Polyphonic piano transcription using non-negative Matrix Factorisation with group sparsity

O'Hanlon, Ken; Plumbley, Mark D.

doi:10.1109/icassp.2014.6854173

Cited by 15 publications

(16 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is usually not true in practice, where the instrument models/sources at test time are unknown and usually do not coincide with the instruments used for training. A majority of experiments with the MAPS dataset train and test model on disjoint instrument types [3], [2], [44]. We thus perform a second set of experiments to compare performance of the different neural network acoustic models in a more realistic setting.…”

Section: A Datasetmentioning

confidence: 99%

An End-to-End Neural Network for Polyphonic Piano Music Transcription

Sigtia

Benetos

Dixon

2016

IEEE/ACM Trans. Audio Speech Lang. Process.

215

178

View full text Add to dashboard Cite

We present a supervised neural network model for polyphonic piano music transcription. The architecture of the proposed model is analogous to speech recognition systems and comprises an acoustic model and a music language model. The acoustic model is a neural network used for estimating the probabilities of pitches in a frame of audio. The language model is a recurrent neural network that models the correlations between pitch combinations over time. The proposed model is general and can be used to transcribe polyphonic music without imposing any constraints on the polyphony. The acoustic and language model predictions are combined using a probabilistic graphical model. Inference over the output variables is performed using the beam search algorithm. We perform two sets of experiments. We investigate various neural network architectures for the acoustic models and also investigate the effect of combining acoustic and music language model predictions using the proposed architecture. We compare performance of the neural network based acoustic models with two popular unsupervised acoustic models. Results show that convolutional neural network acoustic models yields the best performance across all evaluation metrics. We also observe improved performance with the application of the music language models. Finally, we present an efficient variant of beam search that improves performance and reduces runtimes by an order of magnitude, making the model suitable for real-time applications.

show abstract

Section: A Datasetmentioning

confidence: 99%

An End-to-End Neural Network for Polyphonic Piano Music Transcription

Sigtia

Benetos

Dixon

2016

IEEE/ACM Trans. Audio Speech Lang. Process.

215

178

View full text Add to dashboard Cite

show abstract

“…Improvements of ∼ 5% for β (0.5) and ∼ 4% for KL are observed using dictionary tuning. In the case of β (0.5) , all metrics increase by 1% relative to β-NMD using the optimal dictionary, and also by ∼ 2% relative to our previous results [17], using the monotonic descent algorithm.…”

Section: Experiments Cmentioning

confidence: 74%

“…We then employ this approach in a dictionary tuning method, applied to a restructured version of the harmonic dictionary used in [4], whereby the hard harmonic constraint is dropped. Part of this work was described in [17] and is augmented here through monotonic descent algorithm with scale invariant group sparse penalty and further evaluation. We also propose a new onset detector for NMF-based AMT.…”

Section: A Contributions Of This Papermentioning

confidence: 99%

See 1 more Smart Citation

Non-Negative Group Sparsity with Subspace Note Modelling for Polyphonic Transcription

O'Hanlon

Nagano

Keriven

et al. 2016

IEEE/ACM Trans. Audio Speech Lang. Process.

Self Cite

View full text Add to dashboard Cite

Abstract-Automatic Music Transcription (AMT) can be performed by deriving a pitch-time representation through decomposition of a spectrogram with a dictionary of pitch-labelled atoms. Typically, Non-negative Matrix Factorisation (NMF) methods are used to decompose magnitude spectrograms. One atom is often used to represent each note. However, the spectrum of a note may change over time. Previous research considered this variability using different atoms to model specific parts of a note, or large dictionaries comprised of datapoints from the spectrograms of full notes. In this paper the use of subspace modelling of note spectra is explored, with group sparsity employed as a means of coupling activations of related atoms into a pitched subspace.Stepwise and gradient-based methods for non-negative group sparse decompositions are proposed. Finally, a group sparse NMF approach is used to tune a generic harmonic subspace dictionary, leading to improved NMF-based AMT results.

show abstract

“…The proposed approach gives improved performance for the AMT task and is simple to implement, which makes it easy to add to existing NMF-based systems. For future work, similar ideas could be applied to more complicated NMF AMT algorithms involving group sparsity [22] or non-negative dictionary learning. So far, we have used a setting of 0.5 for β which has previously found to work well for AMT, however it is possible that the performance could be further improved by fine-tuning the β parameter for the low-rank approach.…”

Section: A Results and Discussionmentioning

confidence: 99%

Automatic music transcription using low rank non-negative matrix decomposition

O'Brien

Plumbley

2017

2017 25th European Signal Processing Conference (EUSIPCO)

Self Cite

View full text Add to dashboard Cite

Abstract-Automatic Music Transcription (AMT) is concerned with the problem of producing the pitch content of a piece of music given a recorded signal. Many methods rely on sparse or low rank models, where the observed magnitude spectra are represented as a linear combination of dictionary atoms corresponding to individual pitches. Some of the most successful approaches use Non-negative Matrix Decomposition (NMD) or Factorization (NMF), which can be used to learn a dictionary and pitch activation matrix from a given signal. Here we introduce a further refinement of NMD in which we assume the transcription itself is approximately low rank. The intuition behind this approach is that the total number of distinct activation patterns should be relatively small since the pitch content between adjacent frames should be similar. A rank penalty is introduced into the NMD objective function and solved using an iterative algorithm based on Singular Value thresholding. We find that the low rank assumption leads to a significant increase in performance compared to NMD using β-divergence on a standard AMT dataset.

show abstract

Polyphonic piano transcription using non-negative Matrix Factorisation with group sparsity

Cited by 15 publications

References 16 publications

An End-to-End Neural Network for Polyphonic Piano Music Transcription

An End-to-End Neural Network for Polyphonic Piano Music Transcription

Non-Negative Group Sparsity with Subspace Note Modelling for Polyphonic Transcription

Automatic music transcription using low rank non-negative matrix decomposition

Contact Info

Product

Resources

About