My house HAS TO talk to me

yeah i was thinking about that sonos but it is apple and i do not like apple nor itunes .
i hear you on the foscam , the audio it outputs is less then terrible lol but the speaker-out with some 7€ Trust speakers is really nice though and would be least expensive option here. i was also considering putting a usb stick audio device into my vera and try some of this audio-libs from dd-wrt but i find myself in a lot of time-consuming stuff so i dont have time for this atm.prolly it overloads the cpu of the vera. i think personnally the way to go is to have a audio device being triggered over UPnP. alas there is none as of yet. the doorbell is in progress of being accepted by national laws but there allready seem to be some setup problems to overcome and from what it looks one can archieve the same with a scene-controller and a nice relais-switch connected to a zwave switch to open/trigger a allready installed intercom system. Z-wave is not really capable of sending audio/video over its network so that is out of question. must gonna be some standalone device that talks over wireless.

Sonos = Apple? I’ll never touch that. No wonder they cost so much money.


[quote=“S-F, post:22, topic:170890”]Sonos = Apple? I’ll never touch that. No wonder they cost so much money.


Sonos != Apple… Not sure where that came from? Great solution all forms of media supported on almost any library. It is NOT tied to itunes (does support it though). Does NOT support airplay, but you can connect an AirportExpress up to it if you want. Android and Windows are fully supported. Very nice quality and functionality. You get what you pay for here. Pretty much any streaming music source (Pandora, Spotify, LastFM etc… ) is supported.

EDIT: For those who are not developers… != means NOT Equal…

Sonos is not Apple.

OK, gotcha.

@S-F, did you ever figure out how to get your house to talk to you?

I’ve been working on a squeezebox based solution that I can share if anyone is interested. Keep in mind that you don’t actually have to buy squeezeboxes if you don’t want to, the software is open sourced. The other nice thing about my solution is that it doesn’t require using an external server. It is a Windows based solution, but could be replicated easily on Linux with festival or espeak instead of Microsoft SAPI.

I have one real squeezebox (SB Touch) and a bunch of DIY squeezeboxes (pogoplus + nice usb sound cards 96kHz/24bit). The DIY route cost me about $30 a squeezebox ($20 for the pogo, $4 for the sound card, $6 for a sandisk). The sound cards I use (ZR-RSSC) support both 5.1 analog and optical out. It was a little tricky getting the asound.conf right but I am happy to share mine.

My setup is I run Logitech Media Server running on a Windows PC. I then have a perl script that uses the Microsoft SAPI interface to create a wave file which I play on the squeezeboxes. After the file finishes playback the squeezeboxes resume playing the same playlist they were running before at the same position in the song. The latency between executing the command and hearing the speech is very short (between 1-2 seconds, most of that is probably running the TTS).

I only have one music playlist going at a time, so if you wanted to have squeezeboxes playing different songs play a TTS and then resume different songs it would need to be tweaked a bit.

I have been using Vera Alerts, although your solution sounds very interesting. Would you mind giving a little more information? I haven’t been following the Sonos/Squeezebox development as I never plan on owning one of these devices, but what you say sounds very interesting. You can play wave files as well as TTS? How are the commands sent?

Absolutely. So my script actually uses the Microsoft TTS engine to overwrite a wave file (speech.wav) which is already in the Logitech Media Server’s library. This file could easily be generated with espeak or festival, or it could be a pre-existing wave file.

You can play around with Squeezebox without having any additional hardware whatsoever, although I think some of the online streaming options might only work if you have Logitech hardware (or at least a login). But it works great for playing mp3/flac files you own.

There are several software players available: Squeezeplay (which has a Java based GUI and is my recommendation), and Squeezeslave, which is a basic command line client.

Right now I just have this set up as a perl script which takes a single argument from the command line (the text I want spoken). I execute this with EventGhost which is running on my Windows PC. I use my Vera eventghost plugin to generate events which can then trigger TTS.

There is absolutely no reason however why you couldn’t use a simple server that listened for strings to speak. This could be called from Vera directly. I’ll probably do something like that, just still thinking about the best way to implement it.

Interesting stuff.

From what I understand there is already some speech integration for Vera (using Google TTS) which works with the Squeezeboxes and Sonos plugins, plus I have seen some other posts using Festival too.

For me this is a great area of exploration, so please do share in detail what you eventually do, who knows a future Vera release might one day benefit from having an audio output of some description.

I personally have the Sonos set up and the work the team have done over there has been incredible, I’ve tried to contribute a little buy sharing some dynamic TTS messages I’ve built, so if you have chance check it out , as the goal for many of us is a workable speech solution, so if elements can be shared it benefits everyone.

I’m also not stuck on Sonos so will watch this post (or a dedicated one if you wanted to do something specific) with interest, the benefits for Sonos are obvious in some areas, the key one for me was not to have loads of different appliances running, hence it blended in nicely with my current set up.

After achieving speech, a seamless voice recognition would be the next nice to have. :wink:

This is all very interesting. In an ideal world I would have the ability to send TTS or play a .wav from a scene just ilke turning on a light. Sound could come from either my always on but logged off windows server or Vera directly.

Is anyone using HAL to facilitate voice communication and status with Vera?

That is my next goal! The beautiful thing is all the heavy lifting on this stuff has been done, you just need a bit of “glue” to get everything working togeather. I’ve figured some of that out as well, but I’ll post in a different thread once I get a bit further along and work out some kinks. I’m always happy to share what source code I have too.

When you say directly do you mean through speakers hooked up to the Windows server? Because I think that may be an easier problem. I’m jumping through some hoops to use squeezebox because I want sound through a half dozen or so different devices at once. For isntance, if you run EventGhost to run at system boot (it has a checkbox for that option, I haven’t tried it but I expect it works fine), you can set up triggers to run TTS or play wave files with its built in plugins. You can also you the Vera plugin to trigger these based on certain events like a specific device being turned on. It isn’t perfect as the Vera plugin is incomplete, but it might get you halfway there.

[quote=“OtisPreslsy, post:31, topic:170890”]Is anyone using HAL to facilitate voice communication and status with Vera?

I’m not using Hal, but working on something similar using my android as a microphone and the Microsoft SAPI to process speech.

I have used HAL on my desktop and the accuracy of my mic just isn’t enough to make it usable. I understand that the Android client for HAL is pretty good on a phone due to the built in noise canceling functionality phones inherently have.

@ radarengineer, could you elaborate a little on this eventghost/windows TTS and .wav playing you speak of? I have looked at eventghost and it seems to be mostly aimed at home theater functionality. Is there any active development on the eventghost plugin? I’m willing to experiment with this in the coming weeks if it’s under development and ready to go. Otherwise if I can get Vera Alerts to work I’ll be happy. RTS seems to be burning the midnight oil churning out great plugins and he seems to be pretty interested in getting Vera Alerts to a mature state in the very near future.

All of this development is really great but all of the options still seem a little bit like a workaround bandaid to Vera’s limitations. I’m pretty impressed with the efforts and accomplishments of the community but I wish it were easier for developers to make big changes to the way scene creation works.

That is exactly what i have and what you get with the Sonos plugin, i think of it as 'the mouth for Vera. ’

I use it in many ways, from an alert siren (so it plays siren.mp3) when an armed sensor is tripped. To remind me of things to do like ‘The back door is still open’ or to wake me up (as an alarm clock) with "Good Morning, today the high will be 5 celcius)

The TTS element is really great and as the string is just a dynamic URL that’s sent to Google, the logic behind it theoretically (if you break the plugin down and take the bits you need) could work for people running a small PC with a simple mp3 stream radio player.