Where to start?

I am also using Kaldi and PicoTTS (it’s OK, but compared to Google or Amazon it is still really bad :frowning: ).

Wake word is an issue. I am currently using porcupine with one of the universal models. Problem: I don’t really like the wake word choices. All feel weird to say for native german speaker and not easy to pronounce without some German accent. I am currently using “pico voice” as wake word. It recognizes me most of the time and very few false activates.

I am thinking about making my only custom wake word with precise. How did you do it (source of the training data?)

1 Like

Is it possible to use rhasspy with Amazon Pollylike it was with Snips?

Have you tried MaryTTS? I don’t know the quality of the German voices, but I’m using the voice dfki-prudence-hsmm for English and it’s really enjoyable.

1 Like

@DanielW If you want more information on training a custom wakeword model with mycroft precise take a look at this Mycroft Precise Model .
I followed the instructions from mycroft for training your own wakeword model.

What kind of hardware are you running your system on?
How fast does Mary TTS work on a pi4 ?

I wrote down some of my experiences about training a custom wake word here in this post Mycroft Precise Model Problem (computer-en.pb) especially with training a noise resistent model.

This is on a Raspberry Pi 4. I haven’t benchmarked it, but it’s fast enough for me.

I just tried the three German voices but there is no huge difference in quality compared to PicoTTS for me. But it uses more RAM and CPU. (it is OK on a Pi 4. But longer texts will slow down Rhasspy)

Got it up and running. Will first create functionality and try around with different voices later :slight_smile:

Have a look at this:


It explains, how to setup master/satellite with Rhasspy as you are used to with snips.

Of course, no problem, really works well

1 Like

What mics are you using? Respeaker or Matrix or …?

If Respeaker or Matrix have a look at HLC LED Control in this forum. Just to light up your LEDs, triggered by event

I use the PSeye cam on the big Raspi which is working great. Later I want to try my zeros with a two mic platine attached (may check for the name later) which was working great with snips before. I think those satellites have an LED I might start.

I found the respeakers just to expensive when creating multiple devices. Mist are even placed where they are invisible :slight_smile:

I think my other mics are seeed hats if I remember correct.

Hi @koan how did you get your setup to work with MaryTTS.

The play back speed is sped up as if it is fast-forwarded.

Could you give some pointers?

Hi @koan !! :slight_smile: Could you please provide instructions for building/installing that MaryTTS voice dfki-prudence-hsmm? I took a very quick look at the MaryTTS github, and am not certain where to get started. Will I need to install a full MaryTTS environment on my Raspbian OS? Thank you in advance for EVERYTHING you do for this Rhasspy community.

@FredTheFrog are you using docker? synesthesiam released a docker image for maryTTS here: https://rhasspy.readthedocs.io/en/latest/text-to-speech/#marytts

1 Like

Thank you!! Most appreciated.

I didn’t do anything special: I swapped out eSpeak for MaryTTS and it just worked. This doesn’t seem to be a MaryTTS issue, but an issue with the settings for your audio device.

I use the Docker image, works fine here.

1 Like

Hmm :thinking: if my audio settings were the issue, I would think that espeak should also exhibit the same issue, but it does not. Espeak is the only TTS that works (meaning I can hear the words, and not something that sounds like its fast-forwarded). Can anyone point me to some docs to help me start diagnosing this issue. I’m not really sure what keywords to search for.