how to specify soft 'g' phoneme?

[greg]

This is probably another question for Johan, but I'll address it to
the group.

I tried to specify the soft 'g' phoneme (e.g., gin) using 'dZ' as
suggested in the SAMPA chart at http://en.wikipedia.org/wiki/SAMPA_chart_for_English

However the following code produces a 'zh' sound (e.g., vision):

text = "<speak xml:lang=\"en-US\"> <phoneme alphabet=
\"xsampa\" ph=\"dZIn\"/>.</speak>";
mTts.speak(text, TextToSpeech.QUEUE_ADD, null);

What is the proper way to specify a soft 'g' sound? And is there a
better reference for specifying phonemes to SVOX pico than the SAMPA
chart at http://en.wikipedia.org/wiki/SAMPA_chart_for_English ?

Thanks,
Greg

[davemac]

Was there an answer to the question about a good reference for
specifying phonemes? This page looks like it would be helpful if it
had more examples:

http://en.wikipedia.org/wiki/X-SAMPA

as the modifiers look like the ones talked about in earlier posts to
this forum. But since I'm not an expert in speech synthesis I have a
hard time understanding which modifier would be appropriate to create
the "g" sound as in "gin".

- dave

On Sep 24, 2:33 pm, greg <gsgm2...@gmail.com> wrote:
> This is probably another question for Johan, but I'll address it to
> the group.
>
> I tried to specify the soft 'g' phoneme (e.g., gin) using 'dZ' as

> suggested in the SAMPA chart athttp://en.wikipedia.org/wiki/SAMPA_chart_for_English

[greg]

Hi, Dave.

Johan suggested d_Z to get a soft 'g' and it worked. Don't ask me
why: I'm a novice and it still seems a little like magic to me. He
also suggested the same Wikipedia page you referenced. If I gather a
couple dozen examples (e.g., xsampa sequence "d_ZIn" to generate
"gin"), I'll post them to that page.

- Greg