question about length audio file for trainnin

[nguyen duy nam]

hi everybody
i read tutorial http://cmusphinx.sourceforge.net/wiki/tutorialam found:
Optimal length is not less than 5 seconds and not more than 30 seconds.
but i only recode once sound "a" so the audio file only 2s, i do it is true or
false.
thanks

[Nickolay V. Shmyrev]

Do you know the meaning of the word optimal? If no please consult with the
dictionary.

[nguyen duy nam]

thanks, i know.
when i trainning, run ./scripts_pl/RunAll.pl

then terminal show:
**
./scripts_pl/RunAll.pl
MODULE: 00 verify training files
O.S. is case sensitive ("A" != "a").
Phones will be treated as case sensitive.
Phase 1: DICT - Checking to see if the dict and filler dict agrees with the
phonelist file.
Found 41 words using 36 phones
Phase 2: DICT - Checking to make sure there are not duplicate entries in the
dictionary
Phase 3: CTL - Check general format; utterance length (must be positive);
files exist
Phase 4: CTL - Checking number of lines in the transcript should match lines
in control file
Phase 5: CTL - Determine amount of training data, see if n_tied_states seems
reasonable.
Estimated Total Hours Training: 0.321669444444444
This is a small amount of data, no comment at this time
Phase 6: TRANSCRIPT - Checking that all the words in the transcript are in the
dictionary
Words in dictionary: 38
Words in filler dictionary: 3
WARNING: Utterance ID mismatch on line 2: mswav/aw vs b
WARNING: Utterance ID mismatch on line 3: mswav/aa vs ba
WARNING: Utterance ID mismatch on line 4: mswav/b vs bay
WARNING: Utterance ID mismatch on line 5: mswav/c vs bon
WARNING: Utterance ID mismatch on line 6: mswav/d vs c
WARNING: Utterance ID mismatch on line 7: mswav/dd vs chin
WARNING: Utterance ID mismatch on line 8: mswav/e vs d
WARNING: Utterance ID mismatch on line 9: mswav/ee vs e
WARNING: Bad line in transcript:
HAI (hai) I
WARNING: Utterance ID mismatch on line 12: mswav/i vs
WARNING: Utterance ID mismatch on line 13: mswav/k vs i
WARNING: Utterance ID mismatch on line 14: mswav/l vs k
WARNING: Utterance ID mismatch on line 15: mswav/m vs khong
WARNING: Utterance ID mismatch on line 16: mswav/n vs l
WARNING: Utterance ID mismatch on line 17: mswav/o vs m
WARNING: Utterance ID mismatch on line 18: mswav/oo vs mot
WARNING: Utterance ID mismatch on line 19: mswav/ow vs nam
WARNING: Utterance ID mismatch on line 20: mswav/p vs o
WARNING: Utterance ID mismatch on line 21: mswav/q vs p
WARNING: Utterance ID mismatch on line 22: mswav/r vs q
WARNING: Utterance ID mismatch on line 23: mswav/s vs r
WARNING: Utterance ID mismatch on line 24: mswav/t vs s
WARNING: Utterance ID mismatch on line 25: mswav/u vs sau
WARNING: Utterance ID mismatch on line 26: mswav/uw vs t
WARNING: Utterance ID mismatch on line 27: mswav/v vs tam
WARNING: Utterance ID mismatch on line 28: mswav/y vs u
WARNING: Utterance ID mismatch on line 29: so/mot vs v
WARNING: Utterance ID mismatch on line 30: so/hai vs x
WARNING: Utterance ID mismatch on line 31: so/ba vs y
WARNING: Utterance ID mismatch on line 32: so/bon vs aa
WARNING: Utterance ID mismatch on line 33: so/nam vs ee
WARNING: Utterance ID mismatch on line 34: so/sau vs oo
WARNING: Utterance ID mismatch on line 35: so/bay vs aw
WARNING: Utterance ID mismatch on line 36: so/tam vs dd
WARNING: Utterance ID mismatch on line 37: so/chin vs ow
WARNING: Utterance ID mismatch on line 38: so/khong vs uw
Phase 7: TRANSCRIPT - Checking that all the phones in the transcript are in
the phonelist, and all phones in the phonelist appear at least once
WARNING: This phone (SIL) occurs in the phonelist
(/Application/Data/speechtotext/demo/datatest2/etc/words.phone), but not in
any word in the transcription
(/Application/Data/speechtotext/demo/datatest2/etc/words_train.transcription)
Something failed: (/Application/Data/speechtotext/demo/datatest2/scripts_pl/00
.verify/verify_all.pl)

**
can you show me where wrong. thanks

[nguyen duy nam]

forget: words.fileids
file

mswav/a
mswav/aw
mswav/aa
mswav/b
mswav/c
mswav/d
mswav/dd
mswav/e
mswav/ee
mswav/g
mswav/h
mswav/i
mswav/k
mswav/l
mswav/m
mswav/n
mswav/o
mswav/oo
mswav/ow
mswav/p
mswav/q
mswav/r
mswav/s
mswav/t
mswav/u
mswav/uw
mswav/v
mswav/y
so/mot
so/hai
so/ba
so/bon
so/nam
so/sau
so/bay
so/tam
so/chin
so/khong

and file transcription

A (a)
B (b)
BA (ba)
BẢY (bay)
BỐN (bon)
C (c)
CHÍN (chin)
D (d)
E (e)
G (g)
H (h)
HAI (hai) I
I (i)
K (k)
KHÔNG (khong)
L (l)
M (m)
MỘT (mot)
NĂM (nam)
O (o)
P (p)
Q (q)
R (r)
S (s)
SÁU (sau)
T (t)
TÁM (tam)
U (u)
V (v)
X (x)
Y (y)
 (aa)
Ê (ee)
Ô (oo)
Ă (aw)
Đ (dd)
Ơ (ow)
Ư (uw)

and file dic
**
A A
B B
BA B A
BẢY B AR Y
BỐN B OOS N
C C
CHÍN CH IS N
D D
E E
G G
H H
HAI H A I
I I
K K
KHÔNG K H OO N G
L L
M M
MỘT M OOJ T
NĂM N AW M
O O
P P
Q Q
R R
S S
SÁU S AS U
T T
TÁM T AS M
U U
V V
X X
Y Y
 AA
Ê EE
Ô OO
Ă AW
Đ DD
Ơ OW
Ư UW

---

[Nickolay V. Shmyrev]

The script tells you exactly what is wrong. Utterance ids do not match in
transcription flie and fileids file on line 2.

mswav/aw

vs

B   (b)

You should have a transcription for file aw on the second line.

There is incorrect format line in transcription file

HAI (hai) I

It's not correct because there is an extra letter l after utterance id.

[nguyen duy nam]

ok, thanks i passed it. but when i run again so it have error:

./scripts_pl/RunAll.pl
MODULE: 00 verify training files
O.S. is case sensitive ("A" != "a").
Phones will be treated as case sensitive.
Phase 1: DICT - Checking to see if the dict and filler dict agrees with the
phonelist file.
Found 41 words using 36 phones
Phase 2: DICT - Checking to make sure there are not duplicate entries in the
dictionary
Phase 3: CTL - Check general format; utterance length (must be positive);
files exist
Phase 4: CTL - Checking number of lines in the transcript should match lines
in control file
Phase 5: CTL - Determine amount of training data, see if n_tied_states seems
reasonable.
Estimated Total Hours Training: 0.320019444444444
This is a small amount of data, no comment at this time
Phase 6: TRANSCRIPT - Checking that all the words in the transcript are in the
dictionary
Words in dictionary: 38
Words in filler dictionary: 3
Phase 7: TRANSCRIPT - Checking that all the phones in the transcript are in
the phonelist, and all phones in the phonelist appear at least once
WARNING: This phone (SIL) occurs in the phonelist
(/Application/Data/speechtotext/demo/datatest2/etc/words.phone), but not in
any word in the transcription
(/Application/Data/speechtotext/demo/datatest2/etc/words_train.transcription)
MODULE: 01 Train LDA transformation
Skipped (set $CFG_LDA_MLLT = 'yes' to enable)
MODULE: 02 Train MLLT transformation
Skipped (set $CFG_LDA_MLLT = 'yes' to enable)
MODULE: 05 Vector Quantization
Skipped for continuous models
MODULE: 10 Training Context Independent models for forced alignment and VTLN
Skipped: $ST::CFG_FORCEDALIGN set to 'no' in sphinx_train.cfg
Skipped: $ST::CFG_VTLN set to 'no' in sphinx_train.cfg
MODULE: 11 Force-aligning transcripts
Skipped: $ST::CFG_FORCEDALIGN set to 'no' in sphinx_train.cfg
MODULE: 12 Force-aligning data for VTLN
Skipped: $ST::CFG_VTLN set to 'no' in sphinx_train.cfg
MODULE: 20 Training Context Independent models
Phase 1: Cleaning up directories:
accumulator...logs...qmanager...models...
Phase 2: Flat initialize
Phase 3: Forward-Backward
Baum welch starting for 1 Gaussian(s), iteration: 1 (1 of 1)
0% 10% 50% 60% 100%
WARNING: This step had 0 ERROR messages and 1 WARNING messages. Please check
the log file for details.
Normalization for iteration: 1
WARNING: This step had 0 ERROR messages and 33 WARNING messages. Please check
the log file for details.
Current Overall Likelihood Per Frame = 3.5063077764372
Baum welch starting for 1 Gaussian(s), iteration: 2 (1 of 1)
0% 10% 50% 60% 100%
WARNING: This step had 0 ERROR messages and 1 WARNING messages. Please check
the log file for details.
Normalization for iteration: 2
WARNING: This step had 0 ERROR messages and 33 WARNING messages. Please check
the log file for details.
Current Overall Likelihood Per Frame = 4.27383926323921
Convergence Ratio = 0.767531486802013
Baum welch starting for 1 Gaussian(s), iteration: 3 (1 of 1)
0% 10% 50% 60% 100%
WARNING: This step had 0 ERROR messages and 1 WARNING messages. Please check
the log file for details.
Normalization for iteration: 3
WARNING: This step had 0 ERROR messages and 33 WARNING messages. Please check
the log file for details.
Current Overall Likelihood Per Frame = 4.98022429192671
Convergence Ratio = 0.706385028687496
Baum welch starting for 1 Gaussian(s), iteration: 4 (1 of 1)
0% 10% 50% 60% 100%
WARNING: This step had 0 ERROR messages and 1 WARNING messages. Please check
the log file for details.
Normalization for iteration: 4
WARNING: This step had 0 ERROR messages and 33 WARNING messages. Please check
the log file for details.
WARNING: WARNING: NEGATIVE CONVERGENCE RATIO AT ITER 4! CHECK BW AND NORM
LOGFILES
Current Overall Likelihood Per Frame = 4.89577022229552
Training completed after 4 iterations
MODULE: 30 Training Context Dependent models
Phase 1: Cleaning up directories:
accumulator...logs...qmanager...
Phase 2: Initialization
This step had 1 ERROR messages and 0 WARNING messages. Please check the log
file for details.
Phase 3: Forward-Backward
Baum welch starting for iteration: 1 (1 of 1)
0%
This step had 1 ERROR messages and 0 WARNING messages. Please check the log
file for details.
Only 0 parts of 1 of Baum Welch were successfully completed
Parts 1 failed to run!
Training failed in iteration 1
Something failed: (/Application/Data/speechtotext/demo/datatest2/scripts_pl/30
.cd_hmm_untied/slave_convg.pl)

**
can you show me to solve it. thanks indeed.

[nguyen duy nam]

pls, help me.
it not generate files in directory model_parameters/words.cd_cont_untied
do, logs file

INFO: main.c(282): Reading: /Application/Data/speechtotext/demo/datatest2/mode
l_parameters/words.cd_cont_untied/mixture_weights
WARN: "s3io.c", line 256: Unable to open /Application/Data/speechtotext/demo/d
atatest2/model_parameters/words.cd_cont_untied/mixture_weights for reading; No
such file or directory
FATAL_ERROR: "main.c", line 771: Initialization failed

where did i wrong? pls

[Nickolay V. Shmyrev]

where did i wrong? pls

You ignored the warnings in the training tool output. They clearly hint you
that the input data is still not properly prepared.

[nguyen duy nam]

thanks, can you show me what did i wrong?
with my files and steps create them :
file words.txt

a
ă
â
b
c
d
đ
e
ê
g
h
i
k
l
m
l
o
ô
ơ
p
q
r
s
t
u
ư
v
x
y
một
hai
ba
bốn
năm
sáu
bảy
tám
chín
không

end i create file vocab:

text2wfreq < words.txt | wfreq2vocab > words.tmp.vocab

then i delete comment and rename words.tmp.vocab -> words.vocab
file words.vocab

a
b
ba
bảy
bốn
c
chín
d
e
g
h
hai
i
k
không
l
m
một
năm
o
p
q
r
s
sáu
t
tám
u
v
x
y
â
ê
ô
ă
đ
ơ
ư

i create file idngram

text2idngram -vocab words.vocab -idngram words.idngram < words.txt

i create file arpa

idngram2lm -vocab_type 0 -idngram words.idngram -vocab words.vocab -arpa
words.arpa

file words.arpa

#######################################################################

Copyright (c) 1996, Carnegie Mellon University, Cambridge University,

Ronald Rosenfeld and Philip Clarkson

Version 3, Copyright (c) 2006, Carnegie Mellon University

Contributors includes Wen Xu, Ananlada Chotimongkol,

David Huggins-Daines, Arthur Chan and Alan Black

#######################################################################

=============================================================================
=============== This file was produced by the CMU-Cambridge ===============
=============== Statistical Language Modeling Toolkit ===============
=============================================================================
This is a 3-gram language model, based on a vocabulary of 38 words,
which begins "a", "b", "ba"...
This is a CLOSED-vocabulary model
(OOVs eliminated from training data and are forbidden in test data)
Good-Turing discounting was applied.
1-gram frequency of frequency : 1
2-gram frequency of frequency : 1 0 0 0 0 0 0
3-gram frequency of frequency : 1 0 0 0 0 0 0
1-gram discounting ratios : 0.03
2-gram discounting ratios :
3-gram discounting ratios :
This file is in the ARPA-standard format introduced by Doug Paul.

p(wd3|wd1,wd2)= if(trigram exists) p_3(wd1,wd2,wd3)
else if(bigram w1,w2 exists) bo_wt_2(w1,w2)*p(wd3|wd2)
else p(wd3|w2)

p(wd2|wd1)= if(bigram exists) p_2(wd1,wd2)
else bo_wt_1(wd1)*p_1(wd2)

All probs and back-off weights (bo_wt) are given in log10 form.

Data formats:

Beginning of data mark: \data\
ngram 1=nr # number of 1-grams
ngram 2=nr # number of 2-grams
ngram 3=nr # number of 3-grams

\1-grams:
p_1 wd_1 bo_wt_1
\2-grams:
p_2 wd_1 wd_2 bo_wt_2
\3-grams:
p_3 wd_1 wd_2 wd_3

end of data mark: \end\

\data\
ngram 1=38
ngram 2=1
ngram 3=1

\1-grams:
-1.5798 a 0.0000
-1.5798 b 0.0000
-1.5798 ba 0.0000
-1.5798 bảy 0.0000
-1.5798 bốn 0.0000
-1.5798 c 0.0000
-1.5798 chín 0.0000
-1.5798 d 0.0000
-1.5798 e 0.0000
-1.5798 g 0.0000
-1.5798 h 0.0000
-1.5798 hai 0.0000
-1.5798 i 0.0000
-1.5798 k 0.0000
-1.5798 không 0.0000
-1.5798 l 0.0000
-1.5798 m 0.0000
-1.5798 một 0.0000
-1.5798 năm 0.0000
-1.5798 o 0.0000
-1.5798 p 0.0000
-1.5798 q 0.0000
-1.5798 r 0.0000
-1.5798 s 0.0000
-1.5798 sáu 0.0000
-1.5798 t 0.0000
-1.5798 tám 0.0000
-1.5798 u 0.0000
-1.5798 v 0.0000
-1.5798 x 0.0000
-1.5798 y 0.0000
-1.5798 â 0.0000
-1.5798 ê 0.0000
-1.5798 ô 0.0000
-1.5798 ă 0.0000
-1.5798 đ 0.0000
-1.5798 ơ 0.0000
-1.5798 ư -0.4771

\2-grams:
-0.1761 ư <unk> -0.3010 </unk>

\3-grams:
-0.3010 ư <unk> <unk> </unk></unk>

\end\

i create fild DMP

sphinx_lm_convert -i words.arpa -o words.lm.DMP

what did i wrong? thanks

[nguyen duy nam]

i forget:
file dic i make by hand

A A
B B
BA B A
BẢY B AR Y
BỐN B OOS N
C C
CHÍN CH IS N
D D
E E
G G
H H
HAI H A I
I I
K K
KHÔNG K H OO N G
L L
M M
MỘT M OOJ T
NĂM N AW M
O O
P P
Q Q
R R
S S
SÁU S AS U
T T
TÁM T AS M
U U
V V
X X
Y Y
 AA
Ê EE
Ô OO
Ă AW
Đ DD
Ơ OW
Ư UW

file phone i create by:
http://bakuzen.com/extractphoneme.php

A
A
AA
AR
AS
AW
AW
B
B
C
CH
D
DD
E
EE
G
H
H
I
IS
K
K
L
M
M
N
N
O
OO
OO
OOJ
OOS
OW
P
Q
R
S
S
T
T
U
UW
V
X
Y
SIL

file fileids i create by hand

mswav/a
mswav/b
so/ba
so/bay
so/bon
mswav/c
so/chin
mswav/d
mswav/e
mswav/g
mswav/h
so/hai
mswav/i
mswav/k
so/khong
mswav/l
mswav/m
so/mot
so/nam
mswav/o
mswav/p
mswav/q
mswav/r
mswav/s
so/sau
mswav/t
so/tam
mswav/u
mswav/v
mswav/x
mswav/y
mswav/aa
mswav/ee
mswav/oo
mswav/aw
mswav/dd
mswav/ow
mswav/uw

and file transcription i create by hand

A (a)
B (b)
BA (ba)
BẢY (bay)
BỐN (bon)
C (c)
CHÍN (chin)
D (d)
E (e)
G (g)
H (h)
HAI (hai)
I (i)
K (k)
KHÔNG (khong)
L (l)
M (m)
MỘT (mot)
NĂM (nam)
O (o)
P (p)
Q (q)
R (r)
S (s)
SÁU (sau)
T (t)
TÁM (tam)
U (u)
V (v)
X (x)
Y (y)
 (aa)
Ê (ee)
Ô (oo)
Ă (aw)
Đ (dd)
Ơ (ow)
Ư (uw)

thanks

[Nickolay V. Shmyrev]

NFO: main.c(282): Reading: /Application/Data/speechtotext/demo/datatest2/mod
el_parameters/words.cd_cont_untied/mixture_weights
WARN: "s3io.c", line 256: Unable to open /Application/Data/speechtotext/demo/d
atatest2/model_parameters/words.cd_cont_untied/mixture_weights for reading; No
such file or directory
FATAL_ERROR: "main.c", line 771: Initialization failed

where did i wrong? pls

If you want to understand the reason of this error, you need to check all log
files, not just the last one.