Moving to HS3 - Do I need a speech proxy?

[Monk]

Preparing to switch to HS3.
In HS2, I used the "Is Speaking" plugin which was dirt-simple.
I used it to set a HS device to power a relay before speech was allowed to go out to the speakers.

Is there something equally simple for HS3? I really don't need a complicated "feature-rich" speech proxy.

Thanks so much for any advice.

[Pete]

Why complicate this with a speaker proxy?

You do not need a speech proxy to do this.

Just concurrently send out an event to trigger a relay when HS is speaking or purchase an analog on demand low input speaker switch. (~$20 on Ebay and less than a 3rd party plugin)

My very first audio amplifier would auto switch speakers on demand from audio line inputs.

[Monk]

Excellent!
How does the event know when HS is speaking? Also - as specified, I used it to set the relay BEFORE speech was allowed to go out.
Very interested how to do this!

[Monk]

More information -
In reality, I use it to send 12V from my ELK to the Russound in order to trigger the page function.

[Pete]

So in your speech event whatever that is send an Elk command to turn on the 12VDC relay.

or

Personally I would just utilize an integrated line level sensor to do that.

Here is an example of one.

[ATTACH]61367[/ATTACH]

Way back I used to use the Audio Source AMP with the auto audio input to the AB8SS (still active today).

When I went to the CAA series of Russound amps dedicated one source to HS Speech and then used an auto line switcher like the above for speech.

I also put a mini serial server on the Russound serial port for OmniPro 2 and HomeSeer control.

[Monk]

Thanks. Now I am remembering why it's setup the way it is.
The Russound has the sensing ability built in, however, it is too slow to kick in and was chopping of half a word. Sometimes reliable and sometimes not.
That's why I went with the "Is Speaking" because it made it easy to know when speech was going to occur and to delay it slightly until I could supply the 12v to the paging input. It was the easiest and most reliable thing to do at the time, needed no extra events associated except the two and worked flawlessly.

[Pete]

You do not need a proxy to do that....IE: here put delays in sometimes relating to speech and wave files either via GUI triggers or in scripts or you can use a flag whenever speaking. The "flag" above is the speech proxy processing.

The proxy creates more work for Homeseer 3 and it is a slower response time thing.

Speaking Event
1 - Output 007 (12VDC to Russound)
2 - Wait 1,2,3 seconds
3 - Speak the time is?

One second might be all you need depending on how fast you can talk to the Elk M1.

Speaker dot exe is fastest running locally on Homeseer. Running remotely on another Windows PC there is a small delay due to the network. Largest delay is when you speak to a HSTouch client rather than use the speaker dot exe because it saves a wave file of the speech then transfers it over to the client. Here relating to the OmniPro 2 I use door beeps when opening and closing doors. Fastest beeps are on the serially connected OmniPro 2 keypad or Omnitouch consoles. Slowest are via the network connected Omnitouch consoles (Omnitouch 5.7 verus Omnitouch 5.7e screens).

Here made up an all in one two garage door script with flags and such to manage garage door open or close speech and physical triggers. I also utilize the old X10 palm pads (or you can use HSTouch screens to trigger speech events like status, time, weather, news et al.

[morrisdlx]

[Pete]

Welcome to the Homeseer automation Forum morrisdlx.

Lots of information and friendly folks here relating to what it is you want to do with Homeseer audio.

Here over the years have remained in the Microsoft SAPI speech mode using the original Homeseer speaker client integration to my Russound whole house audio. Homeseer clients have always been using MS SAPI / speaker dot exe (with different voice fonts sometimes on different clients). Today running HS3 Pro on Ubuntu Linux and Windows server as an Oracle Virtual box just for Microsoft SAPI. I run 4 speaker dot exe's connection to the main mother ship and the Zee Lite HS box (also running Linux today).

Using events or triggers based on my Leviton Omni Pro 2 combo panel I used same said text to speech events for outdoor perimeter, driveway, garage door, types of events.

I also use old X10 palm pads to trigger speech events like weather or news sometimes.

A variety of physical and virtual triggers and events trigger speech, dings, CCTV, text and email messages. Today you can send that audio clip to your smartphone if you want. Over the last couple of years have attached a NOAA radio output to trigger voice alerts generated by a weather program called WeatherXML.

Over the years with the introduction of Android / iOS clients the Homeseer client changed to play back a speech / audio sourced at the mothership as a wave file sent to the client. The speaker application remained on the wintel HS3 mother ship.

Next came / concurrently integration came to a variety of audio clients / virtual audio clients et al in the form of using a speaker proxy which would catch the text to speech audio and proxy it. There are a few speech proxies available today. One highly regarded speech proxy is Blade's which is mentioned below.

In the last couple of years now we have integration with the Amazon Alexa, Cortana (and Kinect), Google, Siri, et al which you can also now integrate with Homeseer providing you with some great voice recognition, automation control, internet information, et al.

In recap you have many choices relating to what it is you want to do relating to Homeseer speech and automation integration.

Give the options a try; see what fits for you. It all works very well these days.

[freeon]

Pete

When you say you can send an audio clip to your smart phone do you mean via HSTouch? Or is there a way to have SIRI speak the message? If this is possible how do I do this?

[Pete]

Yes

Way long time ago when Rich was integrating HSTouch to Android and iOS clients and Windows HSTouch clients he made it such that the mothership's HSTouch server saved the hsspeaker sound in a wave file and moved it over to whatever client where as the client played back the audio. Do a search for the temporary audio file on every client running HSTouch. It is there.

This doesn't have anything to do with any cloud based text to speech. It turned the Windows, iOS and Android client to an audio proxy of sorts.

The question has come up now relating to creating an RPi speech client and looking at the original HSTouch server / client API if possible.

Personally I would rather source the SAPI TTS on the Windows HSTouch clients using speaker dot exe. You can then have the SAPI clients speak in any available speech font or language. When you proxy the speech from the mothership to the clients you only have the choice of what SAPI speech fonts you are using on the mothership. If using Linux then it is whatever is running on the mothership relating to speech. (machine sounding text to speech).