Some of you may have seen my recent post and video on the cortana/pandorabot and would have heard the completely horrible TTS voice. This morning I installed a Cereproc voice (Giles) on that PC but got the problem that ARC does see it, but will not actually load the Giles TTS and the awful (slow) Hazel voice continues.
I have followed Steve G great tutorial hoping this would allow me to start using Giles, but the problem is that the 4 files Steve suggests to delete, cannot now be deleted and a TrustedInstaller note appears which stops the delete. This must be something that has occurred in the latest win10 software upgrade (anniversary edition)?
Can anyone advise how to get around this? Thanks.
Upgrade to ARC Pro
Unleash your creativity with the power of easy robot programming using Synthiam ARC Pro
Maybe This thread can help you. You can try going through the control panel to change it. You may have to reinstall a voice engine after last round of updates. If you click on the link on page 6 remove the s after http. Don't know how it got there.
There looks to be no way to take ownership of the four TrustedInstaller files (for the Zira TTS in my case) so they then cannot be deleated which is required in Steve G tutorial to be able to load another vendors TTS.
So for (up to date) win 10 users ARC are left with a dreadful Zira TTS that talks so slow it sounds drugged and there seems to be no way now of changing the TTS or the speech settings ie speed etc. Our robots will sound like something out of the eighties then we used phonemes and allophones!
Going thru this www.ez-robot.com/Community/Forum/Thread?threadId=9028&page=6 post #56, I eventually was given the option to download voices, which I did and then I had all the voices I had before including the abilitiy to change the options in the Control.
Hope this helps you out.
Other applications can use all installed voices, including 3rd party voices like Cepstral without issue. Only ARC is stuck on one voice regardless of the selected voice.
Try Roboblather http://zetacentauri.com/software_roboblather.htm to see what I mean. You can select any voice, and modify some of the properties. In ARC, whichever voice I select they all sound the same, although I can change it from male to female and change the speed (which some others can't do, so I think there are some regional differences). They all also sound like an old version of hte voice, not the current much smoother voices used by Cortana and other applications. (note however, Cortana can only use Microsoft voices as far as I can tell. They are better than the one that EZ-Buider is using, but it doesn't see the 3rd party voices).
Alan, you are right the TTS part of ARC seriously needs to be looked at.
Here is working source code for selecting the installed voices. I was able to compile this in VS2015 Community edition. It had some warning, but worked and I can select any installed voice.
There is also a lot of sample code on MSDN for the various speech manipulation functions under dotnet, but I didn't see downloadable samples and wasn't ready to put in the time to figure out how to write a full sample app.
I wonder if the reason I can't get Cepstral Lawrence to work in Cortana is that it is a UK voice and I am in the US with US settings. I can get it to work in the two apps I listed above.
@thetechguru I don't have ARC on a 10 machine, but I do use UK hazel with US Settings/ Region on a different machine. Cortana's voice is UK when testing it. Maybe adding the stock UK English engine will get Cepstral Lawrence working.
An off topic side comment.... I am very pleased to welcome you to the forum. It is rare for us to get someone who joins, apparently reads all the tutorials and back posts, and starts providing valuable hints and tips so quickly (I see you joined a couple months ago, but your activity seems to be mostly very recent). Your input the past few weeks has been very refreshing.