A Lightweight Instrument-Agnostic Model for Polyphonic Note Transcription and Multipitch Estimation

Bittner, Rachel M.; Bosch, Juan; Rubinstein, David; Meseguer-Brocal, Gabriel; Ewert, Sebastian

doi:10.1109/icassp43922.2022.9746549

Cited by 20 publications

(6 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed technique outperforms existing solutions in terms of low/high-frequency separation accuracy. Bittner et al (2022) introduced a lightweight neural network for instrument transcription that supports multiple vocal outputs, which can be extended to many instruments (including the human voice) in this study. The multi-output structure of our model increases frame-level note accuracy by simultaneously predicting frame-level onsets, multiple pitches, and note activations.…”

Section: Pitchmentioning

confidence: 99%

“…Coorevits et al (2019) again returned to the musical performance itself, examining the many effects that changes in musical tempo have on the 'performance state' or the articulation of the performer's movements. Bittner et al (2022) offered a way to understand musical tempo beyond listening to music by developing a Visual project and found that the number of notes per time unit and tempo also mattered. Pressnitzer et al (2000) concluded from a psychoacoustic perspective that psychoacoustic roughness increases when non-tonal orchestral timbres reduce musical tension.…”

Section: Tempomentioning

confidence: 99%

See 1 more Smart Citation

A Literature Review on Music Parameter Extraction and Visualization

Jin,

Zairul,

Salih

2024

IJARBSS

View full text Add to dashboard Cite

Music visualization research is extremely complex and dynamic. Several researchers have applied various methods to persevere in the study of all aspects that make up music. The complexity of music also includes factors such as waveform, frequency, pitch, rhythm, tempo, timbre, and chords. Researchers in recent years have studied the extraction of single elements, visualization, or cross-discipline for these aspects. As far as the current research is concerned, most of the disciplines related to music visualization are focused on computers, psychology, sports science, and other related disciplines. Research on the elements of music itself has focused on music visualization, music element extraction, music association, music emotion, and the study of several important aspects of music, such as waveform, frequency, pitch, rhythm, tempo, timbre, and chord. After reviewing the research, this paper has found that with the continuous development of science and technology, music visualization has a progressive intersection with computer science, artificial intelligence, and neural networks. Thus, future research can continue to interact more with computer science.

show abstract

Section: Pitchmentioning

confidence: 99%

Section: Tempomentioning

confidence: 99%

A Literature Review on Music Parameter Extraction and Visualization

Jin,

Zairul,

Salih

2024

IJARBSS

View full text Add to dashboard Cite

show abstract

“…Indeed, all pieces included in the analyses included multiple pitch lines. For polyphonic pitch estimation, we used pretrained models based on neural networks implemented in the python library basicpitch [26]. For the model input, we calculated the mean of all estimated pitches at each time point in order to receive one aggregated pitch score.…”

Section: Plos Onementioning

confidence: 99%

“…The model is implemented in the python library midi-miner [27]. The midi input required for the estimation of tonal tension was created using the python package basic-pitch [26] that includes a state-of-the-art polyphonic pitch estimation method. According to the model by [11], tonal tension can be quantified using three metrics.…”

Section: Plos Onementioning

confidence: 99%

TenseMusic: An automatic prediction model for musical tension

Barchet,

Rimmele,

Pelofi

2024

PLoS ONE

View full text Add to dashboard Cite

The perception of tension and release dynamics constitutes one of the essential aspects of music listening. However, modeling musical tension to predict perception of listeners has been a challenge to researchers. Seminal work demonstrated that tension is reported consistently by listeners and can be accurately predicted from a discrete set of musical features, combining them into a weighted sum of slopes reflecting their combined dynamics over time. However, previous modeling approaches lack an automatic pipeline for feature extraction that would make them widely accessible to researchers in the field. Here, we present TenseMusic: an open-source automatic predictive tension model that operates with a musical audio as the only input. Using state-of-the-art music information retrieval (MIR) methods, it automatically extracts a set of six features (i.e., loudness, pitch height, tonal tension, roughness, tempo, and onset frequency) to use as predictors for musical tension. The algorithm was optimized using Lasso regression to best predict behavioral tension ratings collected on 38 Western classical musical pieces. Its performance was then tested by assessing the correlation between the predicted tension and unseen continuous behavioral tension ratings yielding large mean correlations between ratings and predictions approximating r = .60 across all pieces. We hope that providing the research community with this well-validated open-source tool for predicting musical tension will motivate further work in music cognition and contribute to elucidate the neural and cognitive correlates of tension dynamics for various musical genres and cultures.

show abstract

“…In the monophonic BWE system, the DDSP model generates a harmonic signal based on a single f 0 estimated from a monophonic pitch estimator [43]. Now that we are in a polyphonic context, we use a state-of-the-art multipitch estimator [45] which outputs a maximum of I different fundamental frequencies f i 0 . Considering that this multi-pitch estimator has a rather good performance, we propose to iteratively use the monophonic DDSP model DDSP-mono-dec in a cyclic manner, as illustrated in Fig.…”

Section: Cyclic Monophonic Decodermentioning

confidence: 99%

Efficient bandwidth extension of musical signals using a differentiable harmonic plus noise model

Grumiaux,

Lagrange

2023

J AUDIO SPEECH MUSIC PROC.

View full text Add to dashboard Cite

The task of bandwidth extension addresses the generation of missing high frequencies of audio signals based on knowledge of the low-frequency part of the sound. This task applies to various problems, such as audio coding or audio restoration. In this article, we focus on efficient bandwidth extension of monophonic and polyphonic musical signals using a differentiable digital signal processing (DDSP) model. Such a model is composed of a neural network part with relatively few parameters trained to infer the parameters of a differentiable digital signal processing model, which efficiently generates the output full-band audio signal. We first address bandwidth extension of monophonic signals, and then propose two methods to explicitly handle polyphonic signals. The benefits of the proposed models are first demonstrated on monophonic and polyphonic synthetic data against a baseline and a deep-learning-based ResNet model. The models are next evaluated on recorded monophonic and polyphonic data, for a wide variety of instruments and musical genres. We show that all proposed models surpass a higher complexity deep learning model for an objective metric computed in the frequency domain. A MUSHRA listening test confirms the superiority of the proposed approach in terms of perceptual quality.

show abstract

A Lightweight Instrument-Agnostic Model for Polyphonic Note Transcription and Multipitch Estimation

Cited by 20 publications

References 16 publications

A Literature Review on Music Parameter Extraction and Visualization

A Literature Review on Music Parameter Extraction and Visualization

TenseMusic: An automatic prediction model for musical tension

Efficient bandwidth extension of musical signals using a differentiable harmonic plus noise model

Contact Info

Product

Resources

About