VoxForge
Hi,
does anybody know which symbols are allowed to use as a phonem abbreviation?
I'm asking because I want to use SAMPA (instead of IPA) as my phonem syntax but I noticed that htk doesn't like digits and also the "?" seems to make some trouble!
And I don't know if the program can make differences between the upper and lower case! Also I'm not sure if the ":" symbol can cause any problems!
Nick
Hi Nick,
Searching the HTK mail archives is helpful sometimes:
htk-developers.mbox 13-Mar-2009 18:12 3.2M
htk-users.mbox 13-Mar-2009 18:12 24M
This post explains things a bit:
>
> Well i am working on a language URDU , which has a phonetic representation
> of the words which are not listed in the IPA , and the first few steps of
> the HTK till the building of the Dictonary requires the use of the Phonetic
> level trascription .
> Is is possible that i can tell the htk that this is a voice that correspond
> to a symbol that i derived myself.
You can use arbitrary strings as phone labels in HTK. I'd suggest you
restrict them to US-ASCII characters though. HTK has no list of
allowed phone symbols (IPA or others) built into it.
You could call your phones p1, p2, p3,... if you really wanted to.
See also section "4.6 Strings and Names" of the HTK book.
Ken