Sphinx3 Decoder Problem

[Anonymous]

Hello,
i trained successfully a very small model with the SphinxTrainer and use it with Sphinx4 without any Problems under windows... I will now use the trained acoustic model with sphinx3 and get the following error:

ERROR: "c:\dokumente und einstellungen\la.domrte2\desktop\speech recognizer\sph
nx3\sphinx3-0.7\src\libs3decoder\libdict\dict.c", line 268: Line 1: dict_add_wo
d (<s>) failed (duplicate?); ignored
ERROR: "c:\dokumente und einstellungen\la.domrte2\desktop\speech recognizer\sph
nx3\sphinx3-0.7\src\libs3decoder\libdict\dict.c", line 268: Line 2: dict_add_wo
d (<sil>) failed (duplicate?); ignored
ERROR: "c:\dokumente und einstellungen\la.domrte2\desktop\speech recognizer\sph
nx3\sphinx3-0.7\src\libs3decoder\libdict\dict.c", line 268: Line 3: dict_add_wo
d (</s>) failed (duplicate?); ignored
INFO: c:\dokumente und einstellungen\la.domrte2\desktop\speech recognizer\sphin
3\sphinx3-0.7\src\libs3decoder\libdict\dict.c(482): 0 words read
FATAL_ERROR: "c:\dokumente und einstellungen\la.domrte2\desktop\speech recogniz
r\sphinx3\sphinx3-0.7\src\libs3decoder\libdict\dict.c", line 509: <sil> must oc
ur (only) in filler dictionary

here my Dict:

io IO_PHONE
nio NIO_PHONE

my Filler.

<s> SIL
<sil> SIL
</s> SIL

my Phone:

IO_PHONE
NIO_PHONE
SIL

[Anonymous]

Hello,
please give your eMail adress so i can send you my trained model and the meta data....

Christoph

[Nickolay V. Shmyrev]

Hm, I can only think that your phones are too long. Is there any chance to get your data to try to reproduce the problem?