W3C home > Mailing lists > Public > www-voice@w3.org > April to June 2004

Re: OT: TTS Engine

From: RJ Auburn <rj@voxeo.com>
Date: Wed, 7 Apr 2004 09:14:24 -0700
Message-Id: <A30A997A-88AE-11D8-A631-0003938E6664@voxeo.com>
Cc: <www-voice@w3.org>
To: "Doug Schepers" <doug@schepers.cc>

Doug,

Sadly I have not seen much in the way for free small footprint TTS 
engines for things like imbedding in an HTML web browser. On the 
commercial side there are a number of options from all the big name 
folks like IBM, ScanSoft and Fonix.

As for the subject of "standard" interfaces to TTS engines there are a 
few approaches out there:

	* SAPI Speech API from Microsoft

	This of course will limit you to windows based PC's but you should be
	able to take advantage of the MS TTS engine for example.
	A lot of commercial engines also plug into SAPI.

	* JSAPI from Sun Microsystems

	JSAPI is the standard speech interface for Java programs. The FreeTTS 
engine
	that was listed here a bit ago plugs into this API. Other then that I 
have
	not seen a lot of implementations out there but it might be worth a 
look
	if you are working in a Java environment (However it sounds like you 
are NOT).

	* MRCP

	MRCP is a network protocol for interfacing with TTS and ASR engines.
	It's more targeted for speech based servers such as VoiceXML or SALT
	platforms but it could be used in smaller applications also.
	It's a protocol not an API so you would need to write a client
	to interface with a engine out there.

None of the above will get you something plugged into a web browser but 
they are things to look at.

As for some examples of complete browser plugins take a look at some of 
the following websites:

http://www.latestsoft.com/products/iereader.htm
http://www.codeproject.com/audio/WebVoicePkg.asp
http://www.edict.com.hk/textTospeech/ttsbrowser.htm

There may be others but there were the first few that came up in a 
google web search for me.


I hope this helps and good luck to you!

	RJ

---
RJ Auburn
CTO, Voxeo Corporation
tel:+1-407-418-1800
http://www.rjauburn.com/


On Apr 07, 2004, at 04:18, Doug Schepers wrote:

>
> Hi, everyone-
>
> Wow.
>
> I did not expect so many replies, so quickly. What a friendly 
> community!
> Many thanks to all of you.
>
> Most of the private replies I have gotten have pointed me to Festival. 
> I'd
> looked at Festival, and I'm pleased with the resulting sound 
> synthesis, but
> I cannot seem to find any browser plugin. There are a few references 
> to it,
> but none that lead to active links. Does anyone have any information 
> about
> this? If neccessary, I might be able to build one myself from the 
> Festival
> code, but I'd rather not.
>
> Not to sound ungrateful, but the solution I hoped to find was a browser
> plug-in, preferrably C/C++ based (so that it has as few requirements as
> possible, which is a bit of a limitation of Java, these days; the 
> excellent
> samples I tried were a bit more of a download than I think most users 
> can
> bear, and were not as seamless as I think they'd need to be for 
> purposes of
> accessibility).
>
> Just to supply as much information as possible, I'm trying to create a
> talking calculator (as a simple prototype) to be viewed in a browser. 
> A very
> early version is here:
> http://www.schepers.cc/testbed/accessibility/SelfDescribingSVG.html
>
> This example uses the Microsoft TTS engine, which is fine, but not
> cross-platform. I guess I'd thought that there would be more XML TTS 
> engines
> by now, as I'm hoping for a generic solution. Ultimately, I want to 
> create a
> much more complex SVG+TTS accessible application, and I want it to be
> available for anyone.
>
> An alternate solution that occurs to me is to host a TTS engine on my 
> site,
> and stream the sound, but that hardly seems like the best solution.
>
> Again, many thanks!
> -Doug
>
Received on Wednesday, 7 April 2004 12:15:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 30 October 2006 12:48:59 GMT