Microphone advise for PI2 & voice control



  • Hello,

    Iam currently improving my PI2 with MM & would like to add a microphone for voice control.

    Can someone suggest me a good HW choice for microphone?

    BR Jörn



  • Obviously depends on what exacty are your needs for the microphone
    I really like my respeaker 4mic array (was about 25$ and works quite well with a bit of background music and distances of over 5m), but there are obviously other options.
    Generally from what I heard Respeaker and Matrix are quite good. Then you basically have to choose depending on the features you need.
    I know quite a few guys that use the reapeaker two mic hat (around 15$) which is a bit cheaper and still works very good. However when you for example play music next to your PI it might make sence to get one that a AEC feature so you can remove a lot of this background music from the recorded audio.

    For multiroom setups it might make sence to have a microphone array with a atmega32u4 for each room so they can directly stream the audio to the main device which does the speech processing



  • @maxbachmann said in Microphone advise for PI2 & voice control:

    respeaker 4mic array

    Hello,

    thanks, respeaker 4mic array looks really cool & can be attached directly to the PI, maybe I order afterwards if current choice is not satisfying.
    I meanwhile ordered a "Kinobo Mini Akiro USB ", which has a nice housing, to be attached to side of mirror.

    Thanks



  • @joern said in Microphone advise for PI2 & voice control:

    respeaker 4mic array

    I have a question can i use 2 microphones on a raspberry pi ?? Either usb or something like the respeaker where both mics are monitored by different modules one for MM-voice and the other for Alexa Pi … (MM-voice will let you choose a mic ID, where it listens to its wake words)
    I Just don’t know how to have raspberry pi’s configuration twitched, which i am sure it needs to be, to have 2 mics working???
    Denis



  • Why wouldn’t you use the same for both? makes not much sence to use two microphone arrays when both modules can just listen to the same. Or do you mean to have mics in multiple rooms? Two respeaker is not really possible since they would use the same GPIO pins, but yes in general you could connect another one over USB. just can’t see why you would want to do this


  • Module Developer

    @richland007
    I have no idea why you want to find a good mic, but with my experience, the cheapest mini USB mic from Amazon is too enough for the most purpose of vocal input.



  • @Sean @richland007 Well I guess that highly depends on what you want to achieve xD
    I would go cracy with my voice recogntion without a decent Mic array.

    But yes when your in a pretty quiet environment and only using it with MM2 --> standing in front of the mirror --> small distance a cheap mic should absolutely be enough.
    My minimal needs for a voice recognition are around 5m (obviously without shouting) distance 😅

    However I can’t think of any purpose where two mics would make sence at one device for two modules. Only thing that would make sence is having one mic in each room streaming the audio to a main device that does the speech recognition work (comparable to Alexa with the echo dots)

    Well for a cheap one my suggestion would be a respeaker two mic hat (9.9$).



  • @maxbachmann @Sean It is my understanding that not 2 separate programs (AlexaPi and MMM-voice) use the same resources which is the one and only usb microphone i have right now.
    Am i right on this??.. cause boy do i wish to be wrong!!
    Has anyone used 2 progs one mic…and i haven’t even tired it but someone said you couldn’t and i think they are right myself… but i will give it a try and hope it works.



  • @richland007 it definetly works. and why should it not? On the spur of the moment I can’t think of any pc connected device that works with only one software (well excluding some proprietary devices that work only with the company’s stuff).
    However for example Mouse, Keyboard, speakers, microphone … all definetly work with multiple programs. The components are always connected to the OS, so any software running on that OS can generally ask it for the informations.
    I mean if I wanted I could even split for example the cable of the microphone and connect t to two computers and it would work ^^
    I never used Alexa Pi, since it is cloud based and I am no big fan of putting my stuff in the cloud, but usually when setting up voice recognitions, you tell which microphone to listen to

    And as an example where you use one muc for 2 programs: when you play video games with a voice chat program in the background and still use the ingame chat 😉


  • Module Developer

    I cannot assure, but there could be a possibility of resource possession.
    With my experience, Hotword catching program should always hold Mic on to listen the hotwords, so when 2 programs are trying to get a mic together, there would become fails.
    I don’t know whether AlexaPi and MMM-Voice are those kinds of program, But without proper pre-processor program for distributing and controlling mic, the two modules will be competing to obtain mic from each other. And with my guess, those two modules are not developed as that way.

    By Example, My modules - MMM-Hotword and MMM-AssistantMk2 both are using one mic together, but when one module is using a mic, the other should release it before. Both modules are made by me, so I can make both to work together. But I cannot think MMM-Voice and AlexaPi both could do the same thing.