Using ATT's Speech Engine

[Mudguppie]

Does anyone here use the AT&T Natural Voices speech engine with HCA? I use it but I'm having problems with the logs not updating after midnight. Apparently, the ATT speech engine starts an executable called "nvproxy.exe" which grabs the HCA log file as one of its handles. When HCA closes the file for updating around midnight, nvproxy.exe grabs the file and won't let go. Therefore HCA cannot log any new events nor can it read the log file without rebooting. I've gotten around this by running MS Process Explorer and deleting the handle from nvproxy.exe which works but only until the next reboot which occurs often because of MS' continual updates.

Thanks.

[KHarms]

I worked with this person for quite some time and never was able to resolve this. Why a speech engine should open the HCA log file - grabbing it when it gets closed at midnight to do the overflow processing - never made any sense.

While I beleive what he sees, I was unable to come up with any plausible scenario why it should be so.

Kimberly Harms

[mjl]

The best I can come up with is that nvproxy might be some sort of reverse proxy software to allow web based sharing of files. [http://www.delegate.org/delegate/nvproxy/] I don't see any indication that it is related to AT&T's software, but it's possible that they used it (I do not use any speech capabilities).

If something in the AT&T software is configured to use or share files in the HCA or HCA log folders, you might want to instead set up a process to periodically copy the files to another folder and have the AT&T software access the files elsewhere ( or configure it not to access that folder if it is not necessary).

Alternatively, you could try configuring the HCA Log file to go someplace different.

Does anyone here use the AT&T Natural Voices speech engine with HCA? I use it but I'm having problems with the logs not updating after midnight. Apparently, the ATT speech engine starts an executable called "nvproxy.exe" which grabs the HCA log file as one of its handles. When HCA closes the file for updating around midnight, nvproxy.exe grabs the file and won't let go. Therefore HCA cannot log any new events nor can it read the log file without rebooting. I've gotten around this by running MS Process Explorer and deleting the handle from nvproxy.exe which works but only until the next reboot which occurs often because of MS' continual updates.

Thanks.

[Mudguppie]

I'm sorry mjl, the name of the executable is actually "nvdesktopproxy.exe" and, I have no idea why it has that file and several HCA directories listed as handles. I have indeed tried to move or rename the files but it seems to know where they are and claims them. I don't know from where it's getting this information however, it appears as a child process of HCA in Process Explorer so maybe it inherits the ownership of that file from HCA...who knows?

[mjl]

Belowis a quote from a newsgroup on the AT&T Natural Voice. This doesn't explain why it is opening the files, but it does explain a bit about what the file does and how it works. I suspect that NVDesttopProxy.exe might proxy file activity to the program it attaches to. Since I'm not familiar with the AT&T Natural Voice product, I'm afraid I probably can't help much more than this. :confused:



(http://www.hcatech.com/forum/posting.php?mode=quote&p=649&sid=a2c9bcbf3f3c7b9a0939f2f6f85ab265) NVDesttopProxy.exe is the program that manages Natural Voices for any
SAPI requests for the voices. I've NEVER seen it get started without some type of voice request causing it. I'd check to see if possibly ReadPlease, ReadingBar, or some other speech program is running, in the system tray, as a toolbar in IE, or somewhere else like that. Also, you may want to set some voice other than an AT&T voice as the windows default voice under the control panel speech icon. For NVDestTopProxy.exe to be started, there is something out there making sapi requests, the challenge is to just figure out exactly what is doing it.


I'm sorry mjl, the name of the executable is actually "nvdesktopproxy.exe" and, I have no idea why it has that file and several HCA directories listed as handles. I have indeed tried to move or rename the files but it seems to know where they are and claims them. I don't know from where it's getting this information however, it appears as a child process of HCA in Process Explorer so maybe it inherits the ownership of that file from HCA...who knows?

[Mudguppie]

Thanks for the reference mjl. The problem is that once nvdesktopproxy.exe receives a voice request from HCA and successfully processes it, it doesn't terminate itself--it stays resident. I'm not sure where to go from here. I'm wondering if there's a way of restricting access to that file by nvdesktopproxy.exe using the XP security model. Can you restrict certain processes' access to certain files? Even if one is a child process of the other? Maybe it's worth it for me to research it. I'll post any successful results I may have.

[jhendric1]

Not directly related to this thread but... is there any advantage to using a speech engine vs playing a wav file? Other than of course having a bunch of wav files saved. For instance, I've created several wavs using AT&Ts online speed engine like "The side gate is open" and have them played when triggered thru an element in vp.

[KHarms]

One good reason to use the Speech element is that you can create the text to be heard “on the fly”. The speech element gives you the ability to compose the text string by substituting data for placeholders. For example you can use:

The time is %_now()%

And it speaks the current dfate and time.

[jhendric1]

Yep, that dawned on me last evening. Chalk it up to not fully realizing what you can do with HCA. I mean my gate is either open or closed, two states two wav files. Now annoucing the time of day is a another story. So now I'm wondering about something like this - http://www.youtube.com/watch?v=qYa777RBT50. Please don't kill me as another HA product is mentioned. I don't have an interest in it... it uses X10.

I see that there are many out there so is there a recommended speech engine for HCA?

[ewelin]

I'm also toying with the idea of adding speech to my automation solution. I just recently purchase a couple of wireless speakers and hope to have a HCA speak a few alerts. I really hate Microsoft Sam that's included with WHS. Anyone have any good recommendations for TTS engines that they are using in their setup?

[Mudguppie]

The ATT TTS works fine with HCA ignoring the above mentioned "feature". I use it to announce the time, date, inside and outside temp (from weather station) and forecast which gets "scraped" from a weather website. It also announces when someone has changed the thermostat setpoint and tells me the new value. There are other announcement applications for which I use this TTS as well. Currently, I use the "Audrey" voice and it's (she's) very intelligible.

[jhendric1]

Here is a link to the latest AT&T speech engine. The 16k voices are really realistic. http://wizzardsoftware.com/att_NV_demo.php. I just type in what I want it to say and it supplies the wav file for it. I save it and use HCA to play it back for whatever. ie: "kids it is 8 o clock! You are late for school!".

Speaking of which, how can HCA say the time without the seconds? %_now()% gives hr:mn:sec and the date. I just want hr:mn.

[ewelin]

You could use something like:
%_formatTime( _now() , "$I:$M $p" )%
The formatting can be adjusted and is outlined in page 189 of the userguide which can be found on the Documentation (http://www.hcatech.com/index.php?sid=documentation)section