MMM-voice

cowboysdude

pocketsphinx_continuous

Try getting rid of pocketsphinx and reinstalling it…

tbbear

Hello,
im cant understand the problem i have:
First i can arecord and aplay without any problems with my usb audio device.
But when i try to use MMM-voice i always get:
(snipet from the debug file)
…
INFO: lm_trie.c(482): Building LM trie
INFO: ngram_search_fwdtree.c(74): Initializing search tree
INFO: ngram_search_fwdtree.c(101): 32 unique initial diphones
INFO: ngram_search_fwdtree.c(186): Creating search channels
INFO: ngram_search_fwdtree.c(323): Max nonroot chan increased to 168
INFO: ngram_search_fwdtree.c(333): Created 32 root, 40 non-root channels, 5 single-phone words
INFO: ngram_search_fwdflat.c(157): fwdflat: min_ef_width = 4, max_sf_win = 25
INFO: continuous.c(307): pocketsphinx_continuous COMPILED ON: Jan 6 2018, AT: 13:45:05

Error opening audio device plughw:0 for capture: Connection refused
FATAL: “continuous.c”, line 245: Failed to open audio device

Please help me !!!

tbbear

@tbbear Its always the same, better look twice before asking. I found one old pulseaudio.h file and this was the problem. Everything works fine now and i can start to work on supporting my MMM-Openhab with MMM-voice.

Mykle1

@tbbear said in MMM-voice:

Its always the same, better look twice before asking.

Haha Robert. It ALWAYS happens that way, doesn’t it? :-)

Glad you found it now get to work. ;-)

AxLed

Hi Strawberry,
do you know, if i can change the language of MMM-voice with Pocketsphinx / Sphinxbase to german?

I found a webseite where i can create .dic and .lm files and they work somehow (i also changed the new words in different .js files of folder MMM-voice).

I am looking forward hearing from you.

Regards

AxLED

strawberry 3.141

@AxLed https://cmusphinx.github.io/wiki/faq/#q-which-languages-are-supported

Peter

I would like to have this module working, but up till now no succes.
It is a bit confusing.
It says MagicMirror/modules/MMM-voice/ installers as a directory, but after cloning it says: MMM-voice-master.
So I tried both directories; MMM-voice and MMM-voice-master.
Both did not work.
Do I have to see a voice symbol on the screen?
I tried the commands: Magic Mirror, Voice, go to sleep.
Nothing happens.
I have a USB microfone connected as microphone 1.
Installing the module seemed to be ok.
Only in the end: step 6/6; No such file or directory.
Any help would be appreciated.
Peter

Mykle1

@Peter

I was in hospital for a few days. :-(

When you clone MMM-voice into your modules folder, the “installers” folder is inside of that. The readme in the repo gives directions for installation

strawberry 3.141

@Peter if you clone the repository the directory is called MMM-voice, only when you manually download it it will be name MMM-voice-master and that isnt cloning.

it looks like youre trying to use an old version, as the current one (since august last year) had just 5 steps in the installer not 6

Peter

Mykle1: hope you are doing allright! You are a good help to me from the beginning of my project.
Strawberry: you are right; I did not clone but downloaded!
Maybe that is the problem.
A working day is waiting, but after that I will try cloning (have to do some research how to).
I hope it will work!
Thanks!
Peter

strawberry 3.141

@Peter type this in your terminal

cd ~/MagicMirror/modules
git clone https://github.com/fewieden/MMM-voice.git

then follow the instructions here

Peter

@strawberry-3.141 cloning went well!
No step 6, so this was the right version.
I can see a microphone on the screen (with the word ERROR).
I still have to find out where the error comes from.
Thanks for all your help!
Peter

Peter

Errorlog:127
Debuglog: pocketsphinx_continuous: error while loading shared libraries: libpocketsphinx.so.3: cannot open shared object file: No such file or directory

Why is it so hard to get this working?
Peter

Mykle1

@Peter said in MMM-voice:

Why is it so hard to get this working?

Maybe this will help you. It has helped a few people get up and running.

Assuming you installed the dependencies as per the readme. At the end of the dependencies installation is a readout for your microphone entry in the config.js. You can see this readout again. in your terminal type: cat /proc/asound/cards. You can verify your microphone this way.

And this should remove the error that you are receiving. Close all your terminals.
Open a new terminal.
type: LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib and Enter

Then type: sudo ldconfig and Enter

Now try it. Good luck

Peter

@Mykle1 magic!
It works!
Thanks!
Peter

Mykle1

@Peter

The real magic comes from Strawberry. :-)

Glad you got it working. :-)

Cyberdale008

Hi folks, usually don’t post. Just read a lot. But I am a lil over my head here. ASE Certified Master Mechanic here, but dang this coding is tough!! This module works fantasic. However I would like to change this ( /opt/vc/bin/tvservice -p && sudo chvt 6 && sudo chvt 7 ) to this ( /opt/vc/bin/vcgencmd display_power 1 ) and ( /opt/vc/bin/tvservice -o’, null ) to this ( /opt/vc/bin/vcgencmd display_power 0 ) I just want the tv to go off, I don’t want the no signal or light off the screen to show up. Is this even possible. Thanks in advance for any help!!

cowboysdude

@Cyberdale008 sure it is…just to the node_helper and change it.

Look for these lines :

} else if (/(WAKE)/g.test(data) && /(UP)/g.test(data)) {
           exec('/opt/vc/bin/tvservice -p && sudo chvt 6 && sudo chvt 7', null);
           this.hdmi = true;
       } else if (/(GO)/g.test(data) && /(SLEEP)/g.test(data)) {
           exec('/opt/vc/bin/tvservice -o', null);

and make your changes :)

AxLed

@strawberry-3.141 said in MMM-voice:

@AxLed https://cmusphinx.github.io/wiki/faq/#q-which-languages-are-supported

@strawberrxyy 3.141

I tried to change the language to german / german voice recognition with some informations from here:
https://mycroft.ai/documentation/language-support/german/
but my skills a not good enough to get it working.

3 more questions:
Is there a command for commandline to test Pocketsphinx and its components, so maybe i can find out where i went in the wrong direction.

Do i need for changing the language model only have to download certain files and bring them in the right structure or do i have to compile something afterwards?

Where could i change the path to other language models in MMM-voice, i only found links to the .lm an .dic file but not to the .../model/en-us/ structure.

AxLED

strawberry 3.141

@AxLed

pocketsphinx_continuous
You need to look for instructions where you downloaded those files
you can specify the language model with the option -hmm in the node dependencie pocketsphinx-continuous-node