I had been running Snips and then it died.
I then tried Mycroft but it didn’t seem to like female or kid voices. Now I haven’t tried the latest release so maybe that’s changed.
I then saw a video on Almond w/ Ada but I can’t find any real details on this setup.
I haven’t looked into Rhasspy.
Does anyone have any details on Almond w/ Ada?
How does Rhasspy compare to Snips? Does it integrate with MQTT so I can reuse some of my old MQTT flows (in nodeRed) I had with Snips?
Rhasspy is well documented! I suggest a read.
I’m hoping for some personal experience with the different options. I read a tonne about Mycroft but not once did I see a reference to the fact it really only worked with deeper voices, also there was no mention of random phantom wake-word detections.
I’m hoping to learn from real-world experiences before I dump even more time into this part of my hobby. I’ve burnt so much time on Snips and Mycroft already. Maybe Mycroft needs a bit more time, or maybe there is something better,
Fair enough, it is a deep dive. Like most of my more advanced protects I spend so much time on the dive that the project remains in my head!
I have installed rhasspy. The docs are good, the developer is helpful. But I haven’t used it ”in anger”. My reference to the docs in my first answer was really just to help you with the mqtt question. I am pretty sure the answer to that is yes but it is a fast moving area.
Id love to be more helpful.
Hey @nickrout I do the same with projects. Sometimes the projects gets so complicated/buried in details that it never sees the light of day.
I’m definitely checking out the docs.
I’ve watched a few videos on Rhasspy and it looks impressive.
How is Rhasspy at handling higher pitched voices?
That I do not know. They have a forum, it might know. I am surprised at your problems with voice pitch, could it be a hardware problem? Bad microphone?
I have the respeaker 2 mic array, it worked really well with Snips, even picked up commands with music playing.
The Mycroft forum indicated there was known issues with higher pitched voices as the majority of voice samples they had were from men