BrightSpeech Voices, do you want them?

[electron]

I was checking the dates right now myself as I could have sworn they started the poll afterwards. I just verified this, and they did start the poll after announcing they can't get the voices.

[Rupp]

Your right!

They obviously were looking for the amount of interest to see if it was profitable to jump in on these voices then.

Has anyones trial run out? I have one day left. Do they simply stop working?

[mikkola]

Yepp, they stop working

[Skibum]

Yes, she just stops speaking to you... No warning, no nothing. Even her clothes are gone from the closet.

[johnwpb]

Without Warning

I got up in the morning... "Good morning Samantha".... The lights came on and the fans went off, but there was no morning report... I rebooted HS to no avail, and THEN realized that I must have reached the end of the trial with Heather. I sure do miss her

The AT&T voice Crystal, that I have re-initiated as my standard voice is just so much more "choppier". I just hope they fix some of the small glitches before it is finally released, such as the date being spoken wrong. Ea: Today is nov (Nahv) twenty-one" Depending on the date for some reason it does say it every now and then correctly. "Today is October 21st, 2004". Just little glitches, but even with the glitches, I would take it any day over AT&T and just alter the scripts to say it correctly

As for HST being able to get the voices, I am confused on this one. In a post on October 29th, Rich posted:

___________
"Sorry guys, I spoke too soon. We could not get the voices added to our current license as I had hoped and expected. So right now we will not be carrying these voices. If we did, the price would be so high you wouldn't buy them anyway!
-Rich
<!-- / message --><!-- sig --> ___________

Then, just 12 hours later, the poll at the top of this thread was started asking if we would be interested in them.

[Gogs]

I have never found a voice that I like, AT&T sound crap over the telephone, oop's replace crap with terrible.

I currently use MSAgent, it's OK throught the house, not brilliant, but useless for the phone.

I would realy appreciate any thought as to voices for use with HSP.

So at this point, and I appologise, I cannot vote as I do not know what the voices are.

[Skibum]

Gogs

As is stated on one of the Brightspeech threads, there is a free 30 day trial of the voices available from www.brightspeech.com . Why not download them and evaluate for yourself. I think that you can see by the results of the poll what most of us think.

[Rupp]

Gogs,
Just a point. I own the AT&T voices as well as the Neovoices and tried the Babel voices. I can say you guys must have one discerning ear. I can't detect that much of difference between any of them. I'm currently using NeoSpeech Kate as well as AT&T Crystal in many of my scripts and they sound basically the same. There are areas where one sounds better than the other but over all they are comparable. I'm beginning to wonder if we are not comparing apples to oranges and some users have different versions of the NEO voices and AT&T voices. I also wonder if the CPU speeds of our PC's can lead to different quality outputs?

[Skibum]

Wow, Rupp. I find the voices VASTLY different!

[pkoslow]

Rupp,

I think sound cards may have something to do with this?

I remember a noticable improvement in the sound of the AT&T 16khz voices when I upgraded to a SB Live 24-bit PCI sound card. Had been using an older 16-bit SB compatible el-cheapo card.

I was just "handing down" the sound card to the server after upgrading some stuff on a workstation and was pleasantly surprised by the change.

Paul

[johnwpb]

Rupp, Please post your voices.

Rupp, and anyone else interested, I have been keeping up on this topic/thread since the very beginning. It does seem that we either have a "discerning ear" or there actually are differences in what our systems sound like with different voices. Oklahoma, as mentioned earlier, was a good example but something I did not hear on my system.

Rupp, could you take a couple moments and upload a .wav/.mp3 to this thread with a comparison of what the voices sound like on your system? Something like the sports page you posted earlier, and maby a weather forcast/briefing.

I do not own the Neo voices and would like to hear that compared with the same text with the Britghespeech voices. I uploaded a comparison, a page or so back in this thread, and the attachment did not make it through the BB upgrade. Heather has died here, so I can not re-post that again unfortunately.

I realize there are demo's for the various voices on the respective pages, but they do not always do the voices justice for the good or bad, when compared to what we are doing with the TTS in the mainstream everyday uses.

I think for all involved, it would be great to hear samples of the different voices in a real world situation to do out own comparisons. Some of us are biased for one reason or another, due to sound cards, SAPI 5 -vs- Sapi 4 ect.

We have all had our opinions on the voices in this thread, it would be great to hear these different voices to allow anyone to make their own comparisons.

[Gogs]

Skibum, thanks for the link I will give it a shot.

Jon00, you are almost, and I say almost, 100% correct, as usual, can't have you getting too carried away here.

I run three computers, two use MS Agent for voice, both have different speach engines and using Merlin they sound completely different. It all depends on the speach engine.

So, who is using what speach engine and how is it performing? My main interest is to use the voice for the HSM, as at present it is unusable.

[sbessel]

I posted some samples of Neo voices here: http://board.homeseer.com/showthread...d=1#post611902

[Rupp]

I would post a wav file of Babel Heather but I can not get Heather to work with HS.speak to file. Any suggestions?

[sbessel]

I would post a wav file of Babel Heather but I can not get Heather to work with HS.speak to file. Any suggestions?

I used a seperate laptop and microphone and recorded 'live' output of the speakers.

That is the only 'real' way to hear the voices, and the actual delays.