Enhanced Tree Clustering with Single Pronunciation Dictionary for Conversational Speech Recognition
Hua Yu and Tanja Schultz
Interactive Systems Lab,Carnegie Mellon University,Pittsburgh,PA15213
hyu@https://www.doczj.com/doc/b316471755.html,
Abstract
Modeling pronunciation variation is key for recogniz-ing conversational speech.Rather than being limited to dictionary modeling,we argue that triphone clustering is an integral part of pronunciation modeling.We propose a new approach called enhanced tree clustering.This approach,in contrast to traditional decision tree based state tying,allows parameter sharing across phonemes. We show that accurate pronunciation modeling can be achieved through ef?cient parameter sharing in the acous-tic https://www.doczj.com/doc/b316471755.html,bined with a single pronunciation dic-tionary,a1.8%absolute word error rate improvement is achieved on Switchboard,a large vocabulary conversa-tional speech recognition task.
1.Introduction
Modeling conversational speech is a major challenge for current speech recognition research.Conversational speech is characterized by rampant pronunciation vari-ations,where accurate pronunciation modeling can lead to high recognition performance.Traditionally,people have tried to add alternative pronunciations to the recog-nition dictionary.Despite extensive investigation,this has yielded only marginal improvement.
What is pronunciation modeling?Is pronunciation modeling a synonym for dictionary modeling?As shown in Figure1,the boundary between pronunciation mod-eling and acoustic modeling is not clear.For a given word,the lexicon is?rst looked up to convert the word into a phoneme sequence,which is subsequently trans-lated into a state sequence using a phonetic decision tree. The state sequence is ultimately used to align with the acoustic observation.Hence the phonetic decision tree,a traditionally acoustic modeling concept,also plays a key role in the mapping from symbolic(phoneme)level to model(state)level.
Subtle pronunciation variations may actually be bet-ter modeled implicitly at the acoustic model level,rather than modeled explicitly at the lexical level.For example, phoneme AX can be alternately realized as phoneme IX in certain words:
AFFECTIONATE AX F EH K SH AX N AX T AFFECTIONATE(2)AX F EH K SH AX N IX T
Figure1:Pronunciation model as a mapping from sym-bolic level to model level
Instead of adding a variant in the dictionary,we can keep the dictionary unchanged,and either augment the mixture model of AX with Gaussians from the mixture model of IX(as in[1]),or simply tie these two models together.
This paper focuses on implicit pronunciation mod-eling using decision tree based tying.In Section2,we will?rst examine the relative merits of different pronun-ciation modeling methods,namely,explicitly adding dic-tionary variants versus implicit modeling using decision tree based state tying.Section3introduces enhanced tree clustering that allows parameter sharing/tying across different phonemes.The new approach is evaluated on Switchboard,a large vocabulary speech recognition (LVCSR)task.Section4presents experiments and dis-cussions.Related research work is reviewed in Section5.
2.Explicit vs.Implicit Pronunciation
Modeling
Early attempts at pronunciation modeling take the form of manually editing a lexicon.For example,to model the?apping of T in BETTER,an alternative lexical entry BETTER(2)is introduced:
BETTER B EH T AXR
BETTER(2)B EH DX AXR
Interestingly,for this kind of pronunciation variation, triphone modeling turns out to be better than dictionary editing.Triphone was originally introduced to model context dependency.As there is a large number of tri-phones in an
LVCSR system,
decision tree based state
tying is widely used to cluster triphones[2].This ensures suf?cient training data for each model as well as better
generalization to unseen contexts.It turns out that many coarticulation rules,such as the?apping of T,can be well captured by triphone models in a purely data-driven fash-ion.This gives us another solution:leave the dictionary under-speci?ed,and use automatic triphone clustering for pronunciation modeling1.The use of mixture model as the underlying distribution is also important,since we are using the same triphone model for both the?apped and un?apped version of T.
Compared to triphone clustering,manually editing a lexicon is both labor intensive and error prone.In the example of the?apping of T,one needs to be extremely careful to make sure that all relevant dictionary entries are modi?ed,while nothing else is erroneously changed. This turns out to be quite dif?cult in reality.Whereas in the automatic solution,it is a lot easier to keep the dictionary simple and consistent.
One could also use automatic procedures to gener-ate pronunciation variants,in order to avoid the pitfall of manual editing.There has been a lot of research in this area.So far,the improvement has been marginal.This could be attributed to several undesirable side effects of pronunciation variants:
First,adding variants increases lexical confusabil-
ity in a recognition dictionary;
Second,if not done properly,adding variants in-
creases model confusability during training.As il-
lustrated in Figure2,when a variant replaces phone
A by phone B,we are distributing to model
B the
data that was originally used to train model A.In
cases where the variant is spurious,model B will
be contaminated with data belonging to A,making
A and
B more confusable.
Figure2:Model Contamination
In summary,due to the complex interaction between lexicon and acoustic modeling,adding pronunciation variants should be exercised with great care.In the next section,we introduce enhanced tree clustering for im-plicit pronunciation modeling.
3.Enhanced Tree Clustering Decision tree based state tying allows parameter shar-ing at leaf nodes of a tree.Typically,one decision tree 1This actually resonates with single pronunciation dictionary[3]is grown for each sub-state(begin/middle/end)of each phone.With50phonemes in the phone set,150sepa-rate trees are built(Figure3(a)).Parameter sharing is not allowed across different phones or sub-states.With en-hanced tree clustering,a single decision tree is grown for all sub-states of all the phones(Figure3(b)).The clus-tering procedure starts with all polyphones at the root. Questions are asked regarding the identity of the cen-ter phone and its neighboring phones,plus the sub-state identity(begin/middle/end).At each node,the question that yields the highest information gain is chosen and the tree is split.This process is repeated until either the tree reaches a certain size or a minimum count threshold is https://www.doczj.com/doc/b316471755.html,pared to the traditional multiple-tree ap-proach,a single tree allows more?exible sharing of pa-rameters.Any nodes can potentially be shared by multi-ple phones,as well as
sub-states.
CH?e
AA?m
AA?b
AA?e
0=obstruent?0=begin?state?
n y n y
?1=syllabic?0=mid?state??1=obstruent?0=end?state? CH?m
CH?b
(b)
(a)
n y
0=vowel?
Figure3:(a)shows the traditional clustering approach: one tree per phone and sub-state.(b)shows the concept of enhanced clustering using a single tree.
In sloppy speech,people don’t differentiate phonemes as much as they do in read speech.Dif-ferent phonemes tend to exhibit more similarity.Single tree clustering is well suited to capture these cross-phone parameter sharing,whereas the traditional approach does not allow such sharing.
Furthermore,sharing parameters across phones al-leviates certain problems in a dictionary,namely,over-speci?cation and inconsistencies.
Examples of these include the handling of T and DX, AX and IX as mentioned before.Some lexicons choose to differentiate them,while others do not.In lexicons that do,they are most often not marked consistently through-out.This is also referred to as the problem of choosing an optimal phoneme set.Unfortunately,such an optimal phoneme set does not exist.By allowing parameter shar-ing across phonemes,we no longer face this tough de-cision:
if phonemes
are indistinguishable under certain
context,they will be allowed to share the same model;if they show suf?cient differences under certain other con-text,they will be allowed to use different models.
Under the same argument,enhanced tree clustering is also a preferable choice for multilingual speech recogni-
tion or non-native speech recognition,where the phone set is not well de?ned and different phones show in-creased similarity.
4.Experiments and Discussions
4.1.Setup
Experiments are performed on the Switchboard(SWB) task.The test set is a1hour subset of the2001Hub5e evaluation set.The full training set includes160hours of SWB data and17hours of CallHome data.We typi-cally use a66hour subset of the160hours of SWB data for fast experimentation.The baseline system is devel-oped using the Janus speech recognition toolkit[4].The front-end uses vocal tract length normalization,cluster-based cepstral mean normalization,and an11-frame con-text window for delta and double-delta.Linear discrimi-nant analysis is applied to reduce feature dimensionality to42,followed by maximum likelihood linear transform. We use a15k vocabulary and a trigram language model trained on SWB and CallHome.
The baseline acoustic model uses a quinphone tree based,two level state tying scheme(described in[5],sim-ilar to soft-tying[6]):24k distributions sharing6k code-books,with a total of74k Gaussians.It has a word error rate(WER)of34.4%[7].Unless otherwise stated,all re-sults reported here are based on?rst-pass decoding,i.e. no adaptation or multi-stage processing.
Computational cost is the main dif?culty for growing a single big tree.As the number of unique quinphones on the Switchboard task is around600k,direct cluster-ing on all of them is quite daunting.The traditional ap-proach doesn’t have this problem,since polyphones are divided naturally according to center phone and sub-state identities.For this reason,we conducted two experiments to investigate the effects of cross-phone tying and cross-substate tying separately.
4.2.Cross-Phone Clustering
We grow six triphone trees for cross-phone clustering: one for each of the begin/middle/end state of vowels and consonants.We could have built three big trees,without differentiating between vowels and consonants.The rea-son is:?rst,we expect little parameter sharing between vowels and consonants;furthermore,separating them re-duces computation.
Initial experimentation gives a small,albeit signi?-cant,improvement(from34.4%to33.9%).As the tree is grown in a purely data-driven fashion,one may won-der how much cross-phone sharing there actually is.It is possible that questions regarding center phones are highly important,therefore they are asked earlier in the tree,re-sulting in a system which is no different from a phonet-ically tied system.We examined the six triphone trees, and found that20%to38%of the leaf nodes(out of a total of24k)are indeed shared by multiple phones.
4.3.Single Pronunciation Dictionary
Motivated by Hain’s work on single pronunciation dictio-naries(SPD)[3],we tried to reduce the number of pro-nunciation variants in the dictionary.The procedure to derive a new lexicon is even simpler than Hain’s.First, we count the frequency of pronunciation variants in the training data.Variants with a relative frequency of less than20%are removed.For unobserved words,we keep only the baseform(which is more or less a random de-cision).Using this procedure,we reduced the dictionary from an average2.2variants per word to1.1variants per word.We are not using strictly single pronunciations,so that we can keep the most popular variants,while prun-ing away spurious ones.For example,the word A has two variants in the resulted dictionary:
A AX
A(2)EY
Simply using SPD with traditional clustering gives a0.3%improvement,which is comparable to Hain’s results.More interestingly,cross-phone clustering re-sponds quite well with SPD.Overall,we achieve a1.3% gain by cross-phone clustering on a single pronunciation dictionary(Table1).
Dictionary Clustering WER(%)
multi-regular34.4
pronunciation cross-phone33.9
single regular34.1
pronunciation cross-phone33.1
Table1:Cross-Phone Clustering Experiment
Note that experiments in Table1are based on the66 hour training set and triphone clustering.The gain holds when we switch to the full180hour training data and quinphone clustering.Due to high computation,we only compared two systems:one with multi-pronunciation lexicon and no cross-phone clustering,and the other with single-pronunciation lexicon and cross-phone clustering. WER improves from33.4%to31.6%,a1.8%absolute gain.
4.4.Discussion
To explain why cross-phone clustering helps more with SPD(from34.1%to33.1%),than with a regular dictio-nary(from34.4%to33.9%),let us consider the(unin-tended)side effects of pronunciation variants.When a variant replaces phone A by phone B,we are distribut-ing to model B the data that was originally used to train model A(Figure2),effectively allowing parameter shar-ing between phones.Hence,cross-phone parameter shar-ing exists even without explicit cross-phone clustering. But this sharing will only contaminate models and
hurt
performance,if those variants are not carefully scruti-nized.As discussed in Section 2,adding pronunciation variants is not as straightforward as it may seem.Changes to a dictionary should be coordinated closely with acous-tic modeling.
?1=consonant?0=high?vowel??1=obstruent??1=L | R | W?
0=high?vowel?
0=front?vowel?n
y
n y
n y
?1=voiced?VOWEL?b Figure 4:Top part of V owel-b tree (beginning state of vowels).“=”questions ask about the immediately left phone,“0=”questions ask about the center phone.
The top portion of the tree for the beginning state of vowels is shown in Figure 4.It is clear that questions about center phone identities are not necessarily preferred over contextual questions.Again,20%to 40%of the leaf nodes are found to be shared by multiple phones.Conso-nants that are most frequently tied together are:DX and HH ,L and W ,N and NG .V owels that are most frequently tied together are:AXR and ER ,AE and EH ,AH and AX .4.5.Cross-Substate Clustering
In this experiment,we build one tree for each phone,which covers all three sub-states.Three new questions are added regarding sub-state identities.Contrary to our experience with cross-phone clustering,we ?nd those three questions to be highly important.They are chosen in most cases as the top two questions.Hence,the re-sulted tree is not any different from three separate trees,as in traditional clustering.
5.Related Work
Pronunciation modeling has received a lot of attention recently.To explain why simply modifying the lexicon does not work as expected,Jurafsky et al.argued that tri-phones can already capture many kinds of pronunciation variations [8],including phone substitution and reduc-tion.Hain questioned the use of pronunciation variants in a recent work called “single pronunciation dictionary”[3].By systematically removing variants,he showed a slight gain over a state-of-the-art Switchboard system.
Recent studies focus on implicit pronunciation mod-eling by leveraging various acoustic modeling mecha-nisms.Saraclar et al.proposed a state level pronuncia-tion model,which tries to add Gaussians from the surface form model to the baseform model.Pronunciation mod-eling at a deeper level allows greater expressive power and modeling resolution.For example,a pronunciation
variant,as a phone sequence,can always be translated
into a state/model sequence,but not vice versa.A ma-jority of state/model sequences cannot be represented as valid phone sequences.To address this,Hain proposed the Hidden Model Sequence Model [9].
6.Conclusion
This paper shows that polyphone clustering is an inte-gral part in pronunciation modeling.Enhanced tree clus-tering is proposed to allow ef?cient parameter sharing across phonemes.This effectively handles blurred phone identities commonly found in conversational speech,in-stead of making a hard decision in the https://www.doczj.com/doc/b316471755.html,bined with a single pronunciation dictionary,the new approach achieves a 1.8%WER reduction on the Switchboard task,over state of the art decision tree based state tying.We be-lieve this approach also holds promise in other tasks,such as multilingual speech recognition and non-native speech recognition.
7.References
[1]M.Saraclar,H.Nock,and S.Khudanpur,“Pronunci-ation modeling by sharing gaussian densities across phonetic models,”Computer Speech and Language ,vol.14,no.2,pp.137–160,April 2000.[2]S.Young,J.Odell,and P.Woodland,“Tree-based
state tying for high accuracy acoustic modelling,”in Proc.ARPA HLT Workshop ,1994.[3]T.Hain,“Implicit pronunciation modelling in ASR,”
in ISCA Pronunciation Modeling Workshop ,2002.[4]Michael Finke,J¨u rgen Fritsch,Petra Geutner,Klaus
Ries,and Torsten Zeppenfeld,“The JanusRTk Switchboard/Callhome 1997evaluation system,”in Proceedings of LVCSR Hub5-e Workshop ,1997.[5]Michael Finke and Ivica Rogina,“Wide context
acoustic modeling in read vs.spontaneous speech,”in Proc.ICASSP ,1997,pp.1743–1746.[6]X.Luo and F.Jelinek,“Probablistic classi?cation of
hmm states for large vocabulary continuous speech recognition,”in Proc.ICASSP ,1999.[7]H.Soltau,H.Yu,F.Metze,C.F¨u gen,Y .Pan,and
S.Jou,“ISL meeting recognition,”in Rich Tran-scription Workshop ,Vienna,V A,2002.[8]D.Jurafsky,W.Ward,J.Zhang,K.Herold,X.Yu,
and S.Zhang,“What kind of pronunciation variation is hard for triphones to model?,”in Proc.ICASSP ,2001.[9]Thomas Hain,Hidden Model Sequence Models for
Automatic Speech Recognition ,Ph.D.thesis,Cam-bridge University,
2001.
1、人称代词顺口溜：人称代词有两类，一类主格一类宾；主格代词本领大，一切动作由它发；宾格代词不动脑，介动之后跟着跑。 2、物主代词顺口溜：物主代词不示弱，带着‘白勺’来捣乱；形容词性物主代，抓住名词不放松； 1、人称代词的主格在句子中作主语或主语补语。一般在句首，动词前。 例如：John waited a while but eventually he went home. 约翰等了一会儿，最后他回家了。 John hoped the passenger would be Mary and indeed it was she. 约翰希望那位乘客是玛丽，还真是她。 说明：在复合句中，如果主句和从句主语相同，代词主语要用在从句中，名词主语用在主句中。在电话用语中常用主格。 例如：When he arrived, John went straight to the bank. 约翰一到就直接去银行了。 I wish to speak to Mary. This is she. 我想和玛丽通话，我就是玛丽。 2、人称代词的宾格在句中作宾语或表语，在动词或介词后。 例如：Do you know him？(作宾语) 你认识他吗？ Who is knocking at the door？It’s me. (作表语) 是谁在敲门？是我。 说明：单独使用的人称代词通常用宾格，即使它代表主语时也是如此。 例如：I like English. Me too. 我喜欢英语。我也喜欢。 3、注意：在动词be 或to be 后的人称代词视其前面的名词或代词而定。 例如：I thought it was she.我以为是她。(主格----主格) I thought it to be her.(宾格----宾格) I was taken to be she.我被当成了她。(主格----主格) They took me to be her.他们把我当成了她。(宾格----宾格) 4、人称代词并列时的排列顺序 1）单数人称代词并列作主语时，其顺序为： 第二人称→第三人称→第一人称 即you and I he/she/it and I you, he/she/it and I 顺口溜：第一人称最谦虚，但若错误责任担，第一人称学当先。 例如：It was I and John that made her angry. 2）复数人称代词作主语时，其顺序为： 第一人称→第二人称→第三人称 即we and you you and they we, you and they
With的用法全解 with结构是许多英语复合结构中最常用的一种。学好它对学好复合宾语结构、不定式复合结构、动名词复合结构和独立主格结构均能起很重要的作用。本文就此的构成、特点及用法等作一较全面阐述，以帮助同学们掌握这一重要的语法知识。 一、 with结构的构成 它是由介词with或without+复合结构构成，复合结构作介词with或without的复合宾语，复合宾语中第一部分宾语由名词或代词充当，第二部分补足语由形容词、副词、介词短语、动词不定式或分词充当，分词可以是现在分词，也可以是过去分词。With结构构成方式如下： 1. with或without-名词/代词+形容词； 2. with或without-名词/代词+副词； 3. with或without-名词/代词+介词短语； 4. with或without-名词/代词 +动词不定式； 5. with或without-名词/代词 +分词。 下面分别举例： 1、 She came into the room，with her nose red because of cold.（with+名词+形容词，作伴随状语）
2、 With the meal over ， we all went home.（with+名词+副词，作时间状语） 3、The master was walking up and down with the ruler under his arm。（with+名词+介词短语，作伴随状语。） The teacher entered the classroom with a book in his hand. 4、He lay in the dark empty house，with not a man ，woman or child to say he was kind to me.（with+名词+不定式，作伴随状语）He could not finish it without me to help him.（without+代词 +不定式，作条件状语） 5、She fell asleep with the light burning.（with+名词+现在分词，作伴随状语） Without anything left in the with结构是许多英 语复合结构中最常用的一种。学好它对学好复合宾语结构、不定式复合结构、动名词复合结构和独立主格结构均能起很重要的作用。本文就此的构成、特点及用法等作一较全面阐述，以帮助同学们掌握这一重要的语法知识。 二、with结构的用法 with是介词，其意义颇多，一时难掌握。为帮助大家理清头绪，以教材中的句子为例，进行分类，并配以简单的解释。在句子中with结构多数充当状语，表示行为方式，伴随情况、时间、原因或条件（详见上述例句）。 1.带着，牵着…… (表动作特征)。如： Run with the kite like this.
with的用法大全----四级专项训练with结构是许多英语复合结构中最常用的一种。学好它对学好复合宾语结构、不定式复合结构、动名词复合结构和独立主格结构均能起很重要的作用。本文就此的构成、特点及用法等作一较全面阐述，以帮助同学们掌握这一重要的语法知识。 一、 with结构的构成 它是由介词with或without+复合结构构成，复合结构作介词with或without的复合宾语，复合宾语中第一部分宾语由名词或代词充当，第二部分补足语由形容词、副词、介词短语、动词不定式或分词充当，分词可以是现在分词，也可以是过去分词。With结构构成方式如下： 1. with或without-名词/代词+形容词; 2. with或without-名词/代词+副词; 3. with或without-名词/代词+介词短语; 4. with或without-名词/代词+动词不定式; 5. with或without-名词/代词+分词。 下面分别举例：
1、 She came into the room，with her nose red because of cold.(with+名词+形容词，作伴随状语) 2、 With the meal over ， we all went home.(with+名词+副词，作时间状语) 3、The master was walking up and down with the ruler under his arm。(with+名词+介词短语，作伴随状语。) The teacher entered the classroom with a book in his hand. 4、He lay in the dark empty house，with not a man ，woman or child to say he was kind to me.(with+名词+不定式，作伴随状语) He could not finish it without me to help him.(without+代词 +不定式，作条件状语) 5、She fell asleep with the light burning.(with+名词+现在分词，作伴随状语) 6、Without anything left in the cupboard， she went out to get something to eat.(without+代词+过去分词，作为原因状语) 二、with结构的用法 在句子中with结构多数充当状语，表示行为方式，伴随情况、时间、原因或条件(详见上述例句)。
Spirited Away C: I?ll miss you, Chihiro. Your best friends, Rile F: Chihiro, Chihiro, we?re almost there. M: This really is the middle of nowhere. I?ve got to go to the next town to shop. F: We?ll just have to learn to like it. Look, Chihiro, there?s your new school. Looks really good, doesn?t it? M: It doesn?t look so bad. C: It?s gonna stink. I liked my old school. Oh````Mom, my flowers are dying. M: I told you not to smother them like that. We?ll put them in water when we get to our new home. C: I finally get a bouquet, and it?s a good-bye present. That?s depressing. M: Daddy bought you a rose for your birthday. Don?t you remember? C: Yes, one. Just one rose isn?t a bouquet. M: Hold on to your card. I?m opening the window and quit whining. It?s fun to move to a new place. It?s an adventure. F: Wait, did I take the wrong turn? That can?t be right. M: Look, there?s our house. It?s that blue one on the end. F: Oh, you?re right. I must have missed the turn-off. This road should get us there. M: Honey, don?t take a short cut. You always get us lost. F: Trust me, it?s gonna work. C: What are these stones? They look like little houses. M: They?re shines. Some people think little spirits live there. C: Dad, I think we?re lost. F: We?re fine. I?ll get a full wheel drive. M: Sit down please, sweetie. Slow down. You?re gonna kill us. F: What?s that? M: What?s this old building? F: It looks like an entrance. M: Honey, get back in the car. We are going to be late. Oh for heaven?s sake. F: This building is not old. It?s fake. These stones are just made of plaster. C: The wind?s pulling us in. M: What is it? F: Come one. Let?s go in. I want to see what?s on the other side. C: I?m not going. It gives me the creeps. F: Don?t be such a scaredy cat, Hiro. Let?s just take a look. M: The movers will get to our house before we do. F: It?s alright. They?ve got the keys. They can start without us. M: Alright. Just a quick look. C: Forget it. I?m not going. Come on you guys, le t?s get out of here. F: Come on honey. It will be fun. C: I?m not going. M: Chihiro, just wait in the car then. C: But mom….Wait for me. F: Everybody watch your step. M: Chihiro, don?t cling like that. You make me trip. -What is this place? -Oh, do you hear that? - It sounds like a train. - We must be near a train station. - Come on. Let?s go and check it out. - What are those weird buildings? - I knew it. It?s an abandoned theme park. See? They built the everywhere in early 90s, then the economy went bad and they all went bankrupt. This must be one of them. - Where are you going? You said just a quick look. Now let?s go back. Did you hear that building? It was moaning. - It?s just the wind. Oh, what a beautiful place. We should have brought our lunch. Then we could have a picnic. - Look, they were planning to put a river here. En~~~ you smell that? Something smells delicious. - Yeah, I?m starving. - Maybe this theme park is still in business. Let?s go. - Chihiro, hurry it up. - Wait a minute! - Over there….This way. - How strange! They?re all restaurants. - Where?s everybody? - Aaa… There it is! Hey, I found it. Hi, you got to see this. In here. - Aaa… Look at this! - Hello in there. Does anybody work here? - Come in Hiro. It looks delicious. - Anybody? - Don?t worry honey. We can pay the bill when they get back. - Good plan. Hey that looks great. - I wonder what this is called. Oh… It?s delicious. Chihiro, you have to taste this. - I don?t want any. We?re gonna get in trouble. Let?s just get out of here. - Don?t worry. You?ve got daddy here. He?s got credit cards and cash. - Chihiro, you have to try this. It?s so tender. - Mustard? - Thank you. - Come on, you guys. You can?t - That?s weird. It?s a bath house. There?s the train. - You shouldn?t be here. Get out of here, now! - What? - It?s almost night. Leave before it gets dark. They are lighting the lamps. Get out of here. You?ve gotta get across the river. Go! I?ll distract them. - What?s up with him? - Mom, dad, come on, quit eating, let?s get out of here. Ahh~~~~Mom, dad, where are you?.....Water?!! What?... I?m dreaming! I?m dreaming! Come on, wake up! Wake up! This is a dream. Just a dream. Away, away. Disappear! Aaa~~~~~? I?m see-through! It?s just a bad dream. - Don?t be afraid! I just wantta help you. - No…No! - Open your mouth and eat this. You must eat some food from this world or else you?ll disappear. - No. - Don?t worry. It won?t turn you into a pig. Chew it and swallow. There you go. You?re all better. See for yourself. - I?m OK. - You see. Now come with me. - Where are my mom and dad? They didn?t really turn into pigs, did they? - You can?t see them now, but you will. Don?t move. That bird?s looking for you. You?ve gotta get out of here. - My legs, I can?t stand up. Help! What do I do? - Calm down. Take a deep breath. In the name of the wind and water within thee, unbind her. Get up. - You have to hold your breath when we across the bridge. Even the tiniest breath will break the spell and then everyone will see you. - I?m scared. - Just stay calm. I?m back from my mission. - Welcome back Master Haku. - Take a deep breath. Hold it. Hang on, almost there. - Master Haku, where?ve you been? What? A human? - Let?s go. - They know you?re here. - I?m sorry. I took a breath.
with用法归纳 （1）“用……”表示使用工具，手段等。例如： ①We can walk with our legs and feet. 我们用腿脚行走。 ②He writes with a pencil. 他用铅笔写。 （2）“和……在一起”，表示伴随。例如： ①Can you go to a movie with me? 你能和我一起去看电影'>电影吗？ ②He often goes to the library with Jenny. 他常和詹妮一起去图书馆。 （3）“与……”。例如： I’d like to have a talk with you. 我很想和你说句话。 （4）“关于，对于”，表示一种关系或适应范围。例如： What’s wrong with your watch? 你的手表怎么了？ （5）“带有，具有”。例如： ①He’s a tall kid with short hair. 他是个长着一头短发的高个子小孩。 ②They have no money with them. 他们没带钱。 （6）“在……方面”。例如： Kate helps me with my English. 凯特帮我学英语。 （7）“随着，与……同时”。例如： With these words, he left the room. 说完这些话，他离开了房间。 [解题过程] with结构也称为with复合结构。是由with+复合宾语组成。常在句中做状语，表示谓语动作发生的伴随情况、时间、原因、方式等。其构成有下列几种情形： 1.with+名词(或代词)+现在分词 此时，现在分词和前面的名词或代词是逻辑上的主谓关系。 例如:1)With prices going up so fast, we can't afford luxuries. 由于物价上涨很快，我们买不起高档商品。（原因状语） 2)With the crowds cheering, they drove to the palace. 在人群的欢呼声中，他们驱车来到皇宫。（伴随情况） 2.with+名词(或代词)+过去分词 此时，过去分词和前面的名词或代词是逻辑上的动宾关系。
独立主格篇 独立主格，首先它是一个“格”，而不是一个“句子”。在英语中任何一个句子都要有主谓结构，而在这个结构中，没有真正的主语和谓语动词，但又在逻辑上构成主谓或主表关系。独立主格结构主要用于描绘性文字中，其作用相当于一个状语从句，常用来表示时间、原因、条件、行为方式或伴随情况等。除名词/代词+名词、形容词、副词、非谓语动词及介词短语外，另有with或without短语可做独立主格，其中with可省略而without不可以。*注：独立主格结构一般放在句首，表示原因时还可放在句末；表伴随状况或补充说明时，相当于一个并列句，通常放于句末。 一、独立主格结构： 1. 名词/代词+形容词 He sat in the front row, his mouth half open. Close to the bank I saw deep pools, the water blue like the sky. 靠近岸时，我看见几汪深池塘，池水碧似蓝天。 2. 名词/代词+现在分词 Winter coming, it gets colder and colder. The rain having stopped, he went out for a walk.
The question having been settled, we wound up the meeting. 也可以The question settled, we wound up the meeting. 但含义稍有差异。前者强调了动作的先后。 We redoubled our efforts, each man working like two. 我们加倍努力，一个人干两个人的活。 3. 名词/代词+过去分词 The job finished, we went home. More time given, we should have done the job much better. *当表人体部位的词做逻辑主语时，不及物动词用现在分词，及物动词用过去分词。 He lay there, his teeth set, his hands clenched, his eyes looking straight up. 他躺在那儿，牙关紧闭，双拳紧握，两眼直视上方。 4. 名词/代词+不定式 We shall assemble at ten forty-five, the procession to start moving at precisely eleven. We divided the work, he to clean the windows and I to sweep the floor.
with用法小结 一、with表拥有某物 Mary married a man with a lot of money . 马莉嫁给了一个有着很多钱的男人。 I often dream of a big house with a nice garden . 我经常梦想有一个带花园的大房子。 The old man lived with a little dog on the lonely island . 这个老人和一条小狗住在荒岛上。 二、with表用某种工具或手段 I cut the apple with a sharp knife . 我用一把锋利的刀削平果。 Tom drew the picture with a pencil . 汤母用铅笔画画。 三、with表人与人之间的协同关系 make friends with sb talk with sb quarrel with sb struggle with sb fight with sb play with sb work with sb cooperate with sb I have been friends with Tom for ten years since we worked with each other, and I have never quarreled with him . 自从我们一起工作以来，我和汤姆已经是十年的朋友了，我们从没有吵过架。 四、with 表原因或理由 John was in bed with high fever . 约翰因发烧卧床。 He jumped up with joy . 他因高兴跳起来。 Father is often excited with wine . 父亲常因白酒变的兴奋。 五、with 表“带来”，或“带有,具有”，在…身上，在…身边之意
comparison的用法解析大全 comparison的意思是比较，比喻，下面我把它的相关知识点整理给大家，希望你们会喜欢！ 释义 comparison n. 比较;对照;比喻;比较关系 [ 复数 comparisons ] 词组短语 comparison with 与…相比 in comparison adj. 相比之下;与……比较 in comparison with 与…比较，同…比较起来 by comparison 相比之下，比较起来 comparison method 比较法 make a comparison 进行比较 comparison test 比较检验 comparison theorem 比较定理 beyond comparison adv. 无以伦比 comparison table 对照表 comparison shopping 比较购物;采购条件的比较调查 paired comp arison 成对比较 同根词 词根： comparing adj. comparative 比较的;相当的 comparable 可比较的;比得上的 adv. comparatively 比较地;相当地 comparably 同等地;可比较地 n.
comparative 比较级;对手 comparing 比较 comparability 相似性;可比较性 v. comparing 比较;对照(compare的ing形式) 双语例句 He liked the comparison. 他喜欢这个比喻。 There is no comparison between the two. 二者不能相比。 Your conclusion is wrong in comparison with their conclusion. 你们的结论与他们的相比是错误的。 comparison的用法解析大全相关文章： 1.by的用法总结大全
《千与千寻》always with me 中文歌词罗马音及日文歌词呼唤心灵深处的某个地方 总想保持令人心动的梦想 悲伤虽然无法尽数 在它对面一定能与你相逢 每次重蹈覆辙时人总是 仅仅知道碧空之蓝 虽然永无止境的道路看起来总在延续 这双手一定可以拥抱光明 别离时 平静的胸怀 虽然从零开始仍要侧耳倾听 活着的不可思议死去的不可思议 花，风，街道都一样 啦啦啦…… 啦啦啦…… 啦啦啦…… 呼唤心灵深处的某个地方 不论何时与我同在去描绘梦想吧 与其道尽悲伤的数目 不如用相同的双唇轻轻歌唱 走向尘封的回忆中总是 听得到不愿忘记的细语 即使是在 被粉碎的镜子上 也会映出崭新的美景 开始的清晨那宁静的窗口 因为将从零开始渐渐被充实 不再追寻大海的彼端 因为那闪光的东西一直就在这里 在我心中被发现 啦啦啦…… 啦啦啦……
1.yo n de i lu / mu ne no do ko ka o ku de 呼んでいる胸のどこか奥で 2、i tsu mo ko ko lo o do lu / yu me wo mi ta i いつも心踊る梦を见たい 3、ka na shi mi wa / ka zo e ki le na i ke le do 悲しみは数えきれないけれど 4、so no mu ko u de ki to / a na ta ni a e lu その向こうできっとあなたに会える 5、ku li ka e su a ya ma chi no / so no ta bi hi to wa 缲り返すあやまちのそのたびひとは 6、ta da a o i so la no / a o i sa wo shi lu ただ青い空の青さを知る 7、ha te shi na ku / mi chi wa tu zu i te mi e lu ke le do 果てしなく道は続いて见えるけれど 8、ko no li yo u te wa / hi ka li wo i da ke lu この両手は光を抱ける 9、sa yo na la no to ki no /shi zu ka na mu ne さよならのときの静かな胸 10、ze lo ni na lu ka la da ga / mi mi wo su ma se lu ゼロになるからだが耳をすませる 11、i ki te i lu fu shi gi / shi n de i ku fu shi gi 生きている不思议死んでいく不思议 ha na mo ka ze mo ma chi mo minna o na ji 12、ららら……（la la la ……） おおお……（o o o ……） るるる……（lu lu lu ……） 13、yo n de i lu / mu ne no do ko ka o ku de 呼んでいる胸のどこか奥で 14、i tsu mo na n do de mo / yu me wo e ga ko u いつも何度でも梦を描こう
介词With 复合结构讲解及练习 with复合结构的作用：with复合结构在句子中作状语，表示原因、时间、条件、伴随、方式等. 1）We sat on the dry grass with our backs to the wall.（作伴随状语） 2）She could not leave with her painful duty unfulfilled.（作原因状语） 3）He lay in bed with his head covered.（作方式状语） 4）Jack soon fell asleep with the light still burning.（作伴随状语） 5）I won't be able to go on holiday with my mother being ill.（作原因状语） 6）He sat with his arms clasped around his knees.（作方式状语） 注：with复合结构在句子中还可以作定语，阅读下面的句子。 1）There was a letter for Lanny with a Swiss stamp on it.（作定语修饰letter） 2）It was a vast stretch of country with cities in the distance.（作定语修饰a stretch of country）1) with +宾语+ 现在（短分词语） When mother went into the house, she found her baby was sleeping in bed, with his lips moving. 当妈妈走进房子的时候，她发现自己的孩子正睡在床上，嘴唇一直在动。 My aunt lives in the room with the windows facing south. 我姑妈住在那间窗户朝南开的房间。 With winter coming on，it's time to buy warm clothes 2)with +宾语+ 过去分词（短语） With more and more forests damaged ,some animals and plants are facing the danger of dying out. 由于越来越多的森林遭到破坏，一些动植物正面临着灭绝的危险。 With his legs broken, he had to lie in bed for a long time. 他双腿都断了，只得长时间躺在床上。 3) with +宾语+ 不定式（短语） * With so many children to look after, the nurse is busy all the time. 有这么多的孩子需要照顾，保育员一直都很忙。 *With a lot of papers to correct, M r. Li didn’t attend the party. 李老师有许多试卷需要批改，所以没有参加聚会。 4) with +宾语+ 副词 * You should read with the radio off. 在看书的时候应该把收音机关掉。 * With the temperature up, we had to open all the windows. 气温上升，我们不得不打开所有的窗户。 5) with +宾语+形容词 *With the window open, I felt a bit cold. 窗户开着，我感到有点冷。 * It was cold outside , the boy ran into the room with his nose red. 外面天气很冷，那个男孩跑进了屋子时，鼻子红红的。 6) with +宾语+ 介词短语 * The woman with a baby in her arms is getting on the bus. 怀里抱着婴儿的那位妇女正在上车。 * John starts to work very clearly in the morning and goes on working until late in the afternoon with a break at midday . 约翰早上开始工作，中午稍作休息后又接着工作到下午稍晚些时候。
哪位大虾能帮我解释一下千与千寻片尾曲always with me 的歌词的意思？ 歌词： 内心深处的呼唤 我想要走进悸动的梦中 虽然悲伤总是会重演 但我一定能在某处与你相逢 人们总是不停犯错 他们只知道天是蓝的 虽然前路渺茫 但他们的双手仍在寻找光明 离别时平静的心 身体归于虚无时的倾听 莫名的生存，莫名的死去 花，风，城市都是如此 啦啦啦~~~ 内心深处的呼唤 让我们不停的画出梦的色彩 比起回忆心中的悲伤 不如用同样的唇轻声歌唱 即使在封锁的回忆中 仍然还有无法忘记的呢喃 即使在粉碎的镜片中 仍然能映出新的景色 晨色初照下的宁静窗台 还有化为虚无的身体 从此我不会越过大洋去寻找 闪耀的所有都在身边 我将自己去追寻 啦啦啦~~~ 我来帮他解答 满意回答 2009-03-29 12:40 这是我一点不成熟的看法，理解的不到位，见谅
宫崎骏大师总是善于刻画孩子内心对世界的看法和对待事物的心理，大家都看过这部电影吧，刚开始的时候，千寻是一个胆小麻木的孩子，这与家长的教育很有关，家长首先就对千寻不是很重视，从简短的几句对白和在通往神明的隧道口就能够看出来，不过虽然千寻对世事比较麻木，但是毕竟她还是生活在象牙塔里的小学生而已，大人们对她自身的想法不重视导致她对自己的不自信和对除自己之外的人的不相信。到这就不难理解这段歌词了。 首先第一句是说有关她的这段非比寻常的经历，对这段故事的不舍，那个如梦一样的世界跟现实世界一样，值得她去怀念。在即将和那个世界作别时，他和小白的对话中，充分显示了他们互相的不舍。其实这个世界，离别才是永恒的，但他们仍然相信他们可以重逢。 第二段，宫崎骏借千寻的角度道破，其实大人们总是太过自私，总是不顾一切的去追逐自己想要的，却忽略了最重要的。 第三段，千寻刚进入那个世界是胆小的，眼睛里充满茫然，而且不懂礼貌，对救了她的锅炉爷爷都没说一句谢谢，去见汤婆婆连门也没敲。可是后来千寻却能把河神给自己的丸子给白龙和无面人吃，足以看出千寻的变化。在回到人类世界之后，千寻回头不舍得望向给自己很多宝贵东西的时候，眼神是那么坚定。这时的千寻实际上已经从神明世界的小千那里继承了某种精神，她已经不再像以前那样迷茫，相信她在这个故事之后，会有她的另一个人生。 之后的几段我感觉和前面的差不多，就是
with[wIT] prep.1.与…(在)一起，带着：Come with me. 跟我一起来吧。/ I went on holiday with my friend. 我跟我朋友一起去度假。/ Do you want to walk home with me? 你愿意和我一道走回家吗 2.(表带有或拥有)有…的，持有，随身带着：I have no money with me. 我没有带钱。/ He is a man with a hot temper. 他是一个脾气暴躁的人。/ We bought a house with a garden. 我们买了一座带花园的房子。/ China is a very large country with a long history. 中国是一个具有历史悠久的大国。3.(表方式、手段或工具)以，用：He caught the ball with his left hand. 他用左手接球。/ She wrote the letter with a pencil. 她用铅笔写那封信。4.(表材料或内容)以，用：Fill the glass with wine. 把杯子装满酒。/ The road is paved with stones. 这条路用石头铺砌。5.(表状态)在…的情况下，…地：He can read French with ease. 他能轻易地读法文。/ I finished my homework though with difficulty. 虽然有困难，我还是做完了功课。6.(表让步)尽管，虽然：With all his money, he is unhappy. 尽管他有钱，他并不快乐。/ With all his efforts, he lost the match. 虽然尽了全力，他还是输了那场比赛。7.(表条件)若是，如果：With your permission, I’ll go. 如蒙你同意我就去。8.(表原因或理由)因为，由于：He is tired with work. 他工作做累了。/ At the news we all jumped with joy. 听到这消息我们都高兴得跳了起来。9.(表时间)当…的时候，在…之后：With that remark, he left. 他说了那话就离开了。/ With daylight I hurried there to see what had happened. 天一亮我就去那儿看发生了什么事。10. (表同时或随同)与…一起，随着：The girl seemed to be growing prettier with each day. 那女孩好像长得一天比一天漂亮。11.(表伴随或附带情况)同时：I slept with the window open. 我开着窗户睡觉。/ Don’t speak with your mouth full. 不要满嘴巴食物说话。12.赞成，同意：I am with you there. 在那点上我同你意见一致。13.由…照看，交…管理，把…放在某处：I left a message for you with your secretary. 我给你留了个信儿交给你的秘书了。/ The keys are with reception. 钥匙放在接待处。14 (表连同或包含)连用，包含：The meal with wine came to ￡8 each. 那顿饭连酒每人8英镑。/ With preparation and marking a teacher works 12 hours a day. 一位老师连备课带批改作业每天工作12小时。15. (表对象或关系)对，关于，就…而言，对…来说：He is pleased with his new house. 他对他的新房子很满意。/ The teacher was very angry with him. 老师对他很生气。/ It’s the same with us students. 我们学生也是这样。16.(表对立或敌对)跟，以…为对手：The dog was fighting with the cat. 狗在同猫打架。/ He’s always arguing with his brother. 他老是跟他弟弟争论。17.(在祈使句中与副词连用)：Away with him! 带他走！/ Off with your clothes! 脱掉衣服！/ Down with your money! 交出钱来！ 【用法】1.表示方式、手段或工具等时(=以，用)，注意不要受汉语意思的影响而用错搭配，如“用英语”习惯上用in English，而不是with English。2.与某些抽象名词连用时，其作用相当于一个副词：with care=carefully 认真地/ with kindness=kindly 亲切地/ with joy=joyfully 高兴地/ with anger=angrily 生气地/ with sorrow=sorrowfully 悲伤地/ with ease=easily 容易地/ with delight=delightedly 高兴地/ with great fluency =very fluently 很流利地3.表示条件时，根据情况可与虚拟语气连用：With more money I would be able to buy it. 要是钱多一点，我就买得起了。/ With better equipment, we could have finished the job even sooner. 要是设备好些，我们完成这项工作还要快些。4.比较with 和as：两者均可表示“随着”，但前者是介词，后者是连词：He will improve as he grows older. 随着年龄的增长，他会进步的。/ People’s ideas change with the change of the times. 时代变了，人们的观念也会变化。5.介词with和to 均可表示“对”，但各自的搭配不同，注意不要受汉语意思的影响而用错，如在kind, polite, rude, good, married等形容词后通常不接介词with而接to。6.复合结构“with+宾语+宾语补足语”是一个很有用的结构，它在句中主要用作状语，表示伴随、原因、时间、条件、方式等；其中的宾语补足语可以是名词、形容词、副词、现在分词、过去分词、不定式、介词短语等：I went out with the windows open. 我外出时没有关窗户。/ He stood before his teacher with his head down. 他低着头站在老师面前。/ He was lying on the bed with all his clothes on. 他和衣躺在床上。/ He died with his daughter yet a schoolgirl. 他去世时，女儿还是个小学生。/ The old man sat there with a basket beside her. 老人坐在那儿，身边放着一个篮子。/ He fell asleep with the lamp burning. 他没熄灯就睡着了。/ He sat there with his eyes closed. 他闭目坐在那儿。/ I can’t go out with all these clothes to wash. 要洗这些衣服，我无法出去了。这类结构也常用于名词后作定语：The boy with nothing on is her son. 没穿衣服的这个男孩子是她儿子。 （摘自《英语常用词多用途词典》金盾出版社） - 1 -