HTK Speech Recognition ASK

A phonetic modeling model for continuous speech recognition with large vocabulary

When adding a new pronunciation, you can only modify dict and gram.

Recognition results to time conversion in HTK

13600000 16320000 hao-1452.207031

Directly divided by 10 of the 7-time Square

Hao pronunciation from 1.36 seconds to 1.632 seconds that means HTK is output in 100 microseconds as the base unit.

HTK bad data or over pruning alarms

Find an e-mail

Ask:

Hi

Omer Moav wrote:

Processing Data:cmu_us_arctic_awb_a0015.cmp; Label Cmu_us_arctic_awb_a0015.lab

Unable to traverse states in 1 frames

WARNING [ -7324] Stepback:file

/home/omergil/downloads/hts-demo_cmu-arctic-awb/cmp/cmu_us_arctic_awb_a0015.cmp

-Bad data or over pruning

In/home/omergil/downloads/htk/bin.linux/herest

For:

The above warning means that cmu_us_arctic_awb_a0015.cmp is corrupted.

I recommend to re-run it.

I find the corresponding alarm file according to the alarm information by looking at my error message because the cut voice is too small, basically only half of the pronunciation, so the training when the recognition will not come out a warning

Three-factor model State binding

... (reason not to join SP Mute model)

WB SP

WB sil

Tc

SP and SIL are responsible for segmentation

SIL th ih s sp m ae n sp ...

Hled.exe-n./lists/triphones1-l ' * '-I./LABELS/WINTRI.MLF mktri.led./LABELS/ALIGNED.MLF

Then becomes

Sil th+ih th-ih+s ih-s sp m+ae m-ae+n ae-n sp ...

Mktri.hed explanations generated by Perl/scripts/maketrihed/lists/monophones1/lists/triphones1

CL./lists/triphones1

TI t_b {(*-b+*,b+*,*-b). TRANSP}

TI t_p {(*-p+*,p+*,*-p). TRANSP}

...

CL indicates the meaning of cloning

TI represents the connection

Herest-b-C/config/config2-i./labels/wintri.mlf-t 250.0 150.0 1000.0-s stats-s train.scp-h./hmms/hmm11/macros-h ./hmms/hmm11/hmmsdef-m./hmms/hmm12./lists/triphones1

-B Store file in binary mode-s stats generate stats file

Constructing Decision Trees

Recognition results Correct accurate

1), the correct recognition rate of words

correct = (n-d-S)%N * 100%

2), Recognition accuracy

Accurate = (n-d-s-i)%N * 100%

N: Number of original script file morphemes

D: The recognition result corresponds to the number of words deleted in the reference sentence script

S: Recognition results correspond to the number of words replaced in the reference sentence script

I: The recognition result corresponds to the number of words inserted in the reference sentence script

In the Gram.txt file, change (Sent-start (< $word >|<SENT-START>) sent-end) to

< $word > Add Sent-start effect in Word pronunciation failed

Successful results $word =a|ai|an|ang|ao|ba|bai|ban|bang|bao|bei|ben|beng|bi|bian|biao|bie|bin|bing|

bo|bu|ca|cai|can|cang|cao|ce|cen|ceng|cha|chai|chan|chang|chao|che|chen|cheng|chi|

chong|chou|chu|chua|chuai|chuan|chuang|chui|chun|chuo|ci|cong|cou|cu|cuan|cui|cun|

cuo|da|dai|dan|dang|dao|de|den|deng|di|dia|dian|diao|die|ding|diu|dong|dou|du|duan|

dui|dun|duo|e|en|eng|er|fa|fan|fang|fei|fen|feng|fo|fu|ga|gai|gan|gang|gao|ge|gei|

gen|geng|gong|gou|gu|gua|guai|guan|guang|gui|gun|guo|ha|hai|han|hang|hao|he|hei|hen|

Heng|hong|hou|hu|hua|huai|huan|huang|hui|hun|huo|ji|jia|jian|jiang|jiao|jie|jin|jing|jiong|jiu|jv|jvan|jve|jvn |ka|kai|kan|kang|kao|ke|ken|keng|kong|kou|ku|kua|kuai|kuan|

kuang|kui|kun|kuo|la|lai|lan|lang|lao|le|lei|leng|li|lia|lian|liang|liao|lie|lin|

ling|liu|lo|long|lou|lu|luan|lve|lun|luo|lv|ma|mai|man|mang|mao|me|mei|men|meng|mi|

mian|miao|mie|min|ming|miu|mo|mou|mu|na|nai|nan|nang|nao|ne|nei|nen|neng|ni|nian|

niang|niao|nie|nin|ning|niu|nong|nu|nve|nv|nuan|nuo|nun|o|ou|pa|pai|pan|pang|pao|

pei|pen|peng|pi|pian|piao|pie|pin|ping|po|pou|pu|qi|qia|qian|qiang|qiao|qie|qin|

qing|qiong|qiu|qv|qvan|qve|qvn|ran|ang|rao|re|ren|reng|ri|rong|rou|ru|rua|ruan|rui|

run|ruo|sa|sai|san|sang|sao|se|sen|seng|sha|shai|shan|shang|shao|she|shen|sheng|shi|

shou|shu|shua|shuai|shuan|shui|shun|shuo|si|song|sou|su|suan|sui|sun|suo|ta|tai|tan|

tang|tao|te|tei|teng|ti|tian|tiao|tie|ting|tong|tou|tu|tuan|tui|tun|tuo|wa|wai|wan|

wang|wei|wen|weng|wo|wu|xi|xia|xian|xiang|xiao|xie|xin|xing|xiong|xiu|xv|xvan|xve|

xvn|ya|yan|yang|yao|ye|yv|yi|yin|ying|yo|yong|you|yvan|yve|yvn|za|zai|zan|zang|zao|

ze|zei|zen|zeng|zha|zhai|zhan|zhang|zhao|zhe|zhen|zheng|zhi|zhong|zhou|zhu|zhua|

Zhuai|zhuan|zhuang|zhui|zhun|zhuo|zi|zong|zou|zu|zuan|zui|zun|zuo|fou|shuang|silence;

(Sent-start < $word > Sent-end)

Yes

L Context-sensitive modeling can be a good solution to

L Refinement Modeling

L Cooperative Pronunciation

L Pronunciation Variation

L Accent

L parameter sharing level

L Model level (Model-level)

L Status level (State-level)

L Mixed Grade (Mixture-level)

L sharing of various other parameters (e.g. transfer matrix, center, variance, mixed weights , etc.)

The production and function of vfloors

HCOMPV has a number of options specified fori t.the-f option causes Avariancefloor macro

(called vfloors) to be generated which are equal to 0.01 times the global variance. This is a vector of

Values which is used to set Afloor on the variances estimated in the subsequent steps.

Generates initialization values for the next phase of variances

D:/tryputong>hvite-h./hmms/hmm12/macros-h./hmms/hmm12/hmmsdef-s test.scp-l

*-I./RESULTS/RECOUT_STEP1.MLF-W wdnet-p 0.0-s 5.0/dict/dict2./lists/tri

Phones1

ERROR [+8231] gethcimodel:cannot find hmm [t-]ei[+???]

FATAL error-terminating Program Hvite

1 Regular Expressions <> one or more [] 0 or one

The beauty of mathematics Series three -- application of Hidden Markov model in language processing

Http://www.google.cn/ggblog/googlechinablog/2006/04/blog-post_1583.html