使用PocketSphinx进行音素识别
我需要在Windows 8桌面上通过麦克风进行实时音素识别。所以我跟着http://cmusphinx.sourceforge.net/wiki/phonemerecognition并在VS2013中从subversion源构建了pocketsphinx_continuous。以管理员身份在命令行中运行它：
D:\_SPHINX\cmusphinx-code-13103-trunk\pocketsphinx\bin\Release\Win32>pocketsphinx_continuous.exe -infile ../../../test/data/goforward.raw -hmm ../../../model/en-us/en-us -allphone ../../../model/en-us/en-us-phone.lm.bin -backtrace yes -beam 1e-20 -pbeam 1e-20 -lw 2.0
INFO: pocketsphinx.c(145): Parsed model-specific feature parameters from ../../../model/en-us/en-us/feat.params
Current configuration:
[NAME] [DEFLT] [VALUE]
-agc none none
-agcthresh 2.0 2.000000e+000
-allphone ../../../model/en-us/en-us-phone.lm.bin
-allphone_ci no no
-alpha 0.97 9.700000e-001
-ascale 20.0 2.000000e+001
-aw 1 1
-backtrace no yes
-beam 1e-48 1.000000e-020
-bestpath yes yes
-bestpathlw 9.5 9.500000e+000
-ceplen 13 13
-cmn current current
-cmninit 8.0 40,3,-1
-compallsen no no
-debug 0
-dict
-dictcase no no
-dither no no
-doublebw no no
-ds 1 1
-fdict ../../../model/en-us/en-us/noisedict
-feat 1s_c_d_dd 1s_c_d_dd
-featparams ../../../model/en-us/en-us/feat.params
-fillprob 1e-8 1.000000e-008
-frate 100 100
-fsg
-fsgusealtpron yes yes
-fsgusefiller yes yes
-fwdflat yes yes
-fwdflatbeam 1e-64 1.000000e-064
-fwdflatefwid 4 4
-fwdflatlw 8.5 8.500000e+000
-fwdflatsfwin 25 25
-fwdflatwbeam 7e-29 7.000000e-029
-fwdtree yes yes
-hmm ../../../model/en-us/en-us
-input_endian little little
-jsgf
-keyphrase
-kws
-kws_delay 10 10
-kws_plp 1e-1 1.000000e-001
-kws_threshold 1 1.000000e+000
-latsize 5000 5000
-lda
-ldadim 0 0
-lifter 0 22
-lm
-lmctl
-lmname
-logbase 1.0001 1.000100e+000
-logfn
-logspec no no
-lowerf 133.33334 1.300000e+002
-lpbeam 1e-40 1.000000e-040
-lponlybeam 7e-29 7.000000e-029
-lw 6.5 2.000000e+000
-maxhmmpf 30000 30000
-maxwpf -1 -1
-mdef ../../../model/en-us/en-us/mdef
-mean ../../../model/en-us/en-us/means
-mfclogdir
-min_endfr 0 0
-mixw
-mixwfloor 0.0000001 1.000000e-007
-mllr
-mmap yes yes
-ncep 13 13
-nfft 512 512
-nfilt 40 25
-nwpen 1.0 1.000000e+000
-pbeam 1e-48 1.000000e-020
-pip 1.0 1.000000e+000
-pl_beam 1e-10 1.000000e-010
-pl_pbeam 1e-10 1.000000e-010
-pl_pip 1.0 1.000000e+000
-pl_weight 3.0 3.000000e+000
-pl_window 5 5
-rawlogdir
-remove_dc no no
-remove_noise yes yes
-remove_silence yes yes
-round_filters yes yes
-samprate 16000 1.600000e+004
-seed -1 -1
-sendump ../../../model/en-us/en-us/sendump
-senlogdir
-senmgau
-silprob 0.005 5.000000e-003
-smoothspec no no
-svspec 0-12/13-25/26-38
-tmat ../../../model/en-us/en-us/transition_matrices
-tmatfloor 0.0001 1.000000e-004
-topn 4 4
-topn_beam 0 0
-toprule
-transform legacy dct
-unit_area yes yes
-upperf 6855.4976 6.800000e+003
-uw 1.0 1.000000e+000
-vad_postspeech 50 50
-vad_prespeech 20 20
-vad_startspeech 10 10
-vad_threshold 2.0 2.000000e+000
-var ../../../model/en-us/en-us/variances
-varfloor 0.0001 1.000000e-004
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wbeam 7e-29 7.000000e-029
-wip 0.65 6.500000e-001
-wlen 0.025625 2.562500e-002
INFO: feat.c(715): Initializing feature stream to type: '1s_c_d_dd', ceplen=13, CMN='current', VARNORM='no', AGC='none'
INFO: cmn.c(143): mean[0]= 12.00, mean[1..12]= 0.0
INFO: acmod.c(164): Using subvector specification 0-12/13-25/26-38
INFO: mdef.c(518): Reading model definition: ../../../model/en-us/en-us/mdef
INFO: mdef.c(531): Found byte-order mark BMDF, assuming this is a binary mdef file
INFO: bin_mdef.c(336): Reading binary model definition: ../../../model/en-us/en-us/mdef
INFO: bin_mdef.c(516): 42 CI-phone, 137053 CD-phone, 3 emitstate/phone, 126 CI-sen, 5126 Sen, 29324 Sen-Seq
INFO: tmat.c(206): Reading HMM transition probability matrices: ../../../model/en-us/en-us/transition_matrices
在最后一个INFO行，Windows 8抛出此错误：
PocketSphinx调试输出或命令行选项有什么问题吗？或者它是一个纯粹的Windows问题？我注意到了这个文件夹：/ bin / Release / Win32。我的Windows 8在Intel NUC上是64位。 Sphinxbase.dll是在调试模式下从subversion编译的，而PacketSphinx只有Release模式。
另外，我在某处读到了音素定时信息可用 - 如何获取？
补充：遵循Nikolay的建议，使用这些参数，我消除了错误，但没有音素：
D:\_SPHINX\pocketsphinx\bin\Debug>pocketsphinx_continuous.exe -infile ../../test/data/goforward.raw -hmm ../../model/en-us/en-us -allphone ../../model/en-us/en-us.lm.dmp -backtrace yes -beam 1e-20 -pbeam 1e-20 -lw 2.0 -debug 3 -verbose yes
INFO: cmd_ln.c(697): Parsing command line:
pocketsphinx_continuous.exe \
-infile ../../test/data/goforward.raw \
-hmm ../../model/en-us/en-us \
-allphone ../../model/en-us/en-us.lm.dmp \
-backtrace yes \
-beam 1e-20 \
-pbeam 1e-20 \
-lw 2.0 \
-debug 3 \
-verbose yes
. . . .
INFO: acmod.c(252): Parsed model-specific feature parameters from ../../model/en-us/en-us/feat.params
INFO: fe_interface.c(177): Current FE Parameters:
INFO: fe_interface.c(178): Sampling Rate: 16000.000000
INFO: fe_interface.c(179): Frame Size: 410
INFO: fe_interface.c(180): Frame Shift: 160
INFO: fe_interface.c(181): FFT Size: 512
INFO: fe_interface.c(183): Lower Frequency: 130
INFO: fe_interface.c(185): Upper Frequency: 6800
INFO: fe_interface.c(186): Number of filters: 25
INFO: fe_interface.c(187): Number of Overflow Samps: 0
INFO: fe_interface.c(188): Start Utt Status: 0
INFO: fe_interface.c(190): Will not remove DC offset at frame level
INFO: fe_interface.c(196): Will not add dither to audio
INFO: fe_interface.c(200): Will apply sine-curve liftering, period 22
INFO: fe_interface.c(203): Will normalize filters to unit area
INFO: fe_interface.c(205): Will round filter frequencies to DFT points
INFO: fe_interface.c(207): Will not use double bandwidth in mel filter
INFO: feat.c(715): Initializing feature stream to type: '1s_c_d_dd', ceplen=13, CMN='current', VARNORM='no', AGC='none'
INFO: cmn.c(143): mean[0]= 12.00, mean[1..12]= 0.0
INFO: acmod.c(171): Using subvector specification 0-12/13-25/26-38
INFO: mdef.c(518): Reading model definition: ../../model/en-us/en-us/mdef
INFO: mdef.c(531): Found byte-order mark BMDF, assuming this is a binary mdef file
INFO: bin_mdef.c(336): Reading binary model definition: ../../model/en-us/en-us/mdef
INFO: bin_mdef.c(516): 42 CI-phone, 137053 CD-phone, 3 emitstate/phone, 126 CI-sen, 5126 Sen, 29324 Sen-Seq
INFO: tmat.c(206): Reading HMM transition probability matrices: ../../model/en-us/en-us/transition_matrices
INFO: acmod.c(124): Attempting to use PTM computation module
INFO: ms_gauden.c(198): Reading mixture gaussian parameter: ../../model/en-us/en-us/means
INFO: ms_gauden.c(292): 42 codebook, 3 feature, size:
INFO: ms_gauden.c(294): 128x13
INFO: ms_gauden.c(294): 128x13
INFO: ms_gauden.c(294): 128x13
INFO: ms_gauden.c(198): Reading mixture gaussian parameter: ../../model/en-us/en-us/variances
INFO: ms_gauden.c(292): 42 codebook, 3 feature, size:
INFO: ms_gauden.c(294): 128x13
INFO: ms_gauden.c(294): 128x13
INFO: ms_gauden.c(294): 128x13
INFO: ms_gauden.c(354): 222 variance values floored
INFO: ptm_mgau.c(476): Loading senones from dump file ../../model/en-us/en-us/sendump
INFO: ptm_mgau.c(500): BEGIN FILE FORMAT DESCRIPTION
INFO: ptm_mgau.c(563): Rows: 128, Columns: 5126
INFO: ptm_mgau.c(595): Using memory-mapped I/O for senones
INFO: ptm_mgau.c(835): Maximum top-N: 4
INFO: phone_loop_search.c(115): State beam -225 Phone exit beam -225 Insertion penalty 0
INFO: dict.c(320): Allocating 4101 * 20 bytes (80 KiB) for word entries
INFO: dict.c(342): Reading filler dictionary: ../../model/en-us/en-us/noisedict
INFO: dict.c(213): Allocated 0 KiB for strings, 0 KiB for phones
INFO: dict.c(345): 5 words read
INFO: dict2pid.c(396): Building PID tables for dictionary
INFO: dict2pid.c(406): Allocating 42^3 * 2 bytes (144 KiB) for word-initial triphones
INFO: dict2pid.c(132): Allocated 21336 bytes (20 KiB) for word-final triphones
INFO: dict2pid.c(196): Allocated 21336 bytes (20 KiB) for single-phone word triphones
INFO: ngram_model_arpa.c(77): No \data\ mark in LM file
INFO: ngram_model_dmp.c(142): Will use memory-mapped I/O for LM file
INFO: ngram_model_dmp.c(196): ngrams 1=19794, 2=1377200, 3=3178194
INFO: ngram_model_dmp.c(242): 19794 = LM.unigrams(+trailer) read
INFO: ngram_model_dmp.c(288): 1377200 = LM.bigrams(+trailer) read
INFO: ngram_model_dmp.c(314): 3178194 = LM.trigrams read
INFO: ngram_model_dmp.c(339): 57155 = LM.prob2 entries read
INFO: ngram_model_dmp.c(359): 10935 = LM.bo_wt2 entries read
INFO: ngram_model_dmp.c(379): 34843 = LM.prob3 entries read
INFO: ngram_model_dmp.c(407): 2690 = LM.tseg_base entries read
INFO: ngram_model_dmp.c(463): 19794 = ascii word strings read
INFO: allphone_search.c(239): Building PHMM net of 137095 phones
INFO: allphone_search.c(312): 29324 nodes, 1958591 links
INFO: allphone_search.c(611): Allphone(beam: -450, pbeam: -450)
INFO: continuous.c(299): pocketsphinx_continuous.exe COMPILED ON: Aug 23 2015, AT: 14:00:33
INFO: cmn_prior.c(131): cmn_prior_update: from < 40.00 3.00 -1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 >
INFO: cmn_prior.c(149): cmn_prior_update: to < 44.50 -4.13 0.15 6.94 4.06 -5.38 -2.56 -3.13 -6.12 -1.20 -7.44 -2.25 0.48 >
INFO: allphone_search.c(852): 214 frames, 214 HMMs (1/fr), 642 senones (3/fr), 214 history entries (1/fr)
INFO: allphone_search.c(865): allphone 0.61 CPU 0.283 xRT
INFO: allphone_search.c(867): allphone 0.62 wall 0.290 xRT
INFO: allphone_search.c(911): Hyp: SIL
INFO: pocketsphinx.c(1133): SIL (-858993460)
word start end pprob ascr lscr lback
SIL 51 264 1.000 -1627 0 0
INFO: allphone_search.c(911): Hyp: SIL
SIL
INFO: cmn_prior.c(131): cmn_prior_update: from < 44.50 -4.13 0.15 6.94 4.06 -5.38 -2.56 -3.13 -6.12 -1.20 -7.44 -2.25 0.48 >
INFO: cmn_prior.c(149): cmn_prior_update: to < 44.50 -4.13 0.15 6.94 4.06 -5.38 -2.56 -3.13 -6.12 -1.20 -7.44 -2.25 0.48 >
INFO: allphone_search.c(852): 0 frames, 0 HMMs (0/fr), 0 senones (0/fr), 0 history entries (0/fr)
INFO: allphone_search.c(651): TOTAL fwdflat 0.61 CPU 0.285 xRT
INFO: allphone_search.c(654): TOTAL fwdflat 0.64 wall 0.298 xRT
获取音素输出的正确命令行参数集是什么？
0 个答案:
没有答案
相关问题
最新问题
- 我写了这段代码，但我无法理解我的错误
- 我无法从一个代码实例的列表中删除 None 值，但我可以在另一个实例中。为什么它适用于一个细分市场而不适用于另一个细分市场？
- 是否有可能使 loadstring 不可能等于打印？卢阿
- java中的random.expovariate()
- Appscript 通过会议在 Google 日历中发送电子邮件和创建活动
- 为什么我的 Onclick 箭头功能在 React 中不起作用？
- 在此代码中是否有使用“this”的替代方法？
- 在 SQL Server 和 PostgreSQL 上查询，我如何从第一个表获得第二个表的可视化
- 每千个数字得到
- 更新了城市边界 KML 文件的来源？