy appearing in court to fight the patent troll. With that fight effectively over, Joshua and Kris are focusing on something new (not sure what specifically).

Michael Lewis has been running the company for two years, so this resignation is more symbolic than anything.

It's a reference to the sentient computer from: https://en.m.wikipedia.org/wiki/The_Moon_Is_a_Harsh_Mistress

I'd normally agree, but Joshua (the founder) has been minimally involved for years outside of the patent lawsuits. With that mostly resolved (in Mycroft's favor), he's moving on to the next thing.

Joshua Montgomery founded Mycroft AI, but Michael Lewis has been running the company for 2 years as CEO. Joshua recently "resigned", though his day to day involvement has been minimal.