Ommited index marks in SSML (Hynek Hanke)

[James Simmons]

Hynek,

I was interested in your theory that punctuation in the SSML tags was 
messing up the callbacks so I changed my program to completely eliminate 
punctuation in words spoken (except for the apostrophe) and to replace 
the word in the SSML tag with a number representing the byte offset from 
the start of the page being read.  I find that this has no effect.  The 
highlighting falls behind at exactly the same spot.  I have attached my 
modified program plus a log showing the tagged text and the results of 
running it.

Thanks,

James Simmons


Message: 1

>Date: Thu, 29 May 2008 13:58:50 +0200
>From: Hynek Hanke <hanke at brailcom.org>
>Subject: Ommited index marks in SSML
>To: Jonathan Duddington <jsd at clara.co.uk>
>Cc: speechd at lists.freebsoft.org
>Message-ID: <483E9A7A.8050402 at brailcom.org>
>Content-Type: text/plain; charset=ISO-8859-2; format=flowed
>
>
>Dear Jonathan, Lukas,
>
>the latest Speech Dispatcher CVS with espeak 1.36.02
>(and very possibly later versions of both as well) are not
>reporting certain index marks. It is easy to reproduce
>for me.
>
>For example in the text
>
><speak>for<mark name="w_for"/> the<mark name="w_the"/> use<mark 
>name="w_use"/> of<mark name="w_of"/> anyone<mark name="w_anyone"/> 
>anywhere<mark name="w_anywhere"/> at<mark name="w_at"/> no<mark 
>name="w_no"/> cost<mark name="w_cost"/> and<mark name="w_and"/> 
>with<mark name="w_with"/> almost<mark name="w_almost"/> no<mark 
>name="w_no"/> restrictions<mark name="w_restrictions"/> whatsoever.<mark 
>name="w_whatsoever."/></speak>
>
>Given this text, "w_restrictions" is never reported by the Speech Dispatcher
>espeak module and I suspect this is because eSpeak never sends
>the callback. This is with the default English voice, no punctuation,
>no capital letters recognition and no other special settings.
>
>A slight modification in the text ''fixes'' the problem. For example
>removing the dot in 'whatsoever.' both in text and in the index mark
>causes w_restrictions to be reported correctly. Perhaps punctuation
>in mark name is not a very good idea, but I didn't find anything in SSML
>documentation that would rule it out. Removing the first word 'For' and
>its index mark 'w_for' also results in correct behavior.
>
>Please, do you have any ideas? It would be useful to test the above
>text in espeak directly (without SD in the middle), but I don't
>know how to do such a test without having to write my own code
>(and possibly introduce the same bug).
>
>Thank you,
>Hynek Hanke
>
>  
>

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: ReadEtextsActivity.py
Url: 
http://lists.freebsoft.org/pipermail/speechd/attachments/20080603/f564f290/attachment-0002.txt
 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: messages.txt
Url: 
http://lists.freebsoft.org/pipermail/speechd/attachments/20080603/f564f290/attachment-0003.txt