Digital Workshop

This topic started on the Advanced Forum but because of its general nature, I decided to post here in case others are interested.

This is not a recommendation. It is just a brief report of my findings.

TTS is of particular interest to me, so following Richard Walter's suggestion I tried Sayzme. I could not install it because one of the required MS files would not install on my system (Win 2K Pro SP4). I decided to research in more depth. After downloading and trying several packages, I settled on Text Aloud from NextUp Technologies http://www.nextup.com/TextAloud/. I downloaded the trial version and then ordered the Pro version.

It works from within most of my text applications (MS Word, WordPerfect, PDF, HTML and RTF files, and for Opus, any text copied to the clipboard is converted to WAV, MP3 and WMF.

The free voices provided with TextAloud are average in that they sound like they are computer-generated, but they are acceptable to me -- just. However, TextAbout is compatible with AT&T's Natural Voice Technology and others. AT&T produces "voices" that I find more acceptable. These voices are available in a variety of styles (male and female) in US English, UK English, Indian Accent English, Latin American Spanish, German and French. I irdered a male UK English voice, and after it arrives (2+ weeks) I'll report back if anyone is interested.

TextAloud Pro costs US$29-95. AT&T voices cost US$35 each.

Thanks mate!
I'm about to begin a project (well, my first project for next year!) that will involve this type of technology.
I'll check this out!

German

I would be very interested to hear the end results of your experiments.

I haven't checked recently, but what licencing and usage restrictions are placed on the 'Natural Voices'?

I did not find any licencing or usage restrictions on the 'Natural Voices'.

I checked the AT&T site, and the TextAloud site through whom I bought the Natural Voices file.

That doesn't mean that there will be no restrictions -- it won't be the first time I've discovered the detail after I've installed a program.

If (when?) I find out more, I'll post the details.

I asked the question. Here are the replies (in italics) togther with my response.

1. In general you can't redistribute audio created with AT&T voices at all. We've seen a few exceptions, but they are generally expensive.


I think you and AT&T have just blown what could have been many sales. I'll post your response and let others decide for themselves.

I also believe that by not prominently posting these limitations where they
can be read BEFORE purchase, you may be running very close to being in
breach of consumer protection laws in many jurisdictions. For example, it is unlikely that courts in England and Australia would enforce such ex post
facto limitations.

We have other voices http://www.nextup.com/neospeech.html

Perhaps, but if I recall correctly, these were US voices.

On the off-chance that someone wants more information on free and low-cost TTS applications, I thought I would post a url I found useful:

http://www.handyarchive.com/free/text-speech/

Hi,

Just for info I append copy of email from the people - wizzardsoftware - who deal with the licences for Natural Voices.

----------

To distribute the intellectual properties of AT&T Natural Voices such as by wav files or the engine and voice font being sent along with your product is our Runtime license. If you wish to distribute the intellectual properties of AT&T Natural Voices with you product then the Runtimes are a minimum of 200 at the starting price of $7.50 USD each. These Runtimes do not expire, so if you use say just one this year you have 199 to have as long as you need until they are used up. The more Runtimes you purchase up front the cheaper the Runtime pricing. Each additional Voice Font used in the same product increase the Runtime pricing 20%. So if you use two Voice Fonts in a product the Runtime pricing would be 200 at $9.00 USD each.

I hope this fits your business model and please let me know how you wish to proceed.

Doug Whitson
Wizzard Software
Senior Technical Sales Consultant
Office 954-678-4155 Ext 212
Fax 954-678-4164

---------------------

Certainly doesn't fit my business model.

John

It is interesting that AT&T does not post the licencing info in any way that it can be readily found. I approached them about this and basically, they said "tough".

I responded that I thought that Bristish and Australian courts would be unlikely to enforce hidden, post-purchase restrictions (there is some evidence to support this). AT&T have remained absolutely silent after this.

On this basis, while their Natural Voices are good, I would not be prepared to even consider using their software. So the US35 does down to one of life's experiences.

This weekend is to be devoted to testing several other TSS engines that I downloaded last night. Some, look interesting. None relies on Natural Voices.

I'm surprised at your reaction.

Do you also feel this way about fonts? I ask because font redistribution is analogous: just because someone publishes their manual using an interesting, proprietary font does not give you the right to redistribute that font -- no matter how much you think it's great and would really help you in your own line of work.

You may want to read the AT&T product's software license closely; over the years, I've worked with many AT&T executives, including intellectual property people, and I'd bet that their software rights are airtight.

Would it be nice of AT&T to let you use their "font" voices? Sure. Would it be a significant revenue stream for AT&T to license these voices to others? Apparently only at the stated rates -- and their accountants have sharper pencils than you or I, so they probably figured out a rate that the market will bear.

I'm surprised at your reaction.

Do you also feel this way about fonts? I ask because font redistribution is analogous: just because someone publishes their manual using an interesting, proprietary font does not give you the right to redistribute that font -- no matter how much you think it's great and would really help you in your own line of work.

You may want to read the AT&T product's software license closely; over the years, I've worked with many AT&T executives, including intellectual property people, and I'd bet that their software rights are airtight.

Would it be nice of AT&T to let you use their "font" voices? Sure. Would it be a significant revenue stream for AT&T to license these voices to others? Apparently only at the stated rates -- and their accountants have sharper pencils than you or I, so they probably figured out a rate that the market will bear.

Robert

As usual, you miss the point completely.

1. For starters I only ever use fonts for which I have a licence to do so, and despite your strange reasoning, font use is not analogous.

2. I do not expect to use the AT&T files for free. They were advertised at a price. I search the vendor's site and AT&T's site for licencing information. Nowhere to be found.

3. After paying for a so called "licence", I then learn about the restrictions which make it useless as a product for me, and I would guess, many people.

4. In most legal jurisdictions (except where you live) courts will not enforce such restrictions. Years ago, Australian courts rejected the "shrink wrapped licence" so favoured of software producers. There is every chance they would do the same with the AT&T licence.

5. I'm sorry, but the fact that you have worked with and know some AT&T executives is completely irrelevant, but perhaps it makes you feel better being able to claim that you know them.

I have no problem with AT&T having a licence and imposing restrictions. That is their right -- but I do think that they should make such restrictions obvious. Incidentally, I checked web sites of several vendors of Natural Voices. Not one mentions the restrictions, but will provide details only if you ask. A majority of vendors of other products seem to me more forthcoming.

As I said before, I won't use Natural Voice. That's not a problem. I hope this thread serves as a warning to others. I will also post a copy of the AT&T licencing restrictions on my website.

"As usual, you miss the point completely."

What a gratuitously rude response!

"despite your strange reasoning, font use is not analogous."

Another gratuitously rude comment. AT&T refers to the software component voices as "voice fonts." They draw the analogy -- which I think is not an unfair one.

"I search the vendor's site and AT&T's site for licencing information. Nowhere to be found."

I've not seen a site that publishes its software license in detail, although if they advertise licensing for re-use, I too (like you) would have expected a definition and price list somewhere.

"In most legal jurisdictions (except where you live) courts will not enforce such restrictions [....] There is every chance they would do the same with the AT&T licence."

There's always a chance. (Except where I live?)

"the fact that you have worked with and know some AT&T executives is completely irrelevant, but perhaps it makes you feel better being able to claim that you know them."

Ah ... one last gratuitous insult?

Historically speaking, AT&T's intellectual property people were among the toughest legal minds in the business. I mentioned them because I've dealt with some of those people, done training in conjunction with them, and dealt with many copyright and fair use issues. They are cautious about their licensing & frankly they've often won international infringement cases.

"I have no problem with AT&T having a licence and imposing restrictions. That is their right -- but I do think that they should make such restrictions obvious."

Well, we can agree at least on that much.

Please do not use this site to insult other users (this is directed at all users - not just the specific ones in this thread) - it is a support and information service and we want it to remain a friendly destination.

What a surprise! I thought few people were interested in this, so was pleased when I started browsing around that it had been taken up here in 'the Lounge'.

Yes, I had a tedious run-in with Alan Costilo of wizzardsoftware. I was prepared to pay $7.50 per licence (even a bit more), but only wanted a few licences, but they wouldn't budge. It had to be a minimum of 200. A shame because their voices are good.

I also looked at Loquendo in Italy. They gave me a beta copy of their software to try. Unfortunately I gave it to my boss to test and he did nothing with it. The company has since folded and I am now working free-lance and am too embarrased to go back to them. They weren't cheap either.

Others were Scansoft and Neospeech but both expensive. Rhetorical in Scotland only did server side engines for the web and telephony(very expensive!), who by the way have been bought out by Scansoft.

All I really want is a plugin that will talk to SAPI 5 and use whatever is installed on the machine, or the free Microsoft voices.

I'll keep looking here, but is someone decides to carry on this thread somwhere else, they might let me know?

Cheers all

Eleonora

ScanSoft seems to be buying quite a few companies these days. I find this disappointing because I watched the superb TextBridge Pro decline in usability and accuracy after ScanSoft took over. I guess the old business adage is still often true: if you can't beat 'em ... buy 'em.