Pitch Histograms in Audio and Symbolic Music Information Retrieval

Abstract

In order to represent musical content, pitch and timing information is utilized in the majority of existing work in Symbolic Music Information Retrieval (MIR). Symbolic representations such as MIDI allow the easy calculation of such information and its manipulation. In contrast, most of the existing work in Audio MIR uses timbral and beat information, which can be calculated using automatic computer audition techniques. In this paper, Pitch Histograms are defined and proposed as a way to represent the pitch content of music signals both in symbolic and audio form. This representation is evaluated in the context of automatic musical genre classification. A multiple-pitch detection algorithm for polyphonic signals is used to calculate Pitch Histograms for audio signals. In order to evaluate the extent and significance of errors resulting from the automatic multiple-pitch detection, automatic musical genre classification results from symbolic and audio data are compared. The comparison indicates that Pitch Histograms provide valuable information for musical genre classification. The results obtained for both symbolic and audio cases indicate that although pitch errors degrade classification performance for the audio case, Pitch Histograms can be effectively used for classification in both cases.

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

“…Alternatively, cover song similarity can be assessed by harmonic, rather than melodic, sequences using so-called chroma features or pitch class profiles (PCP) [26,27,63,82]. These mid-level features might provide a more complete, reliable, and straightforward representation than, e.g., melody estimation, as they do not need to tackle the pitch selection and tracking issues outlined above.…”

“…These, in general, differ from traditional systems of their kind in the sense that they also incorporate tonal information (e.g. [48,61,82,90]). However, these systems might also fail in achieving invariance to key or tempo modifications.…”

Mining Digital Music Score Collections: Melody Extraction and Genre Recognition

“…Thirteen statistical features derived from MIDI data are used for this genre discrimination. In (Tzanetakis et al, 2003), pitch features are extracted both from MIDI data and audio data and used separately to classify music within five genres. Pitch histograms regarding to the tonal pitch are used in (Thom, 2000) to describe blues fragments of the saxophonist Charlie Parker.…”

Discrete wavelet packet transform and ensembles of lazy and eager learners for music genre classification

“…Among the different studies in the literature [3,7,9,21,27,28,29,30,32], the work of Tzanetakis et al [30] is particularly relevant. The authors proposed for the first time the concept of beat-histogram and its implementation using a discrete wavelet transform.…”

“…In fact, extracting a symbolic representation from an arbitrary audio signal (polyphonic transcription) is an open research problem, solved only for simple examples [19,20]. However, recent studies show that it is possible to apply signal processing techniques to extract features from audio files and derive reasonably accurate classification by genre [11,16,21,29,30,34]. Other important examples of signal processing techniques applied to the audio domain include discrimination between speech and music [28], tempo and beat estimation [27,30] and audio retrieval by example [9].…”