Thanks so far.

I have such a mistake.

[STEP 1/5] Installing Packages
dependencies.sh: line 36: sudo: command not found
[STEP 1/5] Installing Packages | Failed

[STEP 1/5] Installing Packages
dependencies.sh: line 36: apt-get: command not found
[STEP 1/5] Installing Packages | Failed

I tried to install packages one at a time but without much effect

If I say “OPEN HELP” it understands “OPEN OF OF” or “OPEN THE”
If I say “MAGIC MIRROR” he understands “MAGIC TO”

The recognition is not as accurate a Google Home or an Amazon Alexa. Take into account that this is not cloud based recognition. You almost have to learn how to speak to MMM-voice correctly to minimize the false positives and negatives. I’ve been using MMM-voice(Hello-Lucy) for a long time. I’ve learned how to speak to her so that my success rate is more than acceptable.

Also, I have a friend in Norway that speaks English perfectly. However, MMM-voice(Hello-Lucy) has some trouble with his Norwegian accent but his Google Home has no trouble at all. That’s the benefit of cloud based recognition.

@egnos said in MMM-voice:

No advice for the purchase?

The little usb microphone that you are using is behind the mirror? I think this will lessen your chances of success. Any sounds behind the mirror will certainly interfere and your voice commands will have difficulty reaching it. I don’t have a USB microphone recommendation but anything you do get you should expose somehow to the front of the mirror so that the user’s voice has the best chance of reaching the microphone unimpeded.

Thanks

• If I say “OPEN HELP” it understands “OPEN OF OF” or “OPEN THE”
• If I say “MAGIC MIRROR” he understands “MAGIC TO”

I think nobody has had these problems … correct?

I’m starting to think that the problem is the quality of the microphone.
I bought this microphone Microphone-Raspberry-Recognition.

I have recording a test and I hear a background noise…
Can someone recommend a good microphone?

Thank you all.

LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib

I wasted a half year ago a couple of weeks with this module because it does not work for me, till now!
Yesterday i’ve reinstalled the Whole MM on a fresh Raspbian system and read all the comments here.

But there is one question still alive, i am in struggle with the commands i am not sure if i say it right.

MAGIC MIRROR -> VOICE = (going to Voice Mode)
(VOICE flashing) -> HIDE MODULES, SHOW MODULES, GOING TO SLEEP, WAKE UP

That works fine.

But what should i do if (VOICE) don’t flashing?
and are there different modes instead of VOICE?
oh and last but not least if i say (in flashing voice mode) Open Help or Close Help nothing happens :/

D

Is there already another Programm running in the background which is accessing the mic?

“Error opening audio device plughw:1 for capture: Connection refused
FATAL: “continuous.c”, line 245: Failed to open audio device”

When i run “cat /proc/asound/cards” i get this
“0 [ALSA ]: bcm2835_alsa - bcm2835 ALSA
bcm2835 ALSA
1 [Set ]: USB-Audio - USB Headphone Set
USB Headphone Set at usb-3f980000.usb-1.5, full speed”

And in the config.js file obviously i have selected the only mic i have right now number 1 USB headphone set and i know it is working just fine.

Let me know how what does that tell you
and again thank you
Denis

as the author of MMM-voice I can tell you, that my module doesn’t support a command called hide clock at all. You are probably using hello-lucy, at least that’s a mod I’m aware of which enhances the commands. You can see the command list here https://github.com/fewieden/MMM-voice#usage

i also had problems due to my accent. you can also look into the dic file how the system expects you to pronounce it

if I speak slowly, i can get ‘magic mirror’
then wake up is ‘wake go’ or ‘wake all’,

hide calendar is ‘hide who’
open help is ‘open go’

the smart-mirror project also uses ‘go to sleep’ and ‘wakeup’

‘hide newsfeed’ is ‘hide to show the’, ‘hide is the’

none of the commands, when recognized, do anything…

hide modules

go to sleep (not on PI so the tvservice doesn’t exist), but no error is logged

I also use another voice enabled smart mirror on this same device.

with a Logitech webcam as the mic.

MMM-voice gets the hotword correct every time, but almost none of the following words.

the same hardware setup works correctly with arecord as the wav source, but uses google speech detection for everything after the hotword.

it appears the system is using the default language model (lm) and dictation (dic) files.

anyone have a suggestion on debugging this?