Speech Recognition Not Working

[jhaney143]

"VR and use of a microphone on Android would be a separate application talking to Homeseer which I have no clue of."

Now I'm really confused. Using HSTouch Designer I choose the "Homeseer: Start listening for command" event for a button. I then push that out to my client(s). It brings up the microphone on the client when pressed and starts listening. There are no external programs and I don't expect that any real VR takes place with the exception of what I have added in HS3 already which does work if I speak into the microphone on my NUC. So for my very simple test I have an event that listens for the word 'test'. If I say it in front of my NUC it catches my speaking and fires off the event that says 'it worked' which I hear on all my clients. I used to be able to do this exact same scenario from my Samsung phone but now it does not.

The same event is also on the sample Android-Landscape project so not sure how it is not supported on Android devices? What is supposed to happen when it fires off on the clients?


Or maybe our verbiage is different on how things are laid out and how I'm using them.

I run HSTouch Designer on my Windows 10 laptop.

I run my HS3 server on a Windows 10 server NUC box.

I run my Speaker client on the same Windows 10 NUC Box as the HS3 server is on.

I run an Android client on a Samsung Galaxy S8.

[jhaney143]

And I know I'm not completely crazy because this thread is exactly what I'm after and worked a few years back: https://forums.homeseer.com/showthread.php?t=174479

[Pete]

Ahh....

@John,

I apologize as I did not read the post above at the time in 2015.

The poll was only related to Android and iOS and I was then using Windows and it wasn't included which is kind of weird as the first TTS/VR that Homeseer used in the 1990's was Microsoft SAPI.

Yes quit playing with using an Android touch client after rooting Android and trying to edit kernel pieces out many years ago.

I have done similar with smartphones always purchasing new unlocked phones and not from cellular telco operator and stripping the telco and android fluff from the cell phone OS's.

Today have 5 SIM accounts and do an a la carte phone of the day use and then shut them off when home except for one SIM that is used as a back up telephone and internet device such that it is always on but never goes anywhere in my home or outside. Today I am not tethered to my smart phones and only do one way texts if I text at all.

Before current telco used AT&T multiple grandfathered accounts from the 1990's (unlimited everything) until I got slammed by AT&T from one month to another.

Since abandoning my efforts relating to using Android went to Microsoft only Homeseer Touch clients many years ago.

Not sure what revision level Android HSTouch is at and if it is still using google voice today.

I am sure that others here will mention the use of HStouch on Android and using Google voice for VR.

[jhaney143]

Thanks for the additional info.

So is there anyone out there reading this that can offer any other pointers? I know that Google and Amazon integration is the hot topic nowadays but my use case is to not use a cloud service.

[Pete]

So just another post relating here to first testing Linux speech on the RPi...

Thinking back here to when I first installed Homeseer 3 on Linux (RPi at the time) googled speech for the RPi over here ==>

RPi Text to Speech (Speech Synthesis)

I installed all of the speech software mentioned and tried one at a time with Homeseer.

It sound exactly like my first speech endeavors in the 1980's with the Commodore 64 and Amiga Computers later on.

I was not impressed.

None the less thinking that Google speech did utilize the cloud anyhow and it was in testing mode at the time and then Google shut it off way back (for testing).

This is where for Homeseer I went back to Microsoft SAPI. BUT it only worked in the Windows version of Homeseer 3.

And just a guesstimate here to see if it worked in an Oracle Windows virtual box on the same computer. The audio output of the Linux box combines the Linux audio to the VB audio such that I hear my SAPI voice fonts fine.

I have collected many SAPI voice fonts over the years.

One of the newer things with Homeseer 3 is to be able to run a plugin remotely to the mothership. The plugin can run in Windows or Linux ....IE:

On a remote Windows computer it would be:

plugin.exe server=mothership_IP

or in a remote Linux computer it would be:

mono plugin.exe server=mothership_IP

It was using these features that I tested a PipoX7 running Windows 10 with Homeseer Touch, Speaker (SAPI) and the Kinect plugin connected to a multitouch openframe monitor. It all worked fine plus I opened up Kodi on the same computer for live television streaming from my HDHomerun boxes. I saw no issues doing this on the in house home network.

Note all of the above only deals with using the Homeseer speaker client for text to speech and VR. Note this works with Windows 10 desktop, tablet but not with Windows 10 mobile. (MS has discontinued it but supports it still)

Now too you have support of the cloud application VR on mobile phones.

HS is building a new mobile application and asking for suggestions here and yesterday showed some sample pictures over here ==> HomeSeer Mobile Feature Requests

Personally here stuck with the Windows side of using SAPI. A few years ago also went to only utilizing VPN to access my home network. I did test this on my Windows one tabletop tablet connecting to my mobile phone hot spot and a hotel's wireless internet.

Today new voice fonts from the old text to speech companies are sold as cloud applications any how. And your choices are limited now to using Windows computer SAPI or cloud or a combo. IE: new MS SAPI Cortana is a cloud application much like the rest you see.

BTW when using your smart phone you connect to the telco's cloud the to the OS cloud (Google, iOS or Windows).
Telco's massage your stuff before sending it off to the OS cloud where the OS cloud massages your stuff some more.

[jhaney143]

My use case is no data turned on for provider. So all data for when the microphone comes up would only go through same network as HS server and Speaker client. I guess when it worked I thought that since the Android client brought up the microphone, using the Homeseer listen function, the voice captured would be funneled through the Android client to the Speaker client directly. Much the same as the .wav files are sent but reverse. But I guess that is not the case from what you are suggesting in last post?

[Pete]

Understood.

Did you try it back in 2015 and did it work for you with the Homeseer touch Android client you were using back then?

Personally I do not know as I never tried it.

I do not see anyone here posting a response relating to using the Homeseer Android client with VR.

I understand that you said you tried it when initially posted and it worked for you.

There have been multiple Wintel, Android and iOS client versions since 2015 (along with multiple Homeseer Designer versions).

There was never a Homeseer Touch client written for Windows Mobile 10. The old MS Mobile was CE based and worked fine with Homeseer Touch for CE.

And today it works fine with Windows Tablet version 10 (Homeseer Touch and speaker client).

The Text to speech always worked for these clients using the wav file copy and playback.

Personally I was not aware of a working voice recognition working with iOS, Wintel and Android Homeseer touch clients.

The only in house on LAN voice recognition that I am aware of was using windows speaker dot exe and Microsoft SAPI. That has worked since the beginning of Homeseer (1,2,3).

Amazon Echo, Google Home, iOS SIRI, Microsoft Cortana VR are all cloud based.

Microsoft SAPI has now been extended to the cloud. You can too locally load Microsoft Kinect so you have both local VR / gesturing. My testing footprint had a Kinect mounted on top of the testing in house multitouch monitor.

[jhaney143]

I didn't try in 2015 as I didn't have HS then. It worked just a few days ago for a few commands and also maybe a year ago when I first tested.

I guess I could go Windows IoT but then of course what I want to do gets limited and costs become bloated because of compatible hardware more than anything. I really just wanted to replace some light switches with inexpensive Android touchscreens so I could have all the functionality of HS as well including voice commands. This way I could do things like turn the light on as default screen, but then also activate voice for that room and do things like show security camera when doorbell rings or whatever other event I have. Kinda like a mini magic mirror for a light switch for lack of a better explanation. The hardware to do an Android touchscreen like this is extremely affordable nowadays. Oh well. Maybe HS is not for me in this case. But thanks for the effort on helping me. Much appreciated!

[Pete]

Many Homeseer 3 Homeseer touch users today are using Android tablets of all sizes and all prices and they work fine for using the text to speech functions.

If VR worked when you tested a few days ago and a year ago then it must be an enabled feature.

This is where other Homeseer Homeseer Touch on Android users would chime in relating to using the features of Android Homeseer Touch VR.

Do you see an option in the Homeseer for Android Touch client for VR?

Here today bringing up Homeseer Designer 3.0.49 (tried the updates and then downgraded back to 3.0.49). It is running on Windows Server and I access it via RDP. Will install the newest HS3 designer on my W7 laptop today to have a look see.

Did a new Android project. There are no TTS nor VR settings in the project defaults using 3.0.49 that I can see. But that is using 3.0.49 HS Designer.

Relating to doing a Windows, Android or iOS design here have never seen any TTS (or VR) options in Homeseer Designer (well up to 3.0.49).

The Homeseer touch clients whatever OS all show up as speaker clients here in HS3 Pro. That said there are no options in the Homeseer Windows Touch clients that relate to TTS or VR. That said though I am using an old Homeseer Windows client (v.36) cuz it worked best for me.

The Android OS settings do have a microphone / VR option and those have always been there in all versions of Android and Android tablets.

The settings in the Android OS have really nothing to do with the Homeseer Touch client.

Many or most of the Homeseer users today are using the Amazon Alexa for VR and remote VR of Homeseer functions today.

You can now also purchase a cheap Windows 10 tablet of all sizes.

Many of these come with a two partition boot such that you can boot in Android and Windows 10.

Here is an 8" el cheapo for $71.43 Windows 10 tablet.

Premium Top Performance NuVision 8-inch Full HD 1920 x 1200 IPS Touchscreen Tablet PC Intel 1.33GHz Atom Z3735F Quad-Core Processor 2GB RAM 32GB eMMC SSD Webcam WIFI Windows 10-Silver

[ATTACH]68603[/ATTACH]

With this device you could run MS SAPI TTS / VR, Homeseer Touch and Homeseer Speaker for Windows. (and the Kinect Plugin).

Here a few years ago tested the Pipo X7 micro computer running Windows 10 connected to a 15-17" multiple touch wide screen open frame monitor with a Kinect to work well with Homeseer Touch, Homeseer Speaker, Microsoft SAPI VR and TTS plus Kinect gesturing.

My 15 Homeseer touch Atom tabletops run SAPI fine for TTS and are a bit slow for VR.

Hardware wise these are very customized though. (BIOS and SSD plus other adds).



Best running OS's though is Windows XPe (Seabios), Android (EFI) or Linux (EFI).