VoiceXML 2.1 Development Guide

<phoneme> element

The phoneme element essentially allows the developer to specify a phoenetic pronounciation for any TTS nested within.Do note that SSML tags are ignored by some of the available TTS engines, while some engines use their own non-compliant markup. Check the Text-To-Speech appendices for additional details.

usage

attributes

alphabet	Data Type: CDATA	Default: Optional
	The alphabet attribute denotes the phoenetic alphabet to be used. The possible choices are: ipa -- International Phoenetic Alphabet worldbet -- Worldbet (Postscript) phonetic alphabet xsampa -- X-SAMPA phoenetic alphabet
ph	Data Type: CDATA	Default: Required
	The ph attribute specifies the phoneme string to be used.

shadow variables

none

parents

none

children

none

code samples

none

additional links

W3C SSML 1.0 Specification

ANNOTATIONS: EXISTING POSTS

kettle

1/24/2007 2:30 AM (EST)

I'd be really greatful if there were a couple of examples utilizing this tag here.

MattHenry

1/24/2007 2:36 PM (EST)

Hi there,

I should be clear in the fact that the hosted Voxeo VXML platforms do not currently support phoneme. However, you may find a useful workaround by employing the <sub> element:

<prompt>
this message brough to you by the
<sub alias="World Wide Web Consortium">W3C</sub>
</prompt>

Hope this helps!

~Matthew henry

jrhunt

7/28/2010 8:17 PM (EDT)

Just thought I would pass this along for anyone using phoneme to get around mispronunciation of unusual words and names: when using the CMU phoneme alphabet, don't forget to put the stress indicators on the vowels.

Consider this example:

<?xml version="1.0"?>
<vxml version="2.1">
<form>
<block>
<prompt>

Let's go to<phoneme alphabet="x-cmu" ph="P ER S AH L AA P AH L IH S">timbuktu.</phoneme>

Would you rather go to <phoneme alphabet="x-cmu" ph="P ER0 S AH2 L AA1 P AH0 L IH0 S">katmandu</phoneme>?

</prompt>
</block>
</form>
</vxml>

The TTS engine will actually say this: "Let's go to timbuktu. Would you rather go to perselopolis?"

The first phoneme fails to compile and the TTS engine falls back to "timbuktu". The second phoneme compile correctly (and does a decent job of stressing the right parts of a made up city name).

For reference, here is the URL of the CMU dictionary:

http://www.speech.cs.cmu.edu/cgi-bin/cmudict

The stress indicators following the vowels have the following meaning:

0 - no special stress or inflection
1 - primary stress
2 - secondary stress

--
James

voxeoJason

7/28/2010 9:02 PM (EDT)

Hi James,

Thanks for passing along that script! I'll be sure to go ahead and forward that over to our doc team as well.

Cheers,

Jason Sewell
Voxeo Support