为什么 pocketsphinx 通过 java 和 kws 返回零假设？通过命令行而不是通过代码工作
Why is pocketsphinx returning a null hypothesis via java with kws? Works via commandline, not via code
我一直在 java 中使用 pocketsphinx。我从各种来源拼凑了这个。
正在尝试通过 pocketsphinx 进行关键字检测。
正如我所说的通过命令行工作：
pocketsphinx_continuous -inmic yes -kws keyphrase.list
其中 keyphrase.list 文件包含：\
abomination /le-20/
我每次都中招。
这是我的 java 代码：
（我试过 le-1 到 le-40）
import javax.sound.sampled.AudioFormat;
import javax.sound.sampled.AudioSystem;
import javax.sound.sampled.DataLine;
import javax.sound.sampled.TargetDataLine;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.nio.ByteBuffer;
import java.nio.ByteOrder;
import java.util.Arrays;
import edu.cmu.pocketsphinx.Decoder;
import edu.cmu.pocketsphinx.Config;
import edu.cmu.pocketsphinx.Hypothesis;
public class Controller {
static {
System.loadLibrary("pocketsphinx_jni");
}
private static ByteArrayOutputStream out;
public static void main(String args[]) {
AudioFormat format = new AudioFormat(44100, 16, 1, true, true);
TargetDataLine targetLine = null;
DataLine.Info targetInfo = new DataLine.Info(TargetDataLine.class, format);
boolean running = true;
try {
targetLine = AudioSystem.getTargetDataLine(format);
targetLine.open();
out = new ByteArrayOutputStream();
int numBytesRead;
byte[] data = new byte[targetLine.getBufferSize() / 5];
Config c = Decoder.defaultConfig();
c.setString("-hmm", "/usr/local/share/pocketsphinx/model/en-us/en-us/");
//c.setString("-lm", "/usr/local/share/pocketsphinx/model/en-us/en-us.lm.bin");
c.setString("-dict", "/usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict");
c.setString("-keyphrase", "abomination");
c.setFloat("-kws_threshold", 1e-1);
Decoder d = new Decoder(c);
d.setRawdataSize(300000);
targetLine.start();
System.out.println("Recorder started");
byte[] b = new byte[4096];
d.startUtt();
System.out.println("Decoder started");
while ((running)) {
int nbytes;
short[] s = null;
nbytes = targetLine.read(b,0,b.length);
ByteBuffer bb = ByteBuffer.wrap(b, 0, nbytes);
s = new short[nbytes/2];
bb.asShortBuffer().get(s);
d.processRaw(s, nbytes/2, false, false);
d.setKws("abomination", );
if (nbytes > 0) {
Hypothesis hypothesis = d.hyp();
if (hypothesis != null) {
System.out.println("------------------------------------------------------");
System.out.println(hypothesis.getHypstr());
System.out.println("------------------------------------------------------");
d.endUtt();
d.startUtt();
}
}
}
}
catch (Exception e) {
System.err.println(e);
}
}
}
代码 运行 没问题。如果 if(hypothesis != null) 由于某种原因永远不会进入。
这是记录的内容：
INFO: cmn_prior.c(99): 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(116): cmn_prior_update: to < 51.22 14.61 -8.72 -0.31 -3.49 0.18 -7.35 8.43 -0.77 7.64 1.41 0.27 -1.82 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 51.22 14.61 -8.72 -0.31 -3.49 0.18 -7.35 8.43 -0.77 7.64 1.41 0.27 -1.82 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 51.80 15.37 -8.77 -0.62 -2.74 0.09 -6.18 10.24 0.14 7.79 2.59 1.86 -3.22 >
更新：更多信息。
这是运行时的输出。 -kws 未设置。
INFO: pocketsphinx.c(152): Parsed model-specific feature parameters from /usr/local/share/pocketsphinx/model/en-us/en-us//feat.params
Current configuration:
[NAME] [DEFLT] [VALUE]
-agc none none
-agcthresh 2.0 2.000000e+00
-allphone
-allphone_ci no no
-alpha 0.97 9.700000e-01
-ascale 20.0 2.000000e+01
-aw 1 1
-backtrace no no
-beam 1e-48 1.000000e-48
-bestpath yes yes
-bestpathlw 9.5 9.500000e+00
-ceplen 13 13
-cmn current current
-cmninit 8.0 40,3,-1
-compallsen no no
-debug 0
-dict /usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict
-dictcase no no
-dither no no
-doublebw no no
-ds 1 1
-fdict
-feat 1s_c_d_dd 1s_c_d_dd
-featparams
-fillprob 1e-8 1.000000e-08
-frate 100 100
-fsg
-fsgusealtpron yes yes
-fsgusefiller yes yes
-fwdflat yes yes
-fwdflatbeam 1e-64 1.000000e-64
-fwdflatefwid 4 4
-fwdflatlw 8.5 8.500000e+00
-fwdflatsfwin 25 25
-fwdflatwbeam 7e-29 7.000000e-29
-fwdtree yes yes
-hmm /usr/local/share/pocketsphinx/model/en-us/en-us/
-input_endian little little
-jsgf
-keyphrase abomination
-kws
-kws_delay 10 10
-kws_plp 1e-1 1.000000e-01
-kws_threshold 1 1.000000e-20
-latsize 5000 5000
-lda
-ldadim 0 0
-lifter 0 22
-lm
-lmctl
-lmname
-logbase 1.0001 1.000100e+00
-logfn
-logspec no no
-lowerf 133.33334 1.300000e+02
-lpbeam 1e-40 1.000000e-40
-lponlybeam 7e-29 7.000000e-29
-lw 6.5 6.500000e+00
-maxhmmpf 30000 30000
-maxwpf -1 -1
-mdef
-mean
-mfclogdir
-min_endfr 0 0
-mixw
-mixwfloor 0.0000001 1.000000e-07
-mllr
-mmap yes yes
-ncep 13 13
-nfft 512 512
-nfilt 40 25
-nwpen 1.0 1.000000e+00
-pbeam 1e-48 1.000000e-48
-pip 1.0 1.000000e+00
-pl_beam 1e-10 1.000000e-10
-pl_pbeam 1e-10 1.000000e-10
-pl_pip 1.0 1.000000e+00
-pl_weight 3.0 3.000000e+00
-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+04
-seed -1 -1
-sendump
-senlogdir
-senmgau
-silprob 0.005 5.000000e-03
-smoothspec no no
-svspec 0-12/13-25/26-38
-tmat
-tmatfloor 0.0001 1.000000e-04
-topn 4 4
-topn_beam 0 0
-toprule
-transform legacy dct
-unit_area yes yes
-upperf 6855.4976 6.800000e+03
-uw 1.0 1.000000e+00
-vad_postspeech 50 50
-vad_prespeech 20 20
-vad_startspeech 10 10
-vad_threshold 2.0 2.000000e+00
-var
-varfloor 0.0001 1.000000e-04
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wbeam 7e-29 7.000000e-29
-wip 0.65 6.500000e-01
-wlen 0.025625 2.562500e-02
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: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/model/en-us/en-us//transition_matrices
INFO: acmod.c(117): Attempting to use PTM computation module
INFO: ms_gauden.c(198): Reading mixture gaussian parameter: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/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 /usr/local/share/pocketsphinx/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(114): State beam -225 Phone exit beam -225 Insertion penalty 0
INFO: dict.c(320): Allocating 138623 * 32 bytes (4331 KiB) for word entries
INFO: dict.c(333): Reading main dictionary: /usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict
INFO: dict.c(213): Allocated 1014 KiB for strings, 1677 KiB for phones
INFO: dict.c(336): 134522 words read
INFO: dict.c(358): Reading filler dictionary: /usr/local/share/pocketsphinx/model/en-us/en-us//noisedict
INFO: dict.c(213): Allocated 0 KiB for strings, 0 KiB for phones
INFO: dict.c(361): 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 42672 bytes (41 KiB) for word-final triphones
INFO: dict2pid.c(196): Allocated 42672 bytes (41 KiB) for single-phone word triphones
INFO: kws_search.c(420): KWS(beam: -1080, plp: -23, default threshold -450, delay 10)
Recorder started
Decoder started
我在此处找到了为什么您不需要 -lm 行的参考资料。
Also if you intend to use kws there is no need to use -lm in arguments. You need to remove:
"-lm", ".../model/hub4wsj_sc_8k_adapt/etc/hub4.5000.DMP",
这就是答案。
如果我更改上面的代码并删除：
c.setString("-keyphrase", "abomination");
并添加：
c.setString("-kws", "/home/pennyworth/keyphrase.list");
现在输出显示 -kws set
-kws /home/pennyworth/keyphrase.list
我在输出中得到了这个：
INFO: kws_search.c(420): KWS(beam: -1080, plp: -23, default threshold -450, delay 10)
尽管如此，结果为空。
我
NFO: cmn_prior.c(99): cmn_prior_update: from < 73.10 11.10 -10.49 1.23 0.67 -1.37 -5.29 5.17 -0.62 3.91 -0.28 2.56 -2.14 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 73.59 11.36 -9.61 2.41 2.13 0.15 -5.09 4.30 1.03 4.46 -0.13 3.36 -0.62 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 73.59 11.36 -9.61 2.41 2.13 0.15 -5.09 4.30 1.03 4.46 -0.13 3.36 -0.62 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 74.65 10.58 -9.76 4.47 3.63 1.19 -5.20 3.74 2.33 4.75 -0.11 3.06 -0.32 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 74.65 10.58 -9.76 4.47 3.63 1.19 -5.20 3.74 2.33 4.75 -0.11 3.06 -0.32 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 77.49 10.99 -8.80 5.45 4.37 2.83 -4.14 4.06 3.48 5.07 -0.41 2.82 -0.35 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 77.49 10.99 -8.80 5.45 4.37 2.83 -4.14 4.06 3.48 5.07 -0.41 2.82 -0.35 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 73.54 9.62 -11.34 3.19 3.30 2.24 -6.61 4.52 1.31 5.99 -1.28 2.24 -0.96 >
我是否假设 kws 不是 return 炒作？有一个很棒的 python 示例，但是关于 kws 的 java 没有什么用。
https://github.com/cmusphinx/pocketsphinx/blob/master/swig/python/test/kws_test.py
这是 pocketsphinx 的 api 文档，http://cmusphinx.sourceforge.net/doc/pocketsphinx/pocketsphinx_8c_source.html
我不知道如何推进这个。我要么没有正确设置解码器，要么发生了其他事情，这就是我得到 null return 的原因。
我不清楚 -kws 与 -keyphrase vx -kws-threshold 的用法。使用 -kws 是否意味着您不需要其他两个，因为它有效地设置了短语和阈值？
更新代码。添加了字节顺序。 （并确保我的 1 不是 l ）
import javax.sound.sampled.AudioFormat;
import javax.sound.sampled.AudioSystem;
import javax.sound.sampled.DataLine;
import javax.sound.sampled.TargetDataLine;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.nio.ByteBuffer;
import java.nio.ByteOrder;
import java.util.Arrays;
import edu.cmu.pocketsphinx.Decoder;
import edu.cmu.pocketsphinx.Config;
import edu.cmu.pocketsphinx.Hypothesis;
public class Controller {
static {
System.loadLibrary("pocketsphinx_jni");
}
private static ByteArrayOutputStream out;
public static void main(String args[]) {
AudioFormat format = new AudioFormat(44100, 16, 1, true, true);
TargetDataLine targetLine = null;
DataLine.Info targetInfo = new DataLine.Info(TargetDataLine.class, format);
boolean running = true;
try {
targetLine = AudioSystem.getTargetDataLine(format);
targetLine.open();
out = new ByteArrayOutputStream();
int numBytesRead;
byte[] data = new byte[targetLine.getBufferSize() / 5];
Config c = Decoder.defaultConfig();
c.setString("-hmm", "/usr/local/share/pocketsphinx/model/en-us/en-us/");
c.setString("-dict", "/usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict");
c.setString("-keyphrase", "abomination");
c.setFloat("-kws_threshold", 1e-20);
//c.setString("-kws", "/home/bruce/keyphrase.list");
Decoder d = new Decoder(c);
d.setRawdataSize(300000);
targetLine.start();
System.out.println("Recorder started");
byte[] b = new byte[4096];
d.startUtt();
System.out.println("Decoder started");
while ((running)) {
int nbytes;
short[] s = null;
nbytes = targetLine.read(b,0,b.length);
ByteBuffer bb = ByteBuffer.wrap(b, 0, nbytes);
s = new short[nbytes/2];
bb.asShortBuffer().get(s);
bb.order(ByteOrder.LITTLE_ENDIAN);
d.processRaw(s, nbytes/2, false, false);
if (nbytes > 0) {
Hypothesis hypothesis = d.hyp();
if (hypothesis != null) {
System.out.println("------------------------------------------------------");
System.out.println(hypothesis.getHypstr());
System.out.println("------------------------------------------------------");
d.endUtt();
d.startUtt();
}
}
}
}
catch (Exception e) {
System.err.println(e);
}
}
}
这是更新后的输出：
INFO: cmn_prior.c(99): cmn_prior_update: from < 67.54 12.23 -8.09 -0.29 0.56 -0.37 -3.25 7.00 -1.97 3.98 -1.87 3.63 -1.49 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 66.41 12.98 -8.67 -0.63 1.35 -0.13 -3.16 7.97 -3.11 3.57 -0.74 3.91 -1.79 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 66.41 12.98 -8.67 -0.63 1.35 -0.13 -3.16 7.97 -3.11 3.57 -0.74 3.91 -1.79 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 66.66 12.12 -10.32 -1.55 0.57 -0.01 -3.20 8.83 -2.47 4.65 0.07 4.54 -2.35 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 66.66 12.12 -10.32 -1.55 0.57 -0.01 -3.20 8.83 -2.47 4.65 0.07 4.54 -2.35 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 68.35 12.82 -9.91 -1.85 0.77 0.18 -2.25 9.05 -1.75 3.84 0.66 5.82 -2.50 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 68.35 12.82 -9.91 -1.85 0.77 0.18 -2.25 9.05 -1.75 3.84 0.66 5.82 -2.50 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 64.05 14.55 -8.14 -0.36 0.64 0.75 -1.96 9.76 -0.03 5.26 1.16 5.03 -1.68 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 64.05 14.55 -8.14 -0.36 0.64 0.75 -1.96 9.76 -0.03 5.26 1.16 5.03 -1.68 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 63.08 15.30 -8.96 0.76 1.05 0.83 -1.40 10.94 -0.69 4.52 -0.80 3.58 -3.18 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 63.08 15.30 -8.96 0.76 1.05 0.83 -1.40 10.94 -0.69 4.52 -0.80 3.58 -3.18 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 62.15 16.49 -10.32 -0.25 1.14 -0.32 -2.32 10.95 -2.12 2.91 -1.31 2.57 -4.05
更新：
更改了以下顺序以反映 https://github.com/cmusphinx/pocketsphinx/blob/master/swig/java/test/DecoderTest.java
bb.order(ByteOrder.LITTLE_ENDIAN);
bb.asShortBuffer().get(s);
一旦启动，就好像没有输入一样，完全没有输出。
INFO: pocketsphinx.c(152): Parsed model-specific feature parameters from /usr/local/share/pocketsphinx/model/en-us/en-us//feat.params
Current configuration:
[NAME] [DEFLT] [VALUE]
-agc none none
-agcthresh 2.0 2.000000e+00
-allphone
-allphone_ci no no
-alpha 0.97 9.700000e-01
-ascale 20.0 2.000000e+01
-aw 1 1
-backtrace no no
-beam 1e-48 1.000000e-48
-bestpath yes yes
-bestpathlw 9.5 9.500000e+00
-ceplen 13 13
-cmn current current
-cmninit 8.0 40,3,-1
-compallsen no no
-debug 0
-dict /usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict
-dictcase no no
-dither no no
-doublebw no no
-ds 1 1
-fdict
-feat 1s_c_d_dd 1s_c_d_dd
-featparams
-fillprob 1e-8 1.000000e-08
-frate 100 100
-fsg
-fsgusealtpron yes yes
-fsgusefiller yes yes
-fwdflat yes yes
-fwdflatbeam 1e-64 1.000000e-64
-fwdflatefwid 4 4
-fwdflatlw 8.5 8.500000e+00
-fwdflatsfwin 25 25
-fwdflatwbeam 7e-29 7.000000e-29
-fwdtree yes yes
-hmm /usr/local/share/pocketsphinx/model/en-us/en-us/
-input_endian little little
-jsgf
-keyphrase abomination
-kws
-kws_delay 10 10
-kws_plp 1e-1 1.000000e-01
-kws_threshold 1 1.000000e-20
-latsize 5000 5000
-lda
-ldadim 0 0
-lifter 0 22
-lm
-lmctl
-lmname
-logbase 1.0001 1.000100e+00
-logfn
-logspec no no
-lowerf 133.33334 1.300000e+02
-lpbeam 1e-40 1.000000e-40
-lponlybeam 7e-29 7.000000e-29
-lw 6.5 6.500000e+00
-maxhmmpf 30000 30000
-maxwpf -1 -1
-mdef
-mean
-mfclogdir
-min_endfr 0 0
-mixw
-mixwfloor 0.0000001 1.000000e-07
-mllr
-mmap yes yes
-ncep 13 13
-nfft 512 512
-nfilt 40 25
-nwpen 1.0 1.000000e+00
-pbeam 1e-48 1.000000e-48
-pip 1.0 1.000000e+00
-pl_beam 1e-10 1.000000e-10
-pl_pbeam 1e-10 1.000000e-10
-pl_pip 1.0 1.000000e+00
-pl_weight 3.0 3.000000e+00
-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+04
-seed -1 -1
-sendump
-senlogdir
-senmgau
-silprob 0.005 5.000000e-03
-smoothspec no no
-svspec 0-12/13-25/26-38
-tmat
-tmatfloor 0.0001 1.000000e-04
-topn 4 4
-topn_beam 0 0
-toprule
-transform legacy dct
-unit_area yes yes
-upperf 6855.4976 6.800000e+03
-uw 1.0 1.000000e+00
-vad_postspeech 50 50
-vad_prespeech 20 20
-vad_startspeech 10 10
-vad_threshold 2.0 2.000000e+00
-var
-varfloor 0.0001 1.000000e-04
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wbeam 7e-29 7.000000e-29
-wip 0.65 6.500000e-01
-wlen 0.025625 2.562500e-02
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: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/model/en-us/en-us//transition_matrices
INFO: acmod.c(117): Attempting to use PTM computation module
INFO: ms_gauden.c(198): Reading mixture gaussian parameter: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/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 /usr/local/share/pocketsphinx/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(114): State beam -225 Phone exit beam -225 Insertion penalty 0
INFO: dict.c(320): Allocating 138623 * 32 bytes (4331 KiB) for word entries
INFO: dict.c(333): Reading main dictionary: /usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict
INFO: dict.c(213): Allocated 1014 KiB for strings, 1677 KiB for phones
INFO: dict.c(336): 134522 words read
INFO: dict.c(358): Reading filler dictionary: /usr/local/share/pocketsphinx/model/en-us/en-us//noisedict
INFO: dict.c(213): Allocated 0 KiB for strings, 0 KiB for phones
INFO: dict.c(361): 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
Recorder started
Decoder started
INFO: dict2pid.c(132): Allocated 42672 bytes (41 KiB) for word-final triphones
INFO: dict2pid.c(196): Allocated 42672 bytes (41 KiB) for single-phone word triphones
INFO: kws_search.c(420): KWS(beam: -1080, plp: -23, default threshold -450, delay 10)
abomination /le-20/
每个开发人员都应该知道浮点格式，它以数字开头 - 1e-1 与 0.1 相同，5e-3 与 0.005 相同。 le-20 和 l 不是浮点数。
d.processRaw(s, nbytes/2, false, false);
d.setKws("abomination", );
此代码不应编译。此外，您不应在处理过程中设置搜索。
I'm not clear on the use of -kws vs -keyphrase vx -kws-threshold. Does using -kws mean you don't need the other two since it effectively sets both the phrase and threshold?
是，关键字列表替换命令行选项
Am I assuming wrong that kws doesn't return a hyp? T
这不是 return 假设，因为您没有将正确的音频传递到解码器。你需要
bb.order(ByteOrder.LITTLE_ENDIAN);`
没有它，它会传递大端数据（java 默认值）。来源中的 Swig 示例有。
INFO: cmn_prior.c(99): cmn_prior_update: from < 73.10 11.10 -10.49 1.23 0.67 -1.37 -5.29 5.17 -0.62 3.91 -0.28 2.56 -2.14 >
CMN 范围为 70 表示端序问题。使用正确的字节序，第一个 CMN 值应该在 30 到 60 之间。
我一直在 java 中使用 pocketsphinx。我从各种来源拼凑了这个。
正在尝试通过 pocketsphinx 进行关键字检测。
正如我所说的通过命令行工作：
pocketsphinx_continuous -inmic yes -kws keyphrase.list
其中 keyphrase.list 文件包含：\
abomination /le-20/
我每次都中招。
这是我的 java 代码：
（我试过 le-1 到 le-40）
import javax.sound.sampled.AudioFormat;
import javax.sound.sampled.AudioSystem;
import javax.sound.sampled.DataLine;
import javax.sound.sampled.TargetDataLine;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.nio.ByteBuffer;
import java.nio.ByteOrder;
import java.util.Arrays;
import edu.cmu.pocketsphinx.Decoder;
import edu.cmu.pocketsphinx.Config;
import edu.cmu.pocketsphinx.Hypothesis;
public class Controller {
static {
System.loadLibrary("pocketsphinx_jni");
}
private static ByteArrayOutputStream out;
public static void main(String args[]) {
AudioFormat format = new AudioFormat(44100, 16, 1, true, true);
TargetDataLine targetLine = null;
DataLine.Info targetInfo = new DataLine.Info(TargetDataLine.class, format);
boolean running = true;
try {
targetLine = AudioSystem.getTargetDataLine(format);
targetLine.open();
out = new ByteArrayOutputStream();
int numBytesRead;
byte[] data = new byte[targetLine.getBufferSize() / 5];
Config c = Decoder.defaultConfig();
c.setString("-hmm", "/usr/local/share/pocketsphinx/model/en-us/en-us/");
//c.setString("-lm", "/usr/local/share/pocketsphinx/model/en-us/en-us.lm.bin");
c.setString("-dict", "/usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict");
c.setString("-keyphrase", "abomination");
c.setFloat("-kws_threshold", 1e-1);
Decoder d = new Decoder(c);
d.setRawdataSize(300000);
targetLine.start();
System.out.println("Recorder started");
byte[] b = new byte[4096];
d.startUtt();
System.out.println("Decoder started");
while ((running)) {
int nbytes;
short[] s = null;
nbytes = targetLine.read(b,0,b.length);
ByteBuffer bb = ByteBuffer.wrap(b, 0, nbytes);
s = new short[nbytes/2];
bb.asShortBuffer().get(s);
d.processRaw(s, nbytes/2, false, false);
d.setKws("abomination", );
if (nbytes > 0) {
Hypothesis hypothesis = d.hyp();
if (hypothesis != null) {
System.out.println("------------------------------------------------------");
System.out.println(hypothesis.getHypstr());
System.out.println("------------------------------------------------------");
d.endUtt();
d.startUtt();
}
}
}
}
catch (Exception e) {
System.err.println(e);
}
}
}
代码 运行 没问题。如果 if(hypothesis != null) 由于某种原因永远不会进入。
这是记录的内容：
INFO: cmn_prior.c(99): 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(116): cmn_prior_update: to < 51.22 14.61 -8.72 -0.31 -3.49 0.18 -7.35 8.43 -0.77 7.64 1.41 0.27 -1.82 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 51.22 14.61 -8.72 -0.31 -3.49 0.18 -7.35 8.43 -0.77 7.64 1.41 0.27 -1.82 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 51.80 15.37 -8.77 -0.62 -2.74 0.09 -6.18 10.24 0.14 7.79 2.59 1.86 -3.22 >
更新：更多信息。
这是运行时的输出。 -kws 未设置。
INFO: pocketsphinx.c(152): Parsed model-specific feature parameters from /usr/local/share/pocketsphinx/model/en-us/en-us//feat.params
Current configuration:
[NAME] [DEFLT] [VALUE]
-agc none none
-agcthresh 2.0 2.000000e+00
-allphone
-allphone_ci no no
-alpha 0.97 9.700000e-01
-ascale 20.0 2.000000e+01
-aw 1 1
-backtrace no no
-beam 1e-48 1.000000e-48
-bestpath yes yes
-bestpathlw 9.5 9.500000e+00
-ceplen 13 13
-cmn current current
-cmninit 8.0 40,3,-1
-compallsen no no
-debug 0
-dict /usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict
-dictcase no no
-dither no no
-doublebw no no
-ds 1 1
-fdict
-feat 1s_c_d_dd 1s_c_d_dd
-featparams
-fillprob 1e-8 1.000000e-08
-frate 100 100
-fsg
-fsgusealtpron yes yes
-fsgusefiller yes yes
-fwdflat yes yes
-fwdflatbeam 1e-64 1.000000e-64
-fwdflatefwid 4 4
-fwdflatlw 8.5 8.500000e+00
-fwdflatsfwin 25 25
-fwdflatwbeam 7e-29 7.000000e-29
-fwdtree yes yes
-hmm /usr/local/share/pocketsphinx/model/en-us/en-us/
-input_endian little little
-jsgf
-keyphrase abomination
-kws
-kws_delay 10 10
-kws_plp 1e-1 1.000000e-01
-kws_threshold 1 1.000000e-20
-latsize 5000 5000
-lda
-ldadim 0 0
-lifter 0 22
-lm
-lmctl
-lmname
-logbase 1.0001 1.000100e+00
-logfn
-logspec no no
-lowerf 133.33334 1.300000e+02
-lpbeam 1e-40 1.000000e-40
-lponlybeam 7e-29 7.000000e-29
-lw 6.5 6.500000e+00
-maxhmmpf 30000 30000
-maxwpf -1 -1
-mdef
-mean
-mfclogdir
-min_endfr 0 0
-mixw
-mixwfloor 0.0000001 1.000000e-07
-mllr
-mmap yes yes
-ncep 13 13
-nfft 512 512
-nfilt 40 25
-nwpen 1.0 1.000000e+00
-pbeam 1e-48 1.000000e-48
-pip 1.0 1.000000e+00
-pl_beam 1e-10 1.000000e-10
-pl_pbeam 1e-10 1.000000e-10
-pl_pip 1.0 1.000000e+00
-pl_weight 3.0 3.000000e+00
-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+04
-seed -1 -1
-sendump
-senlogdir
-senmgau
-silprob 0.005 5.000000e-03
-smoothspec no no
-svspec 0-12/13-25/26-38
-tmat
-tmatfloor 0.0001 1.000000e-04
-topn 4 4
-topn_beam 0 0
-toprule
-transform legacy dct
-unit_area yes yes
-upperf 6855.4976 6.800000e+03
-uw 1.0 1.000000e+00
-vad_postspeech 50 50
-vad_prespeech 20 20
-vad_startspeech 10 10
-vad_threshold 2.0 2.000000e+00
-var
-varfloor 0.0001 1.000000e-04
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wbeam 7e-29 7.000000e-29
-wip 0.65 6.500000e-01
-wlen 0.025625 2.562500e-02
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: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/model/en-us/en-us//transition_matrices
INFO: acmod.c(117): Attempting to use PTM computation module
INFO: ms_gauden.c(198): Reading mixture gaussian parameter: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/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 /usr/local/share/pocketsphinx/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(114): State beam -225 Phone exit beam -225 Insertion penalty 0
INFO: dict.c(320): Allocating 138623 * 32 bytes (4331 KiB) for word entries
INFO: dict.c(333): Reading main dictionary: /usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict
INFO: dict.c(213): Allocated 1014 KiB for strings, 1677 KiB for phones
INFO: dict.c(336): 134522 words read
INFO: dict.c(358): Reading filler dictionary: /usr/local/share/pocketsphinx/model/en-us/en-us//noisedict
INFO: dict.c(213): Allocated 0 KiB for strings, 0 KiB for phones
INFO: dict.c(361): 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 42672 bytes (41 KiB) for word-final triphones
INFO: dict2pid.c(196): Allocated 42672 bytes (41 KiB) for single-phone word triphones
INFO: kws_search.c(420): KWS(beam: -1080, plp: -23, default threshold -450, delay 10)
Recorder started
Decoder started
我在此处找到了为什么您不需要 -lm 行的参考资料。
Also if you intend to use kws there is no need to use -lm in arguments. You need to remove: "-lm", ".../model/hub4wsj_sc_8k_adapt/etc/hub4.5000.DMP",
这就是答案。
如果我更改上面的代码并删除：
c.setString("-keyphrase", "abomination");
并添加：
c.setString("-kws", "/home/pennyworth/keyphrase.list");
现在输出显示 -kws set
-kws /home/pennyworth/keyphrase.list
我在输出中得到了这个：
INFO: kws_search.c(420): KWS(beam: -1080, plp: -23, default threshold -450, delay 10)
尽管如此，结果为空。
我
NFO: cmn_prior.c(99): cmn_prior_update: from < 73.10 11.10 -10.49 1.23 0.67 -1.37 -5.29 5.17 -0.62 3.91 -0.28 2.56 -2.14 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 73.59 11.36 -9.61 2.41 2.13 0.15 -5.09 4.30 1.03 4.46 -0.13 3.36 -0.62 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 73.59 11.36 -9.61 2.41 2.13 0.15 -5.09 4.30 1.03 4.46 -0.13 3.36 -0.62 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 74.65 10.58 -9.76 4.47 3.63 1.19 -5.20 3.74 2.33 4.75 -0.11 3.06 -0.32 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 74.65 10.58 -9.76 4.47 3.63 1.19 -5.20 3.74 2.33 4.75 -0.11 3.06 -0.32 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 77.49 10.99 -8.80 5.45 4.37 2.83 -4.14 4.06 3.48 5.07 -0.41 2.82 -0.35 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 77.49 10.99 -8.80 5.45 4.37 2.83 -4.14 4.06 3.48 5.07 -0.41 2.82 -0.35 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 73.54 9.62 -11.34 3.19 3.30 2.24 -6.61 4.52 1.31 5.99 -1.28 2.24 -0.96 >
我是否假设 kws 不是 return 炒作？有一个很棒的 python 示例，但是关于 kws 的 java 没有什么用。
https://github.com/cmusphinx/pocketsphinx/blob/master/swig/python/test/kws_test.py
这是 pocketsphinx 的 api 文档，http://cmusphinx.sourceforge.net/doc/pocketsphinx/pocketsphinx_8c_source.html
我不知道如何推进这个。我要么没有正确设置解码器，要么发生了其他事情，这就是我得到 null return 的原因。
我不清楚 -kws 与 -keyphrase vx -kws-threshold 的用法。使用 -kws 是否意味着您不需要其他两个，因为它有效地设置了短语和阈值？
更新代码。添加了字节顺序。 （并确保我的 1 不是 l ）
import javax.sound.sampled.AudioFormat;
import javax.sound.sampled.AudioSystem;
import javax.sound.sampled.DataLine;
import javax.sound.sampled.TargetDataLine;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.nio.ByteBuffer;
import java.nio.ByteOrder;
import java.util.Arrays;
import edu.cmu.pocketsphinx.Decoder;
import edu.cmu.pocketsphinx.Config;
import edu.cmu.pocketsphinx.Hypothesis;
public class Controller {
static {
System.loadLibrary("pocketsphinx_jni");
}
private static ByteArrayOutputStream out;
public static void main(String args[]) {
AudioFormat format = new AudioFormat(44100, 16, 1, true, true);
TargetDataLine targetLine = null;
DataLine.Info targetInfo = new DataLine.Info(TargetDataLine.class, format);
boolean running = true;
try {
targetLine = AudioSystem.getTargetDataLine(format);
targetLine.open();
out = new ByteArrayOutputStream();
int numBytesRead;
byte[] data = new byte[targetLine.getBufferSize() / 5];
Config c = Decoder.defaultConfig();
c.setString("-hmm", "/usr/local/share/pocketsphinx/model/en-us/en-us/");
c.setString("-dict", "/usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict");
c.setString("-keyphrase", "abomination");
c.setFloat("-kws_threshold", 1e-20);
//c.setString("-kws", "/home/bruce/keyphrase.list");
Decoder d = new Decoder(c);
d.setRawdataSize(300000);
targetLine.start();
System.out.println("Recorder started");
byte[] b = new byte[4096];
d.startUtt();
System.out.println("Decoder started");
while ((running)) {
int nbytes;
short[] s = null;
nbytes = targetLine.read(b,0,b.length);
ByteBuffer bb = ByteBuffer.wrap(b, 0, nbytes);
s = new short[nbytes/2];
bb.asShortBuffer().get(s);
bb.order(ByteOrder.LITTLE_ENDIAN);
d.processRaw(s, nbytes/2, false, false);
if (nbytes > 0) {
Hypothesis hypothesis = d.hyp();
if (hypothesis != null) {
System.out.println("------------------------------------------------------");
System.out.println(hypothesis.getHypstr());
System.out.println("------------------------------------------------------");
d.endUtt();
d.startUtt();
}
}
}
}
catch (Exception e) {
System.err.println(e);
}
}
}
这是更新后的输出：
INFO: cmn_prior.c(99): cmn_prior_update: from < 67.54 12.23 -8.09 -0.29 0.56 -0.37 -3.25 7.00 -1.97 3.98 -1.87 3.63 -1.49 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 66.41 12.98 -8.67 -0.63 1.35 -0.13 -3.16 7.97 -3.11 3.57 -0.74 3.91 -1.79 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 66.41 12.98 -8.67 -0.63 1.35 -0.13 -3.16 7.97 -3.11 3.57 -0.74 3.91 -1.79 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 66.66 12.12 -10.32 -1.55 0.57 -0.01 -3.20 8.83 -2.47 4.65 0.07 4.54 -2.35 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 66.66 12.12 -10.32 -1.55 0.57 -0.01 -3.20 8.83 -2.47 4.65 0.07 4.54 -2.35 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 68.35 12.82 -9.91 -1.85 0.77 0.18 -2.25 9.05 -1.75 3.84 0.66 5.82 -2.50 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 68.35 12.82 -9.91 -1.85 0.77 0.18 -2.25 9.05 -1.75 3.84 0.66 5.82 -2.50 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 64.05 14.55 -8.14 -0.36 0.64 0.75 -1.96 9.76 -0.03 5.26 1.16 5.03 -1.68 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 64.05 14.55 -8.14 -0.36 0.64 0.75 -1.96 9.76 -0.03 5.26 1.16 5.03 -1.68 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 63.08 15.30 -8.96 0.76 1.05 0.83 -1.40 10.94 -0.69 4.52 -0.80 3.58 -3.18 >
INFO: cmn_prior.c(99): cmn_prior_update: from < 63.08 15.30 -8.96 0.76 1.05 0.83 -1.40 10.94 -0.69 4.52 -0.80 3.58 -3.18 >
INFO: cmn_prior.c(116): cmn_prior_update: to < 62.15 16.49 -10.32 -0.25 1.14 -0.32 -2.32 10.95 -2.12 2.91 -1.31 2.57 -4.05
更新：
更改了以下顺序以反映 https://github.com/cmusphinx/pocketsphinx/blob/master/swig/java/test/DecoderTest.java
bb.order(ByteOrder.LITTLE_ENDIAN);
bb.asShortBuffer().get(s);
一旦启动，就好像没有输入一样，完全没有输出。
INFO: pocketsphinx.c(152): Parsed model-specific feature parameters from /usr/local/share/pocketsphinx/model/en-us/en-us//feat.params
Current configuration:
[NAME] [DEFLT] [VALUE]
-agc none none
-agcthresh 2.0 2.000000e+00
-allphone
-allphone_ci no no
-alpha 0.97 9.700000e-01
-ascale 20.0 2.000000e+01
-aw 1 1
-backtrace no no
-beam 1e-48 1.000000e-48
-bestpath yes yes
-bestpathlw 9.5 9.500000e+00
-ceplen 13 13
-cmn current current
-cmninit 8.0 40,3,-1
-compallsen no no
-debug 0
-dict /usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict
-dictcase no no
-dither no no
-doublebw no no
-ds 1 1
-fdict
-feat 1s_c_d_dd 1s_c_d_dd
-featparams
-fillprob 1e-8 1.000000e-08
-frate 100 100
-fsg
-fsgusealtpron yes yes
-fsgusefiller yes yes
-fwdflat yes yes
-fwdflatbeam 1e-64 1.000000e-64
-fwdflatefwid 4 4
-fwdflatlw 8.5 8.500000e+00
-fwdflatsfwin 25 25
-fwdflatwbeam 7e-29 7.000000e-29
-fwdtree yes yes
-hmm /usr/local/share/pocketsphinx/model/en-us/en-us/
-input_endian little little
-jsgf
-keyphrase abomination
-kws
-kws_delay 10 10
-kws_plp 1e-1 1.000000e-01
-kws_threshold 1 1.000000e-20
-latsize 5000 5000
-lda
-ldadim 0 0
-lifter 0 22
-lm
-lmctl
-lmname
-logbase 1.0001 1.000100e+00
-logfn
-logspec no no
-lowerf 133.33334 1.300000e+02
-lpbeam 1e-40 1.000000e-40
-lponlybeam 7e-29 7.000000e-29
-lw 6.5 6.500000e+00
-maxhmmpf 30000 30000
-maxwpf -1 -1
-mdef
-mean
-mfclogdir
-min_endfr 0 0
-mixw
-mixwfloor 0.0000001 1.000000e-07
-mllr
-mmap yes yes
-ncep 13 13
-nfft 512 512
-nfilt 40 25
-nwpen 1.0 1.000000e+00
-pbeam 1e-48 1.000000e-48
-pip 1.0 1.000000e+00
-pl_beam 1e-10 1.000000e-10
-pl_pbeam 1e-10 1.000000e-10
-pl_pip 1.0 1.000000e+00
-pl_weight 3.0 3.000000e+00
-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+04
-seed -1 -1
-sendump
-senlogdir
-senmgau
-silprob 0.005 5.000000e-03
-smoothspec no no
-svspec 0-12/13-25/26-38
-tmat
-tmatfloor 0.0001 1.000000e-04
-topn 4 4
-topn_beam 0 0
-toprule
-transform legacy dct
-unit_area yes yes
-upperf 6855.4976 6.800000e+03
-uw 1.0 1.000000e+00
-vad_postspeech 50 50
-vad_prespeech 20 20
-vad_startspeech 10 10
-vad_threshold 2.0 2.000000e+00
-var
-varfloor 0.0001 1.000000e-04
-varnorm no no
-verbose no no
-warp_params
-warp_type inverse_linear inverse_linear
-wbeam 7e-29 7.000000e-29
-wip 0.65 6.500000e-01
-wlen 0.025625 2.562500e-02
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: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/model/en-us/en-us//transition_matrices
INFO: acmod.c(117): Attempting to use PTM computation module
INFO: ms_gauden.c(198): Reading mixture gaussian parameter: /usr/local/share/pocketsphinx/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: /usr/local/share/pocketsphinx/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 /usr/local/share/pocketsphinx/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(114): State beam -225 Phone exit beam -225 Insertion penalty 0
INFO: dict.c(320): Allocating 138623 * 32 bytes (4331 KiB) for word entries
INFO: dict.c(333): Reading main dictionary: /usr/local/share/pocketsphinx/model/en-us/cmudict-en-us.dict
INFO: dict.c(213): Allocated 1014 KiB for strings, 1677 KiB for phones
INFO: dict.c(336): 134522 words read
INFO: dict.c(358): Reading filler dictionary: /usr/local/share/pocketsphinx/model/en-us/en-us//noisedict
INFO: dict.c(213): Allocated 0 KiB for strings, 0 KiB for phones
INFO: dict.c(361): 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
Recorder started
Decoder started
INFO: dict2pid.c(132): Allocated 42672 bytes (41 KiB) for word-final triphones
INFO: dict2pid.c(196): Allocated 42672 bytes (41 KiB) for single-phone word triphones
INFO: kws_search.c(420): KWS(beam: -1080, plp: -23, default threshold -450, delay 10)
abomination /le-20/
每个开发人员都应该知道浮点格式，它以数字开头 - 1e-1 与 0.1 相同，5e-3 与 0.005 相同。 le-20 和 l 不是浮点数。
d.processRaw(s, nbytes/2, false, false);
d.setKws("abomination", );
此代码不应编译。此外，您不应在处理过程中设置搜索。
I'm not clear on the use of -kws vs -keyphrase vx -kws-threshold. Does using -kws mean you don't need the other two since it effectively sets both the phrase and threshold?
是，关键字列表替换命令行选项
Am I assuming wrong that kws doesn't return a hyp? T
这不是 return 假设，因为您没有将正确的音频传递到解码器。你需要
bb.order(ByteOrder.LITTLE_ENDIAN);`
没有它，它会传递大端数据（java 默认值）。来源中的 Swig 示例有。
INFO: cmn_prior.c(99): cmn_prior_update: from < 73.10 11.10 -10.49 1.23 0.67 -1.37 -5.29 5.17 -0.62 3.91 -0.28 2.56 -2.14 >
CMN 范围为 70 表示端序问题。使用正确的字节序，第一个 CMN 值应该在 30 到 60 之间。