Thank you! All good - I’m now gong through the words.json - fabulous!

@strawberry-3.141 Thanks for following this up. That debug switch is a handy one to know :)

I’m now running with debug enabled and pouring over the debug log. Forcing a revisit of that portion of the config.js was the best thing I could have done - I realised that I’d left my Microphone testing value set incorrectly before catching some sleep.

Now that I’ve reset that to ‘1’ - my actual microphone value, as reported by the installer, I no longer get an error.log generated or the onscreen error icon - just the No Mode Detected which is the resting state without debug and an additional ‘undefined’ with debug

This leads me to believe my problem is a Raspian/hardware one, rather than with the module - If I go into the audio device settings in the desktop, I can see my Logitech USB microphone (ALSA mixer) in the selectable sound cards but am unable to make it the default, which stays as bcm2835 ALSA.
If I try arecord to test the mic from the terminal, I get
“arecord: main:682: audio open error No such file or directory”

however, if I use 'arecord -D plughw:1,0 test.wav

I get a great wav file recording. Could this be related in some way? I don’t know enough about the code to poke around for the microphone details at this level, to figure out how the installer maps the input device values

progress of a sort…
tried a sudo apt-get update, followed by a reboot

same clean startups but between the ‘initialising’ and ‘error’ states I briefly get a ‘no mode found’ and the 127 error is gone - now I just get a 1 returned in the error log

Hmm, I too am getting the ‘Error’ mic icon message, once it finishes its initialisation phase when the mirror starts up.

I’m started in dev mode - both the web and terminal consoles look clean - the last report I have in the terminal is 'MMM-voice: Starting pocketsphinx, so I don’t know if I should expect anything further messages in that console.

I’ve taken a look in the MMM-voice/error.log which contains the value ‘127’. It’s repeatable - if i stop the mirror, remove the error log and restart, I get the same results. As a test I’ve tried altering the microphone value in the config.js from it’s initially correct value of 1 - It doesn’t affect the error, which leads me to believe it may not be a hardware issue(?)

It’s such a shame - all steps reported to have completed okay, the microphone was detected - nothing looks to have been out of the ordinary.
Is there anything else I can do to test or anywhere else I can look, before I try flattening the SD card and reinstalling?

@strawberry-3.141 Doh! Thank you :)

Hi @strawberry-3.141, I’ve just tried running the dependencies.sh from the latest .git to install MMM-voice and hit this error in step 5;

[STEP 5/6] Installing npm dependencies
dependencies.sh: line 106: cd: /root/MagicMirror/modules/MMM-voice: No such file or directory
npm WARN enoent ENOENT: no such file or directory, open ‘/root/pocketsphinx/package.json’

despite the fact that pocketsphinx seemed to come down and compile okay. I’m logged in as the pi user via ssh and ran the shell script with sudo bash dependancies.sh.

Do you have any suggestions as to what’s happened?

many thanks

Thanks for the welcome both - @cowboysdude, I couldn’t have put it better. That’s what led me here to be honest - the thriving community and the comprehensive, pluggable modules :)

I like the way Evan’s mirror works but I think I need to emulate/build on that using MM2 and cherry-picking some modules as my starting point.

I’ve been playing around with some of the 3rd party modules already (I love the remote control!) and I’ve had a quick try with the facial recognition - I need to sit down for longer with that. I almost got there having gone through the training, then just hit a few snags

I’ve been dithering over which voicecontrol module to run with. I’ll probably try the various offerings one-by-one and see which seems a natural fit

I’ll look forward to seeing the 32" project unfold - that’s a monster! :) I must admit I’ve been thinking that with the workload imposed by the various recognition modules It might pay to move to a small laptop to do the heavy-lifting, but then it seems to make sense to go the Windows 10 route, as the realsense camera, cortana, etc are built in and I’d benefit from gesture controls out of the box, which shifts me away from the MM environment…

I think I’m going to start small and work my way up. As I add more functionality, if the Pi starts to struggle that might signify the point at which I need to build a second, laptop-powered mirror with more whistles and bells

Hi all, Darren here from the UK. I’m 44 and just started my first MagicMirror2 project.

I’ve had a play with Evan Cohen’s Smart Mirror in the past which I like and will at some point possibly try to merge some of the code into my MagicMirror.

I have an Iiyama 27" LED monitor on the way and some 2-way acrylic. In the meantime I’m developing my install on an RPI3

The idea I have for my mirror workflow goes along the lines of;

1. facialrecognition -> detecting someone wakes the mirror up
2. person not recognised -> the mirror displays a default set of modules
3. person is recognised -> the mirror displays a lightweight personalised set of modules (ie calendar, fitbit, time to work if it’s a weekday, time to town if it’s a weekend)
4. speech recognition to bring up other modules as needed

It’s a bit of a mountain to climb given that my JS and Python skills are non-existent at the moment :)