Voice control question



  • so i admit, i’m a total noob with pi and with linux in general. having said that, i’ve managed to install some 3rd party modules and configure them to suit my needs. but what i was hoping for is the wow factor that comes from voice commands.

    ideally, i would be able to command the mirror to show me a picture of something or play a song or a video. but i keep running into problems before i even begin. i plan to give the mirror away as a gift, so credit card registered accounts for voice recognition are not acceptable. there are 8 voice control modules listed in the github, they are as follows:

    assistant - this seems to require payment for google speech, or at least, credit card info.
    hello mirror - this seems to only send voice commands to other modules… i don’t understand what to do with this
    mirrormirroronthewall - this has everything i want, except that it requires credit card for alexa access
    tts - this is just text-to-speech, not what i’m looking for
    voice control - this only does sleep/wake by default. could i use this to display images or video somehow? i don’t know how to make it interact with other modules.
    voice - same as above, i really have no clue where to even begin making this show pictures/videos from the web
    magic mirror voice - the readme is describing major bugs with displaying and electron versions, etc. and claims to have low configurability, doesn’t sound like it’s for me
    voicecontrol - the readme is so short, i don’t even understand what this module does.

    so that’s all of them. none of them seems to be what i need, except the ones that need credit cards. since i plan to give the mirror as a gift, i can’t have it tied to my credit card in any way. is there any other option out there to let me wake/sleep the mirror and display images/videos and play songs on voice command? can one of the above modules be configured to do what i want it to do with no need for payment account registration?

    any help would be appreciated. i’m sorry if i’m asking about things i could just search for. but i’ve spent so many hours reading up on this one thing and i’m just chasing my own tail, getting nowhere.



  • @dogfeet There are modules that will accomplish what you’re looking for. Check out MMM-voice or @Mykle1’s Hello-Lucy. You want a voice module that will send notifications to other modules to trigger actions, like turning off the screen or displaying a particular photos module.

    However, these will all be preset commands (e.g. “show me photos” or “go to sleep”). I’m not sure that even the MM implementations of Google, Alexa, etc, support complex queries and results like “show me photos of giraffes” or “play popular music videos from YouTube,” if that’s what you were hoping.



  • hello-lucy looks like a great starting place. thorough explanation of how to integrate voice control into other modules. but it’s like you say, only preset commands.

    the closest i could find to live searches is this guy’s project, https://medium.com/zenturtle/building-an-intelligent-voice-controlled-mirror-2edbc7d62c9e . he has it displaying wikipedia and wolfram info on command. but it looks like he has that data coming from an a local server and displaying on the mirror, so not really feasible if i’m giving the mirror away.

    i think it just isn’t possible to ask the mirror to show me a picture of a giraffe, or to play me some led zeppelin or something. i had just assumed from the start that there would be such a module, it seems the natural progression of the project given the newfound popularity of devices like google home or amazon echo that do similar things without a display.

    the mirror-mirror-on-the-wall module does indicate youtube functionality in the description, but i get jammed up on the credit card thing. i feel like i’m so close to the ideal solution i can taste it.



  • @dogfeet said in Voice control question:

    hello-lucy looks like a great starting place. thorough explanation of how to integrate voice control into other modules. but it’s like you say, only preset commands.

    the closest i could find to live searches is this guy’s project, https://medium.com/zenturtle/building-an-intelligent-voice-controlled-mirror-2edbc7d62c9e . he has it displaying wikipedia and wolfram info on command. but it looks like he has that data coming from an a local server and displaying on the mirror, so not really feasible if i’m giving the mirror away.

    i think it just isn’t possible to ask the mirror to show me a picture of a giraffe, or to play me some led zeppelin or something. i had just assumed from the start that there would be such a module, it seems the natural progression of the project given the newfound popularity of devices like google home or amazon echo that do similar things without a display.

    the mirror-mirror-on-the-wall module does indicate youtube functionality in the description, but i get jammed up on the credit card thing. i feel like i’m so close to the ideal solution i can taste it.

    Well yeah you would think that would be the natural progression as many have thought but… then there’s reality. No one seems interested in going that route…which makes no sense to me.



  • ya really this is true , the voice recognition in general is a very deep field okay this is frst , n secondly n as my knowledge to make a voice controll with magic mirror untill now you must past with intents n intities n so on …n that will keep ur commands very limited untill you get bored , or u ll past through google assistant or alexa or or any other voice assistant but its only voice …anyway , i have a lot 2 talk about this …



  • I will get my microphone During this Week And then will Start To Look into a combination of Magic Mirror And snips

    Snips seems To Be pretty Nice because
    They allow offline sprechen recognition (I just don’t like big tech companies to Listen To me)
    And in Addition Support more complex commands aswell
    They use MQTT As Communication which should Be possible To Integrator into MagicMirror with The MQTT module quite easily



  • @maxbachmann said in Voice control question:

    I will get my microphone During this Week And then will Start To Look into a combination of Magic Mirror And snips

    Snips seems To Be pretty Nice because
    They allow offline sprechen recognition (I just don’t like big tech companies to Listen To me)
    And in Addition Support more complex commands aswell
    They use MQTT As Communication which should Be possible To Integrator into MagicMirror with The MQTT module quite easily

    that sounds promising. displaying wiki info, google image searches, and youtube videos would add so much functionality. your kid could ask the mirror for a picture of snow white, you could ask it what an echidna looks like, you could tell it to play the beatles and go clean the house to some music, you could ask it albert eistein’s bio from wikipedia and it would display it. i would use that stuff all the time, and the show-off factor would be through the roof.



  • @maxbachmann said in Voice control question:

    Snips

    Snips ai seems to b promesing and easy to use , its really interesting , hey @maxbachmann how you doing now ;)



  • Hi,
    You can give a try to Kalliope.

    The module is here: https://github.com/kalliope-project/MMM-kalliope

    Kalliope is a free and open source assistant with self hosted STT and TTS.

    Link to Kalliope projet: https://github.com/kalliope-project/kalliope

    Have fun !



  • @Krato5 I will receive the mic tomorrow, so will probably start testing around with it on weekend

    Since they released a German version of the Automated speech recognition I will probably especially use this one.
    Although it obviously still has far more problems than the english one (especiall with ä, ö, ü), but I am in their official discord channel and they seem to work on fixes right now

    And still searching for the stuff I will use to build my first mirror aswell ^^


Log in to reply
 

Looks like your connection to MagicMirror Forum was lost, please wait while we try to reconnect.