[prev in list] [next in list] [prev in thread] [next in thread] 

List:       festvox-talk
Subject:    VS: homograph disambiguation
From:       Stefan Pal <stefan () mikrov ! dk>
Date:       2009-11-18 5:26:22
Message-ID: 722192264D735C4DB27BA9FFF338F0010D60453621 () MAILSERVER1 ! valhalla ! local
[Download RAW message or body]

FYI - What am I doing wrong here?

Festival Speech Synthesis System 1.96:beta July 2004
Copyright (C) University of Edinburgh, 1996-2004. All rights reserved.
For details type `(festival_warranty)'
festival> (load "/home/stefan/tts/festival/examples/toksearch.scm")
nil
festival> (tok_search_db "gutenberg")                              
SIOD ERROR: unbound variable : lead
festival>

-----Oprindelig meddelelse-----
Fra: Stefan Pal 
Sendt: 18. november 2009 06:04
Til: 'Nickolay V. Shmyrev'
Cc: festvox-talk@festvox.org
Emne: SV: homograph disambiguation

Ok - but I was just interested in testing festival's disambig - I just need to figure \
out how to run tok_search_db. I don't understand why it complains about the unbound \
variable, i.e. the defined tokens. Actually, I'm just running toksearch.scm as it is, \
so it should work - or?

-----Oprindelig meddelelse-----
Fra: Nickolay V. Shmyrev [mailto:nshmyrev@yandex.ru] 
Sendt: 18. november 2009 00:57
Til: Stefan Pal
Cc: festvox-talk@festvox.org
Emne: Re: homograph disambiguation

В Втр, 17/11/2009 в 13:12 +0100, Stefan Pal пишет:
> Hi,

> I try to run toksearch.scm, but when I run tok_search_db – I get an 
> unbound variable for the tokens, that I defined in my desired tokens 
> list.

> Has anyone successfully built a CART tree for homograph 
> disambiguation?

As any classifier it can be used, the issue is to get proper features.
Also everything depends on type of homographs. How frequent and diverse they are

As for practical way to implement disambig I'd mentions SRILM which has disambig \
command and can distinguish homographs with language model. Probably it's worth to \
look although not so easy to integrate with festival.


[prev in list] [next in list] [prev in thread] [next in thread]