Deployment Rationales for Whole Home Mic Systems

[LilFonky]

I decided to move some of my discussion here from the whole home thread I started in the mic recommendations forum now that this new sub-forum is going.  Thanks to the moderators for setting this up.

Hadood and Nime5ter started a useful side discussion I've been meaning to respond to, so I'd like to pick it back up here.  Just to get the discussion on track, Nime5ter's commented:

My own perspective is that the desire for an open-air mic solution is the main issue.

We have a few Androids (along with some Amulet remotes, some USB mics, etc.). We can, and do, use VoxWav anywhere in our 2-storey house to control VC -- it's not a problem for us because we're not trying to use our Androids as open/area mics. For us, these are a reliable, convenient, and affordable (since we're using them for multiple purposes anyhow) choice.

Someone recently asked James what kind of open mic solution he's using in the "Coming Home from Work" demo video. The answer is simple: None. Every time he's talking to the computer, he has a mobile device or Amulet in his hand. And yet it apparently seemed natural enough that someone watching the video didn't notice that.

For me, it doesn't feel inconvenient or awkward to pick up the mobile or tablet and click a button before talking, anymore than it feels inconvenient to pick up the phone to call someone or walk to the top of the stairs if I need someone to hear me at the bottom of the stairs.

Sometimes, I'll unthinkingly go "old school" and call out to James from another room to ask him to turn down the music or change the living room speakers to the outdoor speakers, etc. Then suddenly I'll think, "Why the heck am I doing that?" Wherever I am, I pick up my phone (or open Tellvox on my laptop if I'm in my office) and issue the command myself. It takes literally a few seconds, and for me, it's a million times simpler and more reliable than having to yell my command at a microphone across the room.

So, the preference to use some kind of always on hands-free open-air mic system is the main challenge.

A reliable hands-free solution seems like it would be particularly important for users with disabilities and mobility challenges, but that wouldn't necessarily have to be open-air. Whenever I think about this issue, that's the problem I am interested in finding a good solution for.

I think these observations are sound, and like James and Nim5ster I'm using multiple Android devices and laptops as controllers in my own test deployment.  And while Nime5ter's point about the convenience having tablets ready to hand is well taken, in my own personal setup using an Android device as a sort of "walkie-talkie" interface with which to interact with my other devices doesn't quite satisfy Occam's razor. This is to say that in addition to voice controls, each of my Android devices already has touch controls running as home screen widgets which I can use to directly control any device in the house.  Using VoxWav requires me to perform all the UI touching I would normally do with my widgets, and then on top of that I have to talk to the tablet, which adds an additional and mostly superfluous step to the process.  So my desire to have an open mic system stems from a more basic desire to add new control options to augment an already working system rather than replace the old touch system with something different. So far, Vangelis's boundary mic model has great appeal for me, I'm just not able to drop wire through the walls since I'm renting, so I require either an RF or an over-the-network solution for audio input.   

On a purely conceptual level, my interest in voice control is a function of my broader interest in the notion of "ubiquitous computing", which is the idea that almost everything in the built environment has a network attached computer of some sort in it or on it, which allows it to interact with most or all of the other objects in the system.  In my fantasy of an an ideally implemented lifestyle automation system I should be able to operate most any device in the system using either a graphical/touch interface, a speech interface, or a combination both.  Further, if done right, one mode of operation should not render another mode trivial just by dent of its existence. And while I'm fantasizing, I shouldn't have to be constantly locking and unlocking and rebooting and turning on and off and logging in and out of things all day long.  I just want to be able to spontaneously get things done in whatever way makes the most sense from within the context of whatever else I happen to be doing at the moment. If my hands are full of groceries, or if I'm chopping celery, or if I'm just too lazy to go grab a tablet out of the other room, then I want to use my voice wherever I'm standing without having to drop what I'm doing.  If it's late at night, or if someone is sleeping, or if a tablet is already in hand, touch controls make more sense. 

Anyway I'd like to hear more from others, not just about which pieces you're all using in your VC roll-out, but also how it should all fit together in the grand scheme of your imagination. 

     

[nime5ter]

This is to say that in addition to voice controls, each of my Android devices already has touch controls running as home screen widgets which I can use to directly control any device in the house.

Fair enough. For me, there are certain tasks for which it simply makes more sense to press a button than to use a voice command. I consider an integrated solution one in which the various elements work together, so that I can use the most efficient solution depending on the context and so that I have various options/redundancies that help to ensure reliability.

What I like about VC compared to relying on a bunch of apps is it is a one-stop shop. I don't have to open one app to control my Vera, and another app to queue a MediaMonkey playlist and otherwise customize my media experience, and another app to add an item to my grocery list, etc. I have VoxWav running, and when I want to queue songs rated 4 stars or higher, switch from TV output to stereo output, turn on a light, check the weather, and remind myself to buy flour, it's one integrated solution -- but not hands-free.

Open-air is noble aim, and everyone here is very interested in seeing how various people address it, so it's great that we have these new threads popping up on the forum. I look forward to reading about more people's setups. It's probably more feasible for those with more disposable income, those who live in fairly quiet environments, etc.

I only know that for our particular lifestyle, any time we've tried an open-air solution it has been a lot more frustrating and significantly less reliable, whereas the method currently used is so fully integrated after four years that I can no longer imagine home life without voice control at the centre of it. That is why I wanted to explain our setup even though it's not an open-air solution.

It's more that I think it's important for new users to think about whether open air is the first part of the equation they want to address, or the most important. There's a pretty big learning curve with VC and a lot that one can do, but it takes time and exploration to become familiarized with it. Much of its versatility and core functionality is overlooked when new users immediately concentrate on the open-air aspect; oftentimes all of their efforts then go into doing some pretty crazy things to get fairly basic performance, while never really learning how VC works.

That said, definitely -- go crazy, innovate, enjoy, and share your vision!

[LilFonky]

What I like about VC compared to relying on a bunch of apps is it is a one-stop shop.

I totally agree.  If I had all my touch controls under a single interface as well I'd be delighted. I wish I could trigger VC commands with touch using an Android widget. Touch Commando, if you will. You can sort of do this with VoxWav, once you've issued a few commands, but that's not really what the app is for...

Open-air is noble aim, and everyone here is very interested in seeing how various people address it, so it's great that we have these new threads popping up on the forum. I look forward to reading about more people's setups. It's probably more feasible for those with more disposable income, those who live in fairly quiet environments, etc.

I'm tantalizingly close, and I've not spent too much money on it either.  So far I've bought an $84 tablet which I will use no matter what, an $11 mic, a $6 trs to trrs adapter, a $12 tablet wall-mount and an $89 Windows 7 mini PC which was not strictly necessary.  I think I have maybe $12 wrapped up in Android apps.  This gets me, at any one time, as many as 2 open mics running on 2 instances of VC plus a copy of VoxWav; I thus have voice control over the lights in three rooms as well as two PCs, and the whole system is mirrored across both boxes so I really only need one PC up at any given time.  Voice command accuracy is acceptable (not perfect), even with the living room windows open (i live near a major freeway) and the TV on in the background.

The challenge at this stage is not voice recognition accuracy, but rather it's keeping everything up and connected over long stretches of time. I've got all the IP stuff pre-mapped inside my router and various things placed inside windows startup folders.  I'm using an app called Wakey to keep my Android tablets from going entirely to sleep. Still, the WO mic connection will drop spontaneously and usually fail to self-reconnect, and VoxWav is similarly touch-and-go though it's less of a problem because of the nature of the app.  I'm using Splashtop to remote administer the PCs so I don't have to get up to fiddle with them, but even so, I'm doing a awful lot of jiggering and juggling just to keep everything on and synced up.  I think a good use for VoxWav in my case is as a "voice interface of last resort", so to speak.  At some point maybe I can use it to run a script to trigger a sequence of events that starts and syncs things in a certain order when I need to do so, but I'm not there yet.

I also got round to testing some of my recording mics with VC to compare them with my $11 eBay mic.  The eBay mic is clearly better than the SM58 and EV 660 in my open-air application, and about on-par with the AKG c1000s. I've yet to try the AT2035 because I don't have an easy way to phantom power the mic without putting a mixer in-line.  I'll be staying with the eBay mic for voice recognition purposes, and I might even buy another before it's all over...


   

[keithj69]

I have finally started my journey into the whole home/open mic adventure.  My main reasoning is actually simple.  I am a computer geek by profession. I am curious as to what I can do with old used equipment and home automation. 
I rely on voxwav for 95% of the way I command my HA.  I am really happy with it, but if I know I can do it another way I am going to do it.  Or fail miserably trying.
Being just me in a small house, I am not worried about convenience of others.   I thought about the tablet/old android phone route, but having cats, leaving thin electrical wires out and about is not an option. 

With that said, my beyerdynamic MPC 22 should be waiting on my when i get home.  Also have a crown pzm-11 that i am not sure about. 
With this new sub-forum (thanks)  I will try to document my setup so others can decide if they want to even tackle things.

[rebelmaveric19]

I have been using my Samsung gear live with my Samsung galaxy S4 (google play edition), rooted and running autovoice with a xposed framework module and tasker that gives google now an api and simply just issue all voice commands by turning my wrist up. Completely hands free for the most part and works OK even in my truck with the windows down or at home with the movie turned up. It sometimes has its hiccups but 95% it works great.

[rebelmaveric19]

Also voxcommando responds back through my phone. Also nice that it is only one device to issue commands. Also use a Sony sbh52 Bluetooth device so I don't have to have the phone right next to me. Talk through the watch and get a response through the sbh52

[LilFonky]

Just in the way of a general update, I've been having more luck recently with the built-in mic on my wall-mounted Asus MeMO pad now that the system has had more time to train up. The setup goes like this: MeMO Pad built-in mic w/ WO Mic server installed in the bedroom -> HTPC in the living room running the WO Mic client using VoiceMeeter for mic mixing inside the box, which will allow me to run a second mic using a separate driver with separate input levels. I'm now settling on what that second mic should be. I'm thinking I may try the Mele F10 Pro, based on positive reviews posted elsewhere on this forum. Also, since VoiceMeeter allows for multiple outputs, I'm running one set of outs to the home theatre in the regular way, with a second set of outputs feeding a little hidden Bluetooth speaker so I can hear feedback sounds from VC even when the home theatre is powered down.  The whole thing can run headless using Splashtop installed on the tablet and HTPC.

The only problems I have with the way the system works are that it takes a lot of management to keep all the networked components connected and running. Also VoiceMeeter limits me to two mics and two sets of outputs and I'd really like to have a third mic for my studio.  The open mic problem has become mostly a non-problem so I'm feeling good about that aspect of things.  The tablet mic responds with good accuracy even from across the room.

All-in-all I feel like I've made pretty good progress tuning everything and I've not spent a lot of money on gear that I wouldn't have spent even without VC in the mix. I do plan to keep testing different devices, so I'll keep posting from time-to-time to let others know of my progress. Hopefully it will help others to get their own VC systems up and going.   

[Haddood]

Thank you that is really good feedback ... As far as I know in voice enter u can mix a third mic using the virtual line. Another option using steroids mix from windows ...

[LilFonky]

HI Haddood-

Thanks for the tips. I've been trying to sort out how to add a hardware mic to the virtual input in VoiceMeeter but I've not yet figured it out. I feel like it's doable, I'm just failing to put the pieces together somehow.

Also, I know nothing about "steroids mix" so if you've got any links to share I'd be happy to see them.

Cheers... 

[nime5ter]

Ha!

My guess is that Haddood might be the victim of auto-correct/auto-complete and that he meant to write "stereo mix"?

Awesome.

[jitterjames]

Steroids mix sounds like it would be highly effective in the short terms but that eventually it would start to exhibit some nasty side-effects...

[Haddood]

 
The genius apple auto correct...

Unless your PC is based on bio electronic I wouldn't suggest using steroid mix , and be careful with

Steroids mix sounds like it would be highly effective in the short terms but that eventually it would start to exhibit some nasty side-effects...

I experimented a bit with VoiceMeeter ... What you need is an audio repeater (VAC comes with one) that receive the signal from the mic and output to another device, in this case virtual input line installed by VoiceMeeter

[LilFonky]

ooooh STEREO AHAHAHA

I thought this might be a reference to the Steroids CLI and I was trying to piece together what this has to do with audio how it all might relate to VC. Dunno how I failed to connect the dots on that one...

I experimented a bit with VoiceMeeter ... What you need is an audio repeater (VAC comes with one) that receive the signal from the mic and output to another device, in this case virtual input line installed by VoiceMeeter

Wow, we're patching stuff all over the place now... 

BTW, anybody played with this thing at all?

http://www.codeproject.com/Articles/482735/TCP-Audio-Streamer-and-Player-VoIP

AFAIK there's no android client for this but it might be fun to experiment with. With an android/ios client this would be perfect for a whole-home, all in-the-box tcp mic mixing solution.  I seriously need to learn to program ...

[Kalle]

Interesting software, the server looks like a bit the TcpMic plugin from VC.

[Kalle]

Funny, VoxWav can connect to the Audio-Stream Server, but I can't see/hear any audio data