Ok dear Heiga,
i checked for the phoneme aa
it appears 143 time
now the problem is in the phoneme boundaries
how can i assign the phonemes boundaries ?how can i determine the starting and the ending for each phoneme?
if this is difficult
can i escape the HInit and go forward to the HERest ?
if this feasible how can i make it and if i made it this will affect to anything else ?
Many Thanks
Tamer Fares
From: "Heiga ZEN (Byung Ha CHUN)" <zen@xxxxxxxxxxxxxxx>
Reply-To: hts-users@xxxxxxxxxxxxxxx
To: hts-users@xxxxxxxxxxxxxxx
Subject: [hts-users:00619] Re: HInit Error
Date: Wed, 11 Apr 2007 22:57:49 +0900
>Hi,
>
>Tamer Fares wrote:
>
>>i made for label file (full) p1^p2-p3+p4=p5 only
>>i made the mono label file too
>> i recorded 60 utterance raw file
>> i run the demo
>> the same error appeared to me
>>
>>system is PLAIN
>>ERROR [+2121] HInit : Too Few Observation Sequences [0]
>>HInit -A -C trn.cnf D -T 1 -S train.scp -m 1 -u tmvw -w 3 -H
>>init.mmf ...... -l aa -o aa state-5_mcp-75_lf0-3.prt
>
>It means that there is no "aa" in the training data.
>>this was happened also when i worked on 50 utterance i added 20
>>additional and the same error appear to me for the phoneme aa
>>can this error belong to something else that it is a few training
>>data for the phoneme aa
>>(can be for example that i included for full label format the
>>p1^p2-p3+p4=p5 only
>
>Please check whether your labes files contain any "aa" or not.
>Of course phone boundaries should be assigned.
>
>Regards,
>
>Heiga ZEN (Byung Ha CHUN)
>
>--
>------------------------------------------------
>Heiga ZEN (in Japanese pronunciation)
>Byung Ha CHUN (in Korean pronunciation)
>
>Department of Computer Science and Engineering
>Nagoya Institute of Technology
>Gokiso-cho, Showa-ku, Nagoya 466-8555 Japan
>
>http://www.sp.nitech.ac.jp/~zen
>------------------------------------------------
>