www.homeseer.com    
 

Go Back   HomeSeer Message Board > HomeSeer Products & Services > HomeSeer Software > HomeSeer Voices

HomeSeer Voices Discussions related to the HomeSeer add-on voices and their use within HomeSeer.

Reply
 
Thread Tools Display Modes
  #41  
Old December 20th, 2017, 10:39 AM
Pete's Avatar
Pete Pete is offline
OverSeer
 
Join Date: Jan 2001
Location: House
Posts: 15,261
Yes it is around 9:30 PM over there.
__________________
- Pete

Automator

HS3 Pro & Lite Edition Beta 3.0.0.435

HS3 Wintel Touch | Ubuntu 16.04 64 bit | Oracle Windows Virtual Box ==> for Wintel only SAPI and HS3 plugins | Speech - Microsoft SAPI - Neospeech - Amazon Echo | Hardware | Haswell Intel iSeries 3 - 16Gb | Pine64 - 2Gb computers | Openpeak Intel Atom SoC tabletop touchscreens (15 HS tabletop tablets) | Touchscreens - Windows embedded POE connected |Light switches - X10,UPB, ZWave and Zigbee | Firewall - PFSense - 2 WAN plus 4 LAN interfaces | Network - Gb managed switches / POE WAP(s) | CCTV - Zoneminder IPHD cams - variety | Audio - Russound - AB8SS | Security - Leviton HAI Omni Pro 2 | Weather - Davis Vantage Vue - MeteoStick - WeeWx | 1-Wire - AAG, Midon and HB | OWFS - Mosquitto - Node Red - Python - RPi Stretch - OpenWRT
Reply With Quote
  #42  
Old December 20th, 2017, 10:48 AM
happnatious1's Avatar
happnatious1 happnatious1 is offline
Seer Master
 
Join Date: Mar 2004
Location: Southern Maryland
Posts: 1,297
Quote:
Originally Posted by abwyatt View Post
Nice! Is that a cloud solution or offline?
Off-line and fast. Its not going out to my HSTouch clients though. Not sure how homeseer handles this part.
Reply With Quote
  #43  
Old December 20th, 2017, 11:39 AM
abwyatt's Avatar
abwyatt abwyatt is offline
Seer Deluxe
 
Join Date: May 2016
Location: Ho Chi Minh City, Vietnam
Posts: 206
Quote:
Originally Posted by happnatious1 View Post
Off-line and fast. Its not going out to my HSTouch clients though. Not sure how homeseer handles this part.
Hmmm ... I just reread this thread and it seems someone else had come to a deadend where you are now. See post No.12 and prior posts.

Ok, need to get shuteye now.
Reply With Quote
  #44  
Old December 20th, 2017, 12:04 PM
lveatch lveatch is offline
Seer Deluxe
 
Join Date: Dec 2012
Location: USA - Illinois
Posts: 179
Quote:
Originally Posted by happnatious1 View Post
Off-line and fast. Its not going out to my HSTouch clients though. Not sure how homeseer handles this part.
You might need to change speak_to_file.sh as well.
__________________
Len


HomeSeer Version: HS3 Pro Edition 3.0.0.435
Linux version: Linux homeseer Ubuntu 16.04 x86_64
Number of Devices: 633
Number of Events: 773

Enabled Plug-Ins
2.0.54.0: BLBackup
2.0.40.0: BLLAN
3.0.0.48: EasyTrigger
30.0.0.36: RFXCOM
3.0.6.2: SDJ-Health
3.0.0.87: weatherXML
3.0.1.190: Z-Wave
Reply With Quote
  #45  
Old December 20th, 2017, 11:15 PM
happnatious1's Avatar
happnatious1 happnatious1 is offline
Seer Master
 
Join Date: Mar 2004
Location: Southern Maryland
Posts: 1,297
Quote:
Originally Posted by lveatch View Post
You might need to change speak_to_file.sh as well.
I modified the speak_to_file.sh and I was able to stream to a HSTouch screen, however she will only complain that she is not licensed. Cant get a response from Cepstral on how to purchase a license or how much it costs. Odd that a software company isn't interested in selling software.
Reply With Quote
  #46  
Old December 21st, 2017, 12:33 AM
happnatious1's Avatar
happnatious1 happnatious1 is offline
Seer Master
 
Join Date: Mar 2004
Location: Southern Maryland
Posts: 1,297
Giving up on Cepstral I did a little more searching and found festival.

https://ubuntuforums.org/showthread.php?t=677277

echo "$1" | festival --tts in speak.sh gets her talking.

text2wave "$1" -o "$2" in speak_to_file.sh should work but it isn't. something must be wrong with my syntax
Reply With Quote
  #47  
Old December 21st, 2017, 12:38 AM
lveatch lveatch is offline
Seer Deluxe
 
Join Date: Dec 2012
Location: USA - Illinois
Posts: 179
Quote:
Originally Posted by happnatious1 View Post
I modified the speak_to_file.sh and I was able to stream to a HSTouch screen, however she will only complain that she is not licensed. Cant get a response from Cepstral on how to purchase a license or how much it costs. Odd that a software company isn't interested in selling software.
Create an account with them, once you get an email with approval, set your initial password and the linux for personal use should be available in the store for purchase. Prices appear to be the same as for windows.
Reply With Quote
  #48  
Old December 21st, 2017, 08:56 AM
happnatious1's Avatar
happnatious1 happnatious1 is offline
Seer Master
 
Join Date: Mar 2004
Location: Southern Maryland
Posts: 1,297
The quality of the flite 2.4 version of US English Female CLB sounds just as good as the Cepstral Amy voice IMHO.

Flite Demo: http://tts.speech.cs.cmu.edu:8082/

Cepstral Demo: https://www.cepstral.com/en/personal/linux
Reply With Quote
  #49  
Old December 21st, 2017, 12:54 PM
abwyatt's Avatar
abwyatt abwyatt is offline
Seer Deluxe
 
Join Date: May 2016
Location: Ho Chi Minh City, Vietnam
Posts: 206
Quote:
Originally Posted by happnatious1 View Post
I modified the speak_to_file.sh and I was able to stream to a HSTouch screen,
What was the syntax that you used for the speak_to_file? I'd like to try it when I get home.
Reply With Quote
  #50  
Old December 21st, 2017, 09:21 PM
happnatious1's Avatar
happnatious1 happnatious1 is offline
Seer Master
 
Join Date: Mar 2004
Location: Southern Maryland
Posts: 1,297
Quote:
Originally Posted by abwyatt View Post
What was the syntax that you used for the speak_to_file? I'd like to try it when I get home.
I think for cepstral I was using: padsp /opt/swift/bin/swift -n Amy "$1" "$2"

I'm still trying to get festival working with HStouch clients.
Reply With Quote
  #51  
Old December 21st, 2017, 09:30 PM
Pete's Avatar
Pete Pete is offline
OverSeer
 
Join Date: Jan 2001
Location: House
Posts: 15,261
Curious how much the Cepstral voice fonts are?

The Cepstral voice fonts sound almost as good as the SAPI voice fonts.

Here still using the old AT&T and Neospeech voice fonts offered by Homeseer during the HS1-2 days. (and more purchased / collected over the years).
Reply With Quote
  #52  
Old December 23rd, 2017, 03:54 AM
abwyatt's Avatar
abwyatt abwyatt is offline
Seer Deluxe
 
Join Date: May 2016
Location: Ho Chi Minh City, Vietnam
Posts: 206
OK, I'm back and ready to continue on my MS SAPI (mis)adventure on an Oracle WinXP VB which Pete has been helping with.

Continuing on from where I was previously with not being able to launch speaker.exe to run (Post #40), I've installed .Net4 which was required. Then installed HS3 for Windows to get some updated dependent dll files for HSCF and HomeSeerAPI.

After updating the dlls, I was able to copy the set of speaker files into seperate folders to run multiple instances of speaker.exe.

Anyway, back to basics with just one speaker.exe running. In the speaker options, I've pointed the speaker to the IP address of the Linux HS3 on the same box. It connects. Yes!

With speaker.exe connected, a speaker client called HS3SPEAKER.DEFAULT appears under Tools/Setup/Network on my Linux HS3. I also connect my mobile HSTouch and a speaker client for the mobile HSTouch also appears. See screengrab.

So far so good.

Now to test TTS from events. I set up an event to output TTS to both of the above client speakers, the HSTouch client speaker and the speaker on WinXP VB. See screengrab.

I get speech output on the WinXP VB. See screengrab.

BUT I don't get any speech output on my mobile HSTouch. Just to test if the HSTouch is OK, I enable flite speech in speech-to-file, and I get flite TTS output on the mobile HSTouch. So, basics are working as it should.

So, how do I get TTS from the MS SAPI/speaker.exe out to the mobile HSTouch speaker client?

Conceptually, this is what baffles me. I understand how TTS on the Linux HS3 gets out to the speaker client on the WinXP VB, but I do not understand how TTS goes out to the mobile HSTouch speaker client with the intervention of the SAPI/Speaker.exe application on the WinXp VB. Does something need to be setup in the Linux speech-to-file?

Andrew
Attached Images
   

Last edited by abwyatt; December 23rd, 2017 at 05:48 AM.
Reply With Quote
  #53  
Old December 23rd, 2017, 10:32 AM
Pete's Avatar
Pete Pete is offline
OverSeer
 
Join Date: Jan 2001
Location: House
Posts: 15,261
Speaker dot exe is only for Windows Clients and did work fine with CE and Windows Mobile up until around version 6.

HSTouch clients like Android and iOS and Windows running by themselves with no sapi or speaker dot exe use a wave file for sound.

IE: The homeseer mothership creates a TTS (or combo) wave file on the mothership and sends it out the the client for playback. It is not SAPI. It is just a playback of the speech and or audio from the mothership. The voice font will be identical to the source on the mothership.

BUT I don't get any speech output on my mobile HSTouch. Just to test if the HSTouch is OK, I enable flite speech in speech-to-file, and I get flite TTS output on the mobile HSTouch. So, basics are working as it should.

So, how do I get TTS from the MS SAPI/speaker.exe out to the mobile HSTouch speaker client?

You can only get the wave file which is a recording of the speech from the mothership.

MS SAPI does not work or is present on iOS or Android mobile clients (or Linux mobile clients or Windows 10 mobile clients). Today there is no Windows 10 Homeseer client.

Test it. Do not use the command line stuff. Send a sentence over to your client and play back the wave file it creates on your client is the only way to test. You should hear the wave file when it is speaking. If TTS is being generated by SAPI then that voice is what you should hear in the wave file.

I can do speaker dot exe on a Windows 10 tablet connected directly to Homeseer locally or via the WAN. Via the WAN it is a VPN tunnel.

IE: Speaker dot exe is only a SAPI program. If you configure your speaker dot exe in the VB as your primary TTS then that voice will be the one you hear on your mobile clients (well not true TTS rather just a recording). No need to play around with the Linux TTS commands.

All of the HS3 3rd party speech clients are basically speech proxies that utilize the wave files and are really not Microsoft SAPI based.

What works remotely for me is a Microsoft tablet using VPN tethered to my windows mobile 10 LTE connection. It is too bad that Microsoft is not pursuing the mobile market anymore. While I have multiple phones here (iOS and Android) prefer to use my MS Mobile phones. I do a smart phone of the day connection with my sim cards and unlocked phones but am not tethered to any particular phone today. (all smart phones are purchased unlocked and I cook (build) them cooking them without the included telco junk - have done this for years now).

BTW - 1 my default speaker dot exe settings have the microphone shut off and all of the extra fluff off and basically you do not see the VB running unless you want to cuz it just works. I load the VB with speaker dot exe before Homeseer loads so that Homeseer sees the speaker program running before loading up internal tts.

BTW -2 HS3 TTS in Linux is just a port over from HS3 running in Mono to SAPI to Wave files. There is no SAPI for Linux. I have tried a couple of MS SAPI programs running in Linux and they are difficult to set up and really not worth the effort. Getting Wine to do SAPI was a PITA and had to throw the kitchen sink to Wine which was also a PITA. I was posting the official methodology on the Wine Wiki while I got it to work it wasn't worth the effort.

The best speech is from Microsoft as they have been developing speech since second version of Windows. And you can have multiple voice fonts in multiple languages for the ultimate in TTS and VR. IE: If you run with Spud's Kinect plugin you can turn on and off Amazon or Google voices. Windows 10 SAPI is the best so far and is quick learning.

Have a read here.

Microsoft Speech API

Note too that speech worked fine with my Commodore C64 / Amiga computers except that it was very robotic sounding in the 1980's.

Last edited by Pete; December 23rd, 2017 at 11:09 AM.
Reply With Quote
  #54  
Old December 23rd, 2017, 11:04 AM
abwyatt's Avatar
abwyatt abwyatt is offline
Seer Deluxe
 
Join Date: May 2016
Location: Ho Chi Minh City, Vietnam
Posts: 206
Quote:
Originally Posted by Pete View Post
Speaker dot exe is only for Windows Clients and did work fine with CE and Windows Mobile up until around version 6.

HSTouch clients like Android and iOS and Windows running by themselves with no sapi or speaker dot exe use a wave file for sound.

IE: The homeseer mothership creates a TTS (or combo) wave file on the mothership and sends it out the the client for playback. It is not SAPI. It is just a playback of the speech and or audio from the mothership. The voice font will be identical to the source on the mothership.

BUT I don't get any speech output on my mobile HSTouch. Just to test if the HSTouch is OK, I enable flite speech in speech-to-file, and I get flite TTS output on the mobile HSTouch. So, basics are working as it should.

So, how do I get TTS from the MS SAPI/speaker.exe out to the mobile HSTouch speaker client?

You can only get the wave file which is a recording of the speech from the mothership.

MS SAPI does not work or is present on iOS or Android mobile clients (or Linux mobile clients or Windows 10 mobile clients). Today there is no Windows 10 Homeseer client.

Test it. Do not use the command line stuff. Send a sentence over to your client and play back the wave file it creates on your client is the only way to test. You should hear the wave file when it is speaking. If TTS is being generated by SAPI then that voice is what you should hear in the wave file.

I can do speaker dot exe on a Windows 10 tablet connected directly to Homeseer locally or via the WAN. Via the WAN it is a VPN tunnel.

IE: Speaker dot exe is only a SAPI program. If you configure your speaker dot exe in the VB as your primary TTS then that voice will be the one you hear on your mobile clients (well not true TTS rather just a recording). No need to play around with the Linux TTS commands.

All of the HS3 3rd party speech clients are basically speech proxies that utilize the wave files and are really not Microsoft SAPI based.
I do understand how the Linux HS3 TTS works using wave files and how speak_to_file controls which TTS application is being used - ie. flite, picowave2, etc. Been using and fiddling around with that for quite some time now. I was also using Voice RSS at one time.

And I also understand that the Linux HS3 TTS system using wave files is completely different to the way MS SAPI works. So in using MS SAPI, we ignore the Linux TTS system altogether, right? No need to discuss it.

So what I understand now is that there is no way to use MS SAPI with HStouch on Android, OR Google Chromecasts such as Google Home which was my original question at Post #29 about whether I would be able to output an event generated TTS to Google Home using MS SAPI.

I'll look into Cepstral for my TTS solution ...

All is not lost with the WinXP VB. I had been thinking of another use for it and that is to run HS3 with Windows only plugins.
Reply With Quote
  #55  
Old December 23rd, 2017, 11:17 AM
Pete's Avatar
Pete Pete is offline
OverSeer
 
Join Date: Jan 2001
Location: House
Posts: 15,261
And I also understand that the Linux HS3 TTS system using wave files is completely different to the way MS SAPI works. So in using MS SAPI, we ignore the Linux TTS system altogether, right?

yes

So what I understand now is that there is no way to use MS SAPI with HStouch on Android, OR Google Chromecasts such as Google Home which was my original question at Post #29 about whether I would be able to output an event generated TTS to Google Home using MS SAPI.

Correct except for using MS SAPI to generate speech wave files such that the HSTouch clients and or all of the 3rd party speech clients work. This is the best because you can utilize different speech fonts (in just about any language today). I was bugging my wife for a while having one touchscreen SAPI speak in French, German, Portuguese, Italian, Spanish, et al..

IE: on reboot here have the speaker clients / VB loading before HS3 loads such that the primary speech is SAPI. (and wave files). If I load HS3 via terminal console on the Linux box then HS3 loads both the Linux and SAPI TTS so you hear both when speaking. If I autoboot via cron, rc.local or systemd and delay the boot of HS3 after the VB loads with speaker dot exe then the speech is only the SAPI and not the Linux TTS. I noticed this just recently as I never really touch the console directly and do everything remotely via SSH.

IE: if I look at the HS3 GUI connected speaker clients I only see the VB running...

Connected Speaker Clients
Speaker HS3W2003VB DEFAULT

I'll look into Cepstral for my TTS solution ...

Cepstral is not using SAPI such that I do not know that HS3 can talk to a proprietary Linux created speech font. IE: that is why flite works but not Cepstral with HS3. All of the Linux based TTS utilize different methodologies for speech. There is not one standard.

Here I am not tethered to my cell phone and primary remote to HS3 is via in house wired touch screens or remotely via VPN (only). Actually not tethered to the internet for my automation here. Never have been. I do tinker a lot with it though.

BTW Windows embedded clients (early on) do not need the internet to work. Android is always connected to the internet and I have had issues cooking up an Android OS that doesn't connect to the internet; so I do not use it for my HSTouch clients. All of today's tablet OS's primary need an internet connection to work. Been playing here with embedded W8 and it needs the internet to work right today.

I cooked up a special build of W10 removing all of the internet connections and the updates kept bringing back the internet connectivity. Today W10 is just a kiosk terminal to the internet just like every mobile OS is (my opinion).

BTW just noticed maybe a bug with Beta V.398. I do not see my 15 HStouch clients list and only one speaker.exe client today. I used to see all of the HSTouch and speaker dot exe clients in a list on the bottom of the network tab.

One thing too is that if you run speaker dot exe or use HSTouch on various client and do TTS from the mothership none of the speech will be sync'd.

Much like any audio distribution via a LAN. (Sonos is another example) such that audio here is distributed via Russound and wired speakers.

Last edited by Pete; December 23rd, 2017 at 12:45 PM.
Reply With Quote
  #56  
Old December 23rd, 2017, 12:51 PM
abwyatt's Avatar
abwyatt abwyatt is offline
Seer Deluxe
 
Join Date: May 2016
Location: Ho Chi Minh City, Vietnam
Posts: 206
Quote:
Originally Posted by Pete View Post
Correct except for using MS SAPI to generate speech wave files such that the HSTouch clients and or all of the 3rd party speech clients work.
But how can MS SAPI generate wave files on the Linux HS3 (and then send it to the speaker clients) from an event such as in the screengrab? It cannot as far as I can see.

While testing, I am starting WinXP VB and Speaker.exe first and then starting up HS3 after. But doesn't seem to make any difference. I have been checking in the Linux HS3 wave folder, and there are no wav files being generated by SAPI. Only from flite.
Attached Images
 

Last edited by abwyatt; December 23rd, 2017 at 01:33 PM.
Reply With Quote
  #57  
Old December 23rd, 2017, 02:22 PM
Pete's Avatar
Pete Pete is offline
OverSeer
 
Join Date: Jan 2001
Location: House
Posts: 15,261
But how can MS SAPI generate wave files on the Linux HS3 (and then send it to the speaker clients) from an event such as in the screengrab? It cannot as far as I can see.


Test it by autobooting your Linux box and running the Wintel VB with SAPI on it first and then Homeseer second.

I take it all back as I listened to the temporary wave file generated on one Android client and the wave file is robotic sounding and not SAPI testing it on V.398 of HS3.

I apologize Andrew as here only use Homeseer Touch on Wintel today and do not use (but play) with Android for my tablets and phones.

Last edited by Pete; December 23rd, 2017 at 02:35 PM.
Reply With Quote
  #58  
Old December 24th, 2017, 12:59 AM
Pete's Avatar
Pete Pete is offline
OverSeer
 
Join Date: Jan 2001
Location: House
Posts: 15,261
Microsoft has been slowly decimating itself and honestly I do not think the Bill Gates cares much anymore.

They own Mono now and they should write a SAPI client that runs in Mono for Linux.

It would be one nice gesture (Mr. Satya Nadella) for the Gipper requested by the Rock.

That said though only concern with current Microsoft CEO Satya Nadella is to keep the ship afloat.

Last edited by Pete; December 24th, 2017 at 01:35 AM.
Reply With Quote
  #59  
Old December 31st, 2017, 01:02 PM
happnatious1's Avatar
happnatious1 happnatious1 is offline
Seer Master
 
Join Date: Mar 2004
Location: Southern Maryland
Posts: 1,297
Quote:
Originally Posted by happnatious1 View Post
Giving up on Cepstral I did a little more searching and found festival.

https://ubuntuforums.org/showthread.php?t=677277

echo "$1" | festival --tts in speak.sh gets her talking.

text2wave "$1" -o "$2" in speak_to_file.sh should work but it isn't. something must be wrong with my syntax
Using the command flite -lv will tell you the names of the voices available on your system, for instance I have slt, and then putting

flite -voice slt -o "$1" -t "$2"

in your speak_to_file.sh will use that voice for streaming to your hs touch clients.

I still cant find a tutorial that adds voices to flite. I've added the clb voice to festival but not flite.
Reply With Quote
Reply

Bookmarks

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is On

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Linux TTS - Voices petez69 HS3-Pi Discussion 11 Yesterday 06:35 AM
Current Linux build? New beta Z-Wave Linux 216 doesn't load AbMagFAb HS3/PRO Beta Releases 5 April 4th, 2015 11:40 PM
Static voices when switching voices. stevene HomeSeer Voices 0 July 22nd, 2007 08:16 PM
NeoSpeech and Asian voices / Fun voices GenevaDude HomeSeer Voices 5 December 29th, 2003 08:16 AM


All times are GMT -4. The time now is 05:15 AM.


Copyright HomeSeer Technologies, LLC