英语翻译Music transcription could be defined as an act oflisteni

英语翻译Music transcription could be defined as an act oflistening to a piece of music and writing down musicnotation for the piece.Transcription of polyphonicmusic (polyphonic pitch recognition) converts anacoustical waveform into a parametric representation,where notes,their pitches,starting times anddurations are extracted from the signal.Transcriptionis a difficult cognitive task and is not inherent inhuman perception of music,although it can belearned.It is also a very difficult problem for currentcomputer systems.Separating notes from a mixture ofother sounds,which may include notes played by thesame or different instruments or simply backgroundnoise requires robust algorithms with performancethat should degrade gracefully when noise increases.In recent years,several transcription systems havebeen developed.Some of them are targeted totranscription of music played on specific instruments[3-5],while others are general transcription systems[1-2].All of them share several commoncharacteristics.In the beginning,they calculate atime-frequency representation of the musical signal.In the next step,the time-frequency representation isrefined by locating partials in the signal.To trackpartials,most systems use ad hoc algorithms such aspeak picking and peak connecting.Partial tracks arethen grouped into notes with different algorithmsrelying on cues such as common onset time andharmonicity.Some authors use templates ofinstrument tones in this process [1,3],as well ashigher-level knowledge of music,such asprobabilities of chord transitions [2].Recognizing notes in a signal is a typical patternrecognition task and we were surprised that fewcurrent systems use machine learning algorithms in the transcription process.Therefore,our motivationwas to develop a transcription system based on neuralnetworks,which have proved to be useful in a varietyof pattern recognition tasks.We tried to avoid explicitsymbolic algorithms,and instead used connectionistapproaches in different parts of our system.The name of our transcription system is SONIC.Transcription is a difficult task,so we put one majorconstraint on the system:it only transcribes pianomusic,so piano should be the only instrument in theanalyzed musical signal.We didn't make any otherassumptions about the signal,such as maximalpolyphony,minimal note length,style of transcribedmusic or the type of piano used.The system takes anacoustical waveform of a piano recording (44.1 kHzsampling rate,16 bit resolution) as its input.Stereorecordings are converted to mono.The output of thesystem is a MIDI file containing the transcription.Notes,their starting times,durations and loudness' areextracted from the signal.额 不要哪种GOOGLE啊 百度啊 上的翻译 哪些网站的翻译很不精准 最好能翻的精准点。WOSHIWANGZHIQI在哪里找的翻译啊