Assistants in record shops are used to receiving "humming queries": a customer comes into the store humming a song he wants, but cannot remember either the title or the artist. Knowledgeable staff are often able to name that tune and make a sale. Hummers, though, can be both off-key and off-track. Frequently, therefore, the cash register stays closed and the customer goes away disappointed. A new piece of software may change this. If Online Music Recognition and Searching (OMRAS) is successful, it will be possible to hum a half-remembered tune into a computer and get a match.

 

OMRAS, which has just been unveiled at the International Symposium on Music Information Retrieval, in Paris, is the brainchild of a group of researchers from the Universities of London, Indiana and Massachusetts. Music-recognition programs exist already, of course. Mobile-phone users, for instance, can dial into a system called Shazam, hold their phones to a source of music, and then wait for the title and artist to be texted back to them.

 

Shazam and its cousins work by matching sounds directly to recordings, several million of them, stored in a central database. For Shazam to make a match, though, the music source must be not just similar to, but actually identical with, one of the filed recordings. OMRAS, by contrast, analyses the music. That means it can make a match between different interpretations of the same piece. According to Mark Sandler, the leader of the British side of the project, the program would certainly be able to match performances of the same work by an amateur and a professional pianist. It should also pass the humming-query test.

 

The musical analysis performed by OMRAS is unlike any that a musicologist would recognise. A tune is first digitised, so that it can be processed. It is then subject to such mathematical indignities as wavelet decomposition, multi-resolution Fourier analysis, polyphase filtering and discrete cosine transformation. The upshot is a mathematical model of the sound that contains the essence of the original, without such distractions as style and quality. That essence can then be compared with a library of known essences and a match made. Unlike Shazam, only one library reference per tune is needed.

 

So far, Dr Sandler and his colleagues have been restricted to modelling classical music. Their 3,000-strong database includes compositions by Bach, Beethoven and Mozart. Worries about copyright mean that they have not yet gained access to company archives of pop music, though if the companies realise that the consequence of more humming queries being answered is more sales, this may change. On top of that, OMRAS could help to prevent accidental copyright infringements, in which a composer lifts somebody else's work without realising his inspiration is second-hand. Or, more cynically, it will stop people claiming that any infringement was accidental. There is little point in doing that when a quick check on the Internet could have set your mind at rest that your magnum opus really was yours.

 

注（1）：本文选自Economist；10/19/2002, p77, 2/3p, 1c；

注（2）：本文习题命题模仿对象2000年真题text 3第1题（1），2001年真题text 4第2题（2），2004年真题text 3第4题（3）；2003年真题text 1第4题（4），2002年真题text 3第5题（5）；

 

 

[A] a comparison of two music-recognition programs

[B] an introduction of a new software

[C] a survey of the music recognition and searching market

[D] an analysis of the functions of music recognition softwares

 

 

[A] its ability to analyze music

[B] its large database

[C] its matching speed

[D] its ability to match music of different pieces

 

 

[A] last step

[B] final result

[C] goal

[D] program

 

 

[A] OMRAS will facilitate copyright infringements

[B] OMRAS researchers are fans of classical music

[C] composers can get more inspiration with the help of OMRAS

[D] music companies are yet to realize the value of OMRAS

 

 

[A] optimistic

[B] uncertain

[C] indifferent

[D] skeptical

 

答案：BABDA

 

本篇文章是一篇说明文，介绍了一款最新发布的“联机音乐识别和查询系统”。第一段通过一个生动的例子介绍这种系统的功能。第二段和第三段将这种系统和其他的产品的工作原理进行了比较。第四段介绍了这种新产品的音乐分析方法。最后一段介绍了有关音乐版权问题以及这个系统在版权领域所能发挥的作用。

 

query[5kwiEri] n. 询问

off-key（唱歌）跑调

off-track 唱错曲子

unveil[Qn5veil] v. 使公诸于众

symposium[sim5pEuziEm] n. （专家、学者的）研讨会，专题讨论会，座谈会

retrieval[ri5tri:vEl] n. 检索；从内存或其它存储设备中获取信息的过程

brainchild[5breIntFaIld] n. 脑力劳动成果（指计划、发明等）

text[tekst] v. 以文本形式发送

musicologist[7mju:zi5kClEdVist] n. 音乐学者

digitize[5didVitaiz] v. [计]将资料数字化

wavelet[5weivlit] n. 微[子, 弱, 小]波

decomposition[7di:kCmpE5ziFEn] n. 分解

multi-resolution  n. 多重分辨率

Fourier analysis  傅立叶分析

polyphase[5pClifeiz] adj.多相的

filtering[5filtEriN] n. 过滤, 滤除

discrete[dis5kri:t] adj. 离散的

cosine[5kEusain] n. [数]余弦

transformation[7trAnsfE5meiFEn] n. 变化, 转化

upshot[5QpFCt] n. 结果

infringement[in5frindVmEnt] n. 侵权侵害一项权利或特权

magnum opus[mA^nEm 5EJpEs] n. <拉> 巨著, 杰作, 代表作, 主要作品

 

主体句式：worries mean that …

结构分析：这是一个复杂句，句中包含一个that引导的宾语从句，这个从句中有一个词组gain access to，意思是“可以进入”，此外，句中还有一个由though引导的让步状语从句，在这个从句里又有一个if引导的条件状语从句，而在这个条件状语从句里又有一个that引导的宾语从句。

句子译文：对于版权的担忧意味着他们还无法进入各公司的流行音乐资料库，虽然如果公司意识到回答更多的哼唱问询就可以带来更多销量的话，这种状况也许会有所改变。

2. There is little point in doing that when a quick check on the Internet could have set your mind at rest that your magnum opus really was yours.

主体句式：There is little point…

结构分析：这是一个复杂句，句子主体结构是一个惯用表达“there is little point in doing somethi （责任编辑：城市网）