Is "ng~" a phone for the language?It seems the observation for phone "ng~" is too few to train an initial model. You can delete the "ng~" from your monophone list, and copy some "close" phone's model as the initial one for "ng~".--On Thu, Jan 24, 2013 at 5:25 PM, Naresh kumar <elluru.nareshkumar@xxxxxxxxx> wrote:
Dear All,I am building Telugu HTS Voice using 10 hours of speech data. While Intialization of HMM parameters, I got the following error.
../../tool/bin/HInit -A -C /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/config/general.conf -D -V -H /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/hmm.5-state/VarianceFloor/vFloor.cmp.mmf -I /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/list/mono.mlf -M /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/hmm.5-state/SegmentalKmeans -S /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/Analysis/train.tel_lenina_01.unilex -T 1 -l ng~ -m 1 -o ng~ -u tmvw -w 3 /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/hmm.5-state/VarianceFloor/cmp.prt
../../tool/bin/HInit -A -C /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/config/general.conf -D -V -H /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/hmm.5-state/VarianceFloor/vFloor.cmp.mmf -I /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/list/mono.mlf -M /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/hmm.5-state/SegmentalKmeans -S /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/Analysis/train.tel_lenina_01.unilex -T 1 -l ng~ -m 1 -o ng~ -u tmvw -w 3 /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/hmm.5-state/VarianceFloor/cmp.prt
HTK Configuration Parameters[7]
Module/Tool Parameter Value
# MAXSTDDEVCOEF 10
# DURVARFLOORPERCENTILE 1.000000
# APPLYDURVARFLOOR TRUE
# VFLOORSCALESTR Vector 5 0.01 0.01 0.01 0.01 0.01
# APPLYVFLOOR TRUE
# NATURALWRITEORDER TRUE
# NATURALREADORDER TRUE
HTK Version Information
Module Version Who Date : CVS Info
HInit 3.4 CUED 25/04/06 : $Id: HInit.c,v 1.8 2008/05/30 07:19:10 zen Exp $
HShell 3.4 CUED 25/04/06 : $Id: HShell.c,v 1.11 2008/06/24 03:19:08 zen Exp $
HMem 3.4 CUED 25/04/06 : $Id: HMem.c,v 1.7 2008/05/30 07:19:16 zen Exp $
HLabel 3.4 CUED 25/04/06 : $Id: HLabel.c,v 1.7 2008/06/24 03:19:09 zen Exp $
HMath 3.3 CUED 28/04/05 : $Id: HMath.c,v 1.11 2008/05/30 07:19:15 zen Exp $
HSigP 3.4 CUED 25/04/06 : $Id: HSigP.c,v 1.4 2008/05/30 07:19:16 zen Exp $
HWave 3.4 CUED 25/04/06 : $Id: HWave.c,v 1.7 2008/06/24 03:19:08 zen Exp $
HAudio 3.4 CUED 25/04/06 : $Id: HAudio.c,v 1.5 2008/05/30 07:19:15 zen Exp $
HVQ 3.4 CUED 25/04/06 : $Id: HVQ.c,v 1.4 2008/05/30 07:19:15 zen Exp $
HModel 3.4 CUED 25/04/06 : $Id: HModel.c,v 1.30 2008/06/24 03:19:08 zen Exp $
HParm 3.4 CUED 25/04/06 : $Id: HParm.c,v 1.10 2008/06/24 03:19:08 zen Exp $
HTrain 3.4 CUED 25/04/06 : $Id: HTrain.c,v 1.17 2008/11/06 06:45:38 uratec Exp $
HUtil 3.4 CUED 25/04/06 : $Id: HUtil.c,v 1.21 2008/11/06 06:50:25 uratec Exp $
Initialising HMM /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/hmm.5-state/VarianceFloor/cmp.prt . . .
States : 2 3 4 5 6 (width)
Mixes s1: 1 1 1 1 1 ( 120 )
( 120 )
Num Using: 0 0 0 0 0
Mixes s2: 2 2 2 2 2 ( 1 )
( MSD <= 1 )
Num Using: 0 0 0 0 0
Mixes s3: 2 2 2 2 2 ( 1 )
( MSD <= 1 )
Num Using: 0 0 0 0 0
Mixes s4: 2 2 2 2 2 ( 1 )
( MSD <= 1 )
Num Using: 0 0 0 0 0
Mixes s5: 1 1 1 1 1 ( 63 )
( 63 )
Num Using: 0 0 0 0 0
Parm Kind: USER
Number of owners = 1
SegLab : ng~
maxIter : 20
epsilon : 0.000100
minSeg : 1
Updating : Means Variances MixWeights/DProbs TransProbs
- system is PLAIN
ERROR [+2121] HInit: Too Few Observation Sequences [0]
FATAL ERROR - Terminating program ../../tool/bin/HInit
--------------------------------------------------
Realtime 0:00:15
--------------------------------------------------
Error in this command : ../../tool/bin/HInit -A -C /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/config/general.conf -D -V -H /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/hmm.5-state/VarianceFloor/vFloor.cmp.mmf -I /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/list/mono.mlf -M /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/hmm.5-state/SegmentalKmeans -S /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/Analysis/train.tel_lenina_01.unilex -T 1 -l ng~ -m 1 -o ng~ -u tmvw -w 3 /home/naresh/HTS_voices/hts_dir/Lenina_voice/built_voices/tel_lenina_01.unilex.39-order.5-shift.spec-1.f0-4/hmm.5-state/VarianceFloor/cmp.pr
--
Regards
Naresh Kumar
B4-104,Speech and Vision Lab
IIIT-H
Best regards!
Xi Wang (汪曦)