Speaker Suggestions

Hey everyone,
I just signed up to get some more suggestions for speakers for a Linux-Mycroft Box. Since the Mark II is still not available and I didn’t want to pay via Creditcard (please add Paypal…) I build a setup on some old hardware that was left over.
But I’m still looking for decent speakerphones / mic/speaker combo. I tried Jabra 510 and 710, but sorry, this sound is not anywhere close to what I want to listen to. I don’t mind the weather forecast, but listening to music is a pain.

So anyone has any suggestions? I’d love some Bose or Ultimate Ears device, but I’m not sure I can use those as a mic on a linux machine or the rpi.
Or do I need to look for a mic only + get some simple speaker which offers better sound than the Jabra? I use a pretty cheap JBL wireless/bluetooth speaker for my notebook right now and sound is a lot better than with the Jabra. But no mic…

Any ideas?

I’d use a separate mic and speakers if you’re looking for audio quality. Also mic position can be effective in improving sensitivity, which may not be as helpful for the speakers.

1 Like

+1 on the separate mic and speakers

Not sure what hardware you’re running it on, however if sound quality is the aim, a good amplifier is pretty important.

thanks for your replies. Well, “sound quality” is pretty relative, right? I was just a bit disappointed in the Jabra 710 speaker/mic. I’m ok with my bluetooth JBL wireless speaker, small speaker, nothing special, 50€ max - but way better sound than the 710 via USB.

What do you guys think of respeakers for the PI? Something like this: https://www.amazon.de/Seeed-Studio-respeaker-4-mic-Raspberry/dp/B076SSR1W1

We’ve found the ReSpeaker Mic Array v2.0 to be quite good

It’s more expensive but we think it’s worth it. We’re even adding it as an option in the Picroft installer.

I use a cheap BT-Speaker with AUX-In which sounds ok.
(Disclaimer: The following Options could be overpriced but where simply available at home)

As Microphone Option I started with the Matrix Voice. It worked ok, but while the wakeup was recognized very good, everything after that only worked when very close. As the hardware SHOULD be perfect I personally expect a driver issue.

Currently I use a Yeti blue microphone. It is originally designed for podcasts but is currently not in use. The big benefit is that you can change the characteristics of the mic (e.g. 180` degree or 360 degree). That is really close to perfection. It seems to recognize me even better than the original Google Home.

Buying that only for this purpose would most likely very overpriced, but any good microphone like the Auna with a Cardioid or omnidirectional should be fine. You should be able to find those for a fair price at eBay

1 Like

I ordered the ReSpeaker Array V2, thanks for the tipp.
I’ll test it with my new RPI4, as soon as it is arriving…

1 Like

@gez-mycroft In regards to the Respeaker Mic Array 2 (which is what I am starting with) any suggestions on which firmware to use, the 1 channel or the 6 channel? Do you know if all the advanced features are enabled in the hardware or if they have to be enabled somehow on the rpi3?

We’re using the 1 channel firmware.

You can see the exact setup steps we’re using for the next Picroft image here:

1 Like

Thats really great news because I think Respaeker Mic array is one of the best speakers out there. Is it now also part of the Buster image for Picroft or do I need to create now my own Picroft image from the github master ?

Why dont you use the 6 channel firmware ? The respeaker has 6 mic channels as input.

This was based on both a recommendation from our chats with Seeed (the company behind ReSpeaker) and our own testing. The 1 channel firmware performed significantly better. I’m sure that will change in the future as the other firmware versions are improved.


Just looking at that TODO line in the linked bash script. Why using lsusb for that? You really need udev for that type of things.

Seeed uses udev for their pulseaudio configs as well;

But you could easily (I believe) use udev to check if it is plugged in or not.

Great. Thanks. Very good information.

There is no need for the 6-channel as 1-channel contains the preprocessed audio - which is all you need for the Mycroft application. See also Seeedstudio Wiki

Yeah that makes sense, are you somewhat familiar with udev?

If you feel like submitting a PR… :smile: