Can I change TTS voice selection from HS3?

[jetkit]

I know I can go to the Speaker and change voices there, but can I do that from HS3?

Only solution so far is to have multiple speaker instances.

[ITguyHS]

I know I can go to the Speaker and change voices there, but can I do that from HS3?

Only solution so far is to have multiple speaker instances.

If you just want a temporary change from within an event, you can use the "voice" tag within the HS Speak action. Example:

<voice required="name = Microsoft Sam"> This is Home sear speaking

[jetkit]

Thank you!

[Kitar]

MS Voices

HS3 Pro W10 Home

I am new to Windows 10 and I can't find any mention of Microsoft Sam. I can see Microsoft Zira Mobile, Laura Mobile, Julie Mobile, Ayumi Mobile and several other "Mobile" voices but nothing like Sam.

So in my testing I discovered what must be a bit of a bug.
Tested with the following line in the speech field of the event editor:

Test1
<voice required="name = Microsoft Zira Mobile"> this is a test.
This resulted with no speech output at all.

Test2
<voice required="name = Microsoft Sam"> this is a test.
This results in HS reciting a different text strings that I had deleted.

Test3
this is a test.
Results in proper text to speech in a very good voice.

Edit:
Further research shows Sam as an XP voice. Different Win versions have different voices installed. Still no idea how I can use alternate voices in an event or even find out what voices I have that are usable.

[ITguyHS]

Still no idea how I can use alternate voices in an event or even find out what voices I have that are usable.

If you go to the Windows 10 "Time and Language Settings" and look under the Text-to-speech section, the voices you see there are what is available on your system. I think the basic W10 install only has Mira and Mark - at least that's all I see on my W10 that upgraded from 8.1. My HS3 system is still on an XP system, so I can't test to see if the example works on W10.

[Kitar]

Windows 10 TTS voices = David Zira and Eva

If you go to the Windows 10 "Time and Language Settings" and look under the Text-to-speech section, the voices you see there are what is available on your system. I think the basic W10 install only has Mira and Mark - at least that's all I see on my W10 that upgraded from 8.1. My HS3 system is still on an XP system, so I can't test to see if the example works on W10.

This is a wee bit off topic but I will add my 2 cents here because I found this thread looking for TTS information.

Thanks for the help. I found that there are 2 English desktop voices installed in Win 10 Home, Microsoft David Desktop and Microsoft Zira Desktop. I checked the MS store and did not see any additional voices for sale. This machine has Win 10 pre-installed, not upgraded.

So typing the <voice required="name = Microsoft Zira Desktop"> This is Home sear speaking - results in a correct TTS voice change from David.

In my search I found an online registry tweak that reveals the MS Eva voice. Will try when time permits.

Interesting that under the Time and Language Settings all I see is the Mobile voices and in my tests they do not play through HS. The way I revealed the Desktop voices was in the config of the Homeseer Speaker app under the Options tab. In Win 10 you can also enter "text to speech" in the search box. Pressing enter pops up the menu.

[MNB]

Dumb question, I understand how to change the Win 10 TTS voice and I understand how to make HS3 speak events change the voice but where does HS3 store the default voice as it seems to be always using "Microsoft Dave" thanxs in advance?

Update: I found it! Its under the HomeSeer Speaker app (helps if you read the post above), it also show the current voices in my case they are; VoiceCB which sounds like "Microsoft Dave:, Microsoft David Desktop", and Microsoft Zira Desktop. Which I find interesting as the Win 10 TTS voices are the same except instead of "Desktop" it is "Mobile". Not sure why?

[Pete]

The TTS/VR Windows 10 voice fonts are Microsoft SAPI voice fonts which Homeseer has utilized since the beginning of Homeseer in the 1990's.

A little historical on SAPI and before SAPI

Dr. James Baker laid out the description of a speech understanding system called DRAGON in 1975. In 1982 he and Dr. Janet Baker, his wife, founded Dragon Systems to release products centered around their voice recognition prototype. He was President of the company and she was CEO.

DragonDictate was first released for DOS, and utilized hidden Markov models, a probabilistic method for temporal pattern recognition. At the time, the hardware was not powerful enough to address the problem of word segmentation, and DragonDictate was unable to determine the boundaries of words during continuous speech input. Users were forced to enunciate one word at a time, each clearly separated by a small pause. DragonDictate was based on a trigram model, and is known as a discrete utterance speech recognition engine.

SAPI 1
The first version of SAPI was released in 1995, and was supported on Windows 95 and Windows NT 3.51. This version included low-level Direct Speech Recognition and Direct Text To Speech APIs which applications could use to directly control engines, as well as simplified 'higher-level' Voice Command and Voice Talk APIs.

These companies are so good today that they can create a voice font that sounds just like you with some training.

Early on Homeseer sold AT&T and Neospeech SAPI voice fonts. Today you can purchase individual SAPI voice fonts.

Googling some found AT&T voice fonts. (they were much more reasonable when HS sold them).

AT&T SDK kit

Here running SAPI in a VB on the Homeseer Ubuntu box. I have collected SAPI 32 and 64 bit voice fonts since the beginning of Homeseer.

I am currently running 4 instances of speaker dot exe on the VB and can use the voice fonts you see.

Note too that Windows 10 SAPI is much improved over older versions of SAPI. Integration with Cortana and the cloud and Kinect and a variety of MS voices will provide an ultimate AI / cloud connection sometime soon.

Note too that Homeseer uses MS SAPI voices. Homeseer today does not sell 3rd party SAPI voices today such that you have to purchase they a la carte these days. Googling last price for Heather (very popular voice font) was $246 which is a bit extreme. Note too that some voice fonts sound very nice and natural and some are not so nice and natural.

Have a read here:

Windows 7 Voices

[ATTACH]61826[/ATTACH]

[tahl]

I have the voice Mark loaded; how do I get access to it?

I can get Microsoft David and Microsoft Zira ... but I don't see an option to get Mark even though he's on the computer.

Thanks!

[Pierre]

In Windows 10, you will get the SAPI voice with :
C:\Windows\SysWOW64\Speech\SpeechUX\sapi.cpl

[tahl]

Thanks.

I learned something new as well -- files with the cpl extension.

Do you know if other voices can be loaded that HS will be able to use? I used Cepstral voices in the past....

Best,
Tahl

[randy]

.cpl extensions are for Control Panel actions.

Windows 10 uses 64-bit voices by default. Homeseer uses 32-bit voices. When you go to Windows settings for speech voices, you will only see the 64-bit controls like this:



For HomeSeer voices you need to launch the 32-bit control panel with a shortcut to this path: "C:\Windows\SysWOW64\Speech\SpeechUX\sapi.cpl". That will get you this panel:



You will note that the voices are different between the two panels. Using the 32-bit panel you can select among the two or more voices for a default voice. Whatever voice you select will be the default for all HomeSeer TTS, whether on the speaker client on the server, tablets or a speaker.exe on a remote client. This voice will also be the default for proxies like the Sonos plug-in. We chose Zira for standard announcements and use David at a slightly elevated level for alerts and warnings. If you select Zira as the default voice, you can change it using SAPI tags in the text string such as

"<voice required="name = Microsoft David Desktop">Warning <silence msec='500'/> the refrigerator freezer compartment is $$DSR:1766:" 

This announcement will speak the text with the temperature device value. The silence for 1/2 second is to create a slight pause in the speech after the alert tone and after speaking "Warning". It works fine on the speaker.exe on the server, Android tablets and my Sonos proxy.