1、编译cmuclmtk

http://www.cnblogs.com/libtool: line 479: CDPATH: command not found

http://www.cnblogs.com/libtool: line 1308: func_opt_split: command notfound

libtool: Version mismatch error.  This is libtool 2.2.8,but the

libtool: definition of this LT_INIT comes from an olderrelease.

libtool: You should recreate aclocal.m4 with macros fromlibtool 2.2.8

libtool: and run autoconf again.

make[3]: *** [arpa_bo_ng_prob.lo] error 1

make[3]: Leaving directory`/home/speech/sphinx/cmuclmtk/src/liblmest'

make[2]: *** [all-recursive] error 1

make[2]: Leaving directory`/home/speech/sphinx/cmuclmtk/src'

make[1]: *** [all-recursive] error 1

make[1]: Leaving directory `/home/speech/sphinx/cmuclmtk'

make: *** [all] error 2

解决办法:

1、rm －rf aclocal.m4；

autoreconf

2、将已安装的libtool覆盖到编译目录

./configure

cp/usr/bin/libtool /home/speech/sphinx/cmuclmtk

make

make install

3、cd/usr/local/bin

./sphinx3-simple

………

INFO: srch_time_switch_tree.c(350): Time for buildingtrees, 0.0290 CPU 0.0446 Clk

INFO: srch_time_switch_tree.c(372): Lextrees(0), 1nodes(filler)

INFO: srch_time_switch_tree.c(372): Lextrees(1), 1nodes(filler)

INFO: srch_time_switch_tree.c(372): Lextrees(2), 1nodes(filler)

INFO: vithist.c(168): Initializing Viterbi-history module

INFO:  Initialization of srch_t, report:

INFO:   OperationMode = 4, Operation Name = fwdtree

INFO:

INFO: s3_decode.c(259): Input data will NOT be byteswapped

INFO: s3_decode.c(264): Partial hypothesis WILL be dumped

press ENTER to start recording

2、Soundrecorder不起作用

pocketsphinx_continuous-lm 2916.lm -dict 2916.dic

INFO: ngram_search_fwdtree.c(186):Creating search tree

INFO: ngram_search_fwdtree.c(191): before:0 root, 0 non-root channels, 12 single-phone words

INFO: ngram_search_fwdtree.c(324): after:max nonroot chan increased to 142

INFO: ngram_search_fwdtree.c(333): after:6 root, 14 non-root channels, 11 single-phone words

INFO: ngram_search_fwdflat.c(153):fwdflat: min_ef_width = 4, max_sf_win = 25

INFO: continuous.c(261): pocketsphinx_continuousCOMPILED ON: Feb 18 2011, AT: 10:04:56

FATAL_ERROR:"continuous.c", line 135: cont_ad_calib failed

原因：Soundrecorder不起作用。

SPHINXBASE_EXPORT int32cont_ad_calib (cont_ad_t *cont)

Calibrate the silence filter. 校准静音滤波器

Calibrate the silence filter.

Calibration to determine an initialsilence threshold. This function can be called any number of times. It shouldbe called at least once immediately after cont_ad_init. The silence thresholdis also updated internally once in a while, so this function only needs to becalled in the middle if there is a definite change in the recordingenvironment. The application is responsible for makingsure that the raw audio source is turned on before the calibration.Returnvalue: 0 if successful, <0 otherwise.

Parameters:cont       In: objectpointer returned by cont_ad_init

Definition at line 1022 of filecont_ad_base.c.

References cont_ad_t::ad,cont_ad_t::adbuf, cont_ad_calib(), cont_ad_t::headfrm,cont_ad_t::n_calib_frame, cont_ad_t::n_frm, cont_ad_t::pow_hist,cont_ad_t::spf, and cont_ad_t::thresh_update.

Referenced by cont_ad_calib().

3、vmware+安装ubuntu10.10，soundrecorder使用正常

./pocketsphinxcontinuous

INFO: ngram_search_fwdtree.c(186): Creating searchtree

INFO: ngram_search_fwdtree.c(191): before: 0 root,0 non-root channels, 12 single-phone words

INFO: ngram_search_fwdtree.c(324): after: maxnonroot chan increased to 142

INFO: ngram_search_fwdtree.c(333): after: 6 root,14 non-root channels, 11 single-phone words

INFO: ngram_search_fwdflat.c(153): fwdflat:min_ef_width = 4, max_sf_win = 25

ad_oss.c(103): Failed to open audiodevice(/dev/dsp): No such file or directory

FATAL_ERROR: "continuous.c",line 259: ad_open_dev failed

解决办法：

Sudo apt-get install libasound2-dev

recompiled sphinxbase

…………..

4、my_db.transtription存在“/n”

ODULE: 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 phonelistfile.

Found 16 words using 16 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

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

WARNING: CTL line does not parse correctly:

Phase 4:CTL - Checking number of lines in the transcript should match lines in controlfile

Phase 5:CTL - Determine amount of training data, see if n_tied_states seems reasonable.

Total Hours Training: 0.000789957264957265

Thisis a small amount of data, no comment at this time

Phase 6:TRANSCRIPT - Checking that all the words in the transcript are in thedictionary

Words in dictionary: 13

Words in filler dictionary: 3

Phase 7:TRANSCRIPT - Checking that all the phones in the transcript are in thephonelist, and all phones in the phonelist appear at least once

Something failed:(/sphinx/MyTrain/scripts_pl/00.verify/verify_all.pl)