[mary-dev] Adding allophones features

Fabio Tesser fabio.tesser at gmail.com
Wed Oct 27 18:30:30 CEST 2010

Thanks Ingmar,

Vowel length is not a distinctive function in Italian, this is the 
reason because I removed that feature. From another point of view the 
stressed feature is what make a difference between our phonetic simbol 
/a/ and /a1/. In our diphones system was essential the existence of both 
version of the vowel (stressed and unstressed) in the phonetic alphabet. 
The long consonant feature represent a consonant gemination, that is 
different from Vowel lenght.

I think some phonetic features can be language dependent, and, if 
possible, it is better to give informative/explicative names to phonetic 
The relative NLP and prosodic modules must then use the language 
specific definitions.
Anyway I will take this opportunity to think back about the phoneset and 
the relative features.

Thanks for the advices and warnings!


On 10/27/2010 02:49 PM, Ingmar Steiner wrote:
> Hi Fabio,
> basically you just insert features and define the phones in the allophones file. You may want to exercise restraint in adding features at this level that could otherwise be, or perhaps already are, implemented at a different one. (For example, I see that allophones.it.xml contains a feature "stressed"... I don't know much about Italian phonology, but perhaps this is already covered by the "stressed" feature of syllables...) On the other hand, existing AllophoneSets (for German and English, for example) implement a vowel length feature "vlng", which does not apply to consonants. Perhaps you would want to collapse this with your "long" feature for consonants?
> The XML file is loaded and parsed by AllophoneSet. You may want to extend the FeatureProcessorManager for Italian and provide byte-valued feature processors for any new phonological features you implement. Please be aware that there are several places throughout the code where certain phonological features (or even allophone names) are hard-coded, or implicitly assumed to contain certain values. These should of course be more flexible, but I'm afraid we currently don't have the time to clean this up... Feel free to update the it-branch accordingly if you run into any problems in this respect!
> Best wishes,
> /**
>   * Ingmar Steiner
>   * Researcher, Language Technology
>   * German Research Center for Artificial Intelligence
>   *
>   * Campus D3 1 +1.18
>   * D-66123 Saarbrücken
>   * Germany
>   * Phone: ++49-681-857-75-5263 (NEW!)
>   * Email: ingmar.steiner at dfki.de
>   *
>   * Deutsches Forschungszentrum für Künstliche Intelligenz GmbH
>   * Trippstadter Straße 122, D-67663 Kaiserslautern, Germany
>   * Geschäftsführung:
>   * Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster (Vorsitzender)
>   * Dr. Walter Olthoff
>   * Vorsitzender des Aufsichtsrats:
>   * Prof. Dr. h.c. Hans A. Aukes
>   * Amtsgericht Kaiserslautern, HRB 2313
>   */
> On 26 Oct 2010, at 19:25, Fabio Tesser wrote:
>> Hi,
>> I would like to add some phonetic features to the allophones.it.xml for
>> Italian to respect to allophones.de.xml.
>> Does exist a file (schema?) that contains the phonetic features and
>> attributes permitted for each language? Or I simply insert features and
>> attributes in the allophones file?
>> Thanks,
>> Fabio.
>> _______________________________________________
>> Mary-dev mailing list
>> Mary-dev at dfki.de
>> http://www.dfki.de/mailman/cgi-bin/listinfo/mary-dev

More information about the Mary-dev mailing list