Multipitch Estimation of Piano Sounds Using a New Probabilistic Spectral Smoothness Principle

By Valentin Emiya, Roland Badeau, Bertrand David. IEEE Transactions on Audio, Speech, and Language Processing, 2010

373 citations recorded by the source corpus.

A new method for the estimation of multiple concurrent pitches in piano recordings is presented. It addresses the issue of overlapping overtones by modeling the spectral envelope of the overtones of each note with a smooth autoregressive model. For the background noise, a moving-average model is used and the combination of both tends to eliminate harmonic and sub-harmonic erroneous pitch estimations. This leads to a complete generative spectral model for simultaneous piano notes, which also explicitly includes the typical deviation from exact harmonicity in a piano overtone series. The pitch set which maximizes an approximate likelihood is selected from among a restricted number of possible pitch combinations as the one. Tests have been conducted on a large homemade database called MAPS, composed of piano recordings from a real upright piano and from high-quality samples.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

View Paper: https://www.semanticscholar.org/paper/c9d26a0d690b354ff4e2853fb41a36a3fe81ae47

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

Invariances and Data Augmentation for Supervised Music Transcription (2017, background): “Disklavier piano is wired in this way, and has lead to the creation of datasets such as MAPS [5].”
Neural Music Synthesis for Flexible Timbre Control (2018, methodology): “de, which are also used in the MAPS Database [25]; these MIDI files are recorded from actual performances and contain expressive timing and velocity information.”
An End-to-End Neural Network for Polyphonic Piano Music Transcription (2015, methodology): “We evaluate the proposed model on the MAPS dataset [43].”
Enabling Factorized Piano Music Modeling and Generation with the MAESTRO Dataset (2018, methodology): “After training on the MAESTRO training split for 670k steps, we achieved state of the art results described in table 4 for the MAPS dataset.”
Automatic Music Transcription: An Overview (2019, methodology): “The largest publicly available data set [11] contains several hours of piano music—but all recorded on only seven different synthesizer-based and real pianos.”
Onsets and Frames: Dual-Objective Piano Transcription (2017, methodology): “We use the MAPS dataset [9] which contains audio and corresponding annotations of isolated notes, chords, and complete piano pieces.”
Automatic music transcription: challenges and future directions (2013, methodology): “problem can then be viewed as a maximum a posteriori (MAP) estimation problem [43]:”