Melody Transcription From Music Audio: Approaches and Evaluation

Abstract

Abstract-Although the process of analyzing an audio recording of a music performance is complex and difficult even for a human listener, there are limited forms of information that may be tractably extracted and yet still enable interesting applications. We discuss melody-roughly, the part a listener might whistle or hum-as one such reduced descriptor of music audio, and consider how to define it, and what use it might be. We go on to describe the results of full-scale evaluations of melody transcription systems conducted in 2004 and 2005, including an overview of the systems submitted, details of how the evaluations were conducted, and a discussion of the results. For our definition of melody, current systems can achieve around 70% correct transcription at the frame level, including distinguishing between the presence or absence of the melody. Melodies transcribed at this level are readily recognizable, and show promise for practical applications.

Audio Cover Song Identification and Similarity: Background, Approaches, Evaluation, and Beyond

“…As a first processing step, these systems need to extract the This is a preliminary draft. The original publication is available at www.springerlink.com: plase go to http://www.springerlink.com/content/a02r21125nw63551/ predominant melody from the raw audio signal [62]. Melody extraction is strongly related to pitch tracking, which itself has a long and continuing history [13].…”

“…However, in the context of complex mixtures, the pitch tracking problem becomes further complicated because, although multiple pitches may be present at the same time, at most just one of them will be the melody. This and many other facets [62] make melody extraction from real-world audio signals a difficult task (see Section 1.4). To refine the obtained representation, cover detection systems usually need to combine a melody extractor with a voice/non-voice detector and other post-processing modules in order to achieve a more reliable representation [68,78,79].…”

Transcribing Debussy's Syrinx dynamics through Linguistic Description: The MUDELD algorithm

“…Although much work has been done in this regard, mostly solving the monophonic case, there are still problems with polyphonic music and particular instruments. The work by Poliner [5], for instance, discusses and evaluates several approaches for transcribing the melody.…”

“…The task is complicated because the exact audio transcription is hardly representable as a meaningful score and thus, some simplifications and educated decisions shall be made to fit the audio in an understandable and realistic score. Whilst most efforts in automatic music transcription have been focused on pitch (which is considered completely solved in the monophonic case, and very advanced in the polyphonic one) [5] and duration, little effort has been done on trying to recover the original dynamic indications (marking the aforementioned deviations in volume) in the score. In fact, to the best of our knowledge there are no works on this matter.…”

Automatic music transcription: challenges and future directions

“…The related problem of melody transcription, i.e. the estimation of the predominant pitch, usually performed by a solo instrument or a lead singer, is not addressed in this paper; for an overview of melody transcription approaches the reader can refer to [108]. Also, the field of content-based music information retrieval, which refers to automated processing of music for search and retrieval purposes and includes the AMT problem, is discussed in [22].…”