Previous work has shown that neural architectures are able to perform optical music recognition (OMR) on monophonic and homophonic music with high accuracy. However, piano and orchestral scores frequently exhibit polyphonic passages, which add a second dimension to the task. Monophonic and homophonic music can be described as homorhythmic, or having a single musical rhythm. Polyphonic music, on the other hand, can be seen as having multiple rhythmic sequences, or voices, concurrently. We first introduce a workflow for creating large-scale polyphonic datasets suitable for end-to-end recognition from sheet music publicly available on the MuseScore forum. We then propose two novel formulations for end-to-end polyphonic OMR -- one treating the problem as a type of multi-task binary classification, and the other treating it as multi-sequence detection. Building upon the encoder-decoder architecture and an image encoder proposed in past work on end-to-end OMR, we propose two novel decoder models -- FlagDecoder and RNNDecoder -- that correspond to the two formulations. Finally, we compare the empirical performance of these end-to-end approaches to polyphonic OMR and observe a new state-of-the-art performance with our multi-sequence detection decoder, RNNDecoder.
A Query-By-Humming (QBH) system constitutes a particular case of music information retrieval where the input is a user-hummed melody and the output is the original song which contains that melody. A typical QBH system consists of melody extraction and candidate melody retrieval. For melody extraction, accurate note transcription is the key enabling technology. However, current transcription methods are unable to definitively capture the melody and address inaccuracies in user-hummed queries. In this paper, we incorporate Total Variation Regularization (TVR) to denoise queries. This approach accounts for user error in humming without loss of meaningful data and reliably captures the underlying melody. For candidate melody retrieval, we employ a deep learning approach to time series classification using a Fully Convolutional Neural Network. The trained network classifies the incoming query as belonging to one of the target songs. For our experiments, we use Roger Jang's MIR-QBSH dataset which is the standard MIREX dataset. We demonstrate that inclusion of TVR denoised queries in the training set enhances the overall accuracy of the system to 93% which is higher than other state-of-the-art QBH systems.
The task of identifying emotions from a given music track has been an active pursuit in the Music Information Retrieval (MIR) community for years. Music emotion recognition has typically relied on acoustic features, social tags, and other metadata to identify and classify music emotions. The role of lyrics in music emotion recognition remains under-appreciated in spite of several studies reporting superior performance of music emotion classifiers based on features extracted from lyrics. In this study, we use the transformer-based approach model using XLNet as the base architecture which, till date, has not been used to identify emotional connotations of music based on lyrics. Our proposed approach outperforms existing methods for multiple datasets. We used a robust methodology to enhance web-crawlers' accuracy for extracting lyrics. This study has important implications in improving applications involved in playlist generation of music based on emotions in addition to improving music recommendation systems.
Music preference was reported as a factor, which could elicit innermost music emotion, entailing accurate ground-truth data and music therapy efficiency. This study executes statistical analysis to investigate the distinction of music preference through familiarity scores, response times (response rates), and brain response (EEG). Twenty participants did self-assessment after listening to two types of popular music's chorus section: music without lyrics (Melody) and music with lyrics (Song). We then conduct a music preference classification using a support vector machine (SVM) with the familiarity scores, the response rates, and EEG as the feature vectors. The statistical analysis and SVM's F1-score of EEG are congruent, which is the brain's right side outperformed its left side in classification performance. Finally, these behavioral and brain studies support that preference, familiarity, and response rates can contribute to the music emotion experiment's design to understand music, emotion, and listener. Not only to the music industry, the biomedical, and healthcare industry can also exploit this experiment to collect data from patients to improve the efficiency of healing by music.
Frequency tagging is a powerful approach to investigate the neural processing of sensory features, and is recently adapted to study the neural correlates of superordinate structures, i.e., chunks, in complex sequences such as speech and music. The nesting of sequence structures, the necessity to control the periodicity in sensory features, and the low-frequency nature of sequence structures pose new challenges for data analysis and interpretation. Here, I discuss how to interpret the frequency of a sequential structure, and factors that need to be considered when analyzing the periodicity in a signal. Finally, a safe procedure is recommended for the analysis of frequency-tagged responses.
Close your eyes and listen to music, one can easily imagine an actor dancing rhythmically along with the music. These dance movements are usually made up of dance movements you have seen before. In this paper, we propose to reproduce such an inherent capability of the human-being within a computer vision system. The proposed system consists of three modules. To explore the relationship between music and dance movements, we propose a cross-modal alignment module that focuses on dancing video clips, accompanied on pre-designed music, to learn a system that can judge the consistency between the visual features of pose sequences and the acoustic features of music. The learned model is then used in the imagination module to select a pose sequence for the given music. Such pose sequence selected from the music, however, is usually discontinuous. To solve this problem, in the spatial-temporal alignment module we develop a spatial alignment algorithm based on the tendency and periodicity of dance movements to predict dance movements between discontinuous fragments. In addition, the selected pose sequence is often misaligned with the music beat. To solve this problem, we further develop a temporal alignment algorithm to align the rhythm of music and dance. Finally, the processed pose sequence is used to synthesize realistic dancing videos in the imagination module. The generated dancing videos match the content and rhythm of the music. Experimental results and subjective evaluations show that the proposed approach can perform the function of generating promising dancing videos by inputting music.
Musical instruments recognition is a widely used application for music information retrieval. As most of previous musical instruments recognition dataset focus on western musical instruments, it is difficult for researcher to study and evaluate the area of traditional Chinese musical instrument recognition. This paper propose a traditional Chinese music dataset for training model and performance evaluation, named ChMusic. This dataset is free and publicly available, 11 traditional Chinese musical instruments and 55 traditional Chinese music excerpts are recorded in this dataset. Then an evaluation standard is proposed based on ChMusic dataset. With this standard, researchers can compare their results following the same rule, and results from different researchers will become comparable.
This paper addresses the problem of cross-modal musical piece identification and retrieval: finding the appropriate recording(s) from a database given a sheet music query, and vice versa, working directly with audio and scanned sheet music images. The fundamental approach to this is to learn a cross-modal embedding space with a suitable similarity structure for audio and sheet image snippets, using a deep neural network, and identifying candidate pieces by cross-modal near neighbour search in this space. However, this method is oblivious of temporal aspects of music. In this paper, we introduce two strategies that address this shortcoming. First, we present a strategy that aligns sequences of embeddings learned from sheet music scans and audio snippets. A series of experiments on whole piece and fragment-level retrieval on 24 hours worth of classical piano recordings demonstrates significant improvement. Second, we show that the retrieval can be further improved by introducing an attention mechanism to the embedding learning model that reduces the effects of tempo variations in music. To conclude, we assess the scalability of our method and discuss potential measures to make it suitable for truly large-scale applications.
While there are many music datasets with emotion labels in the literature, they cannot be used for research on symbolic-domain music analysis or generation, as there are usually audio files only. In this paper, we present the EMOPIA (pronounced `yee-m\`{o}-pi-uh') dataset, a shared multi-modal (audio and MIDI) database focusing on perceived emotion in pop piano music, to facilitate research on various tasks related to music emotion. The dataset contains 1,087 music clips from 387 songs and clip-level emotion labels annotated by four dedicated annotators. Since the clips are not restricted to one clip per song, they can also be used for song-level analysis. We present the methodology for building the dataset, covering the song list curation, clip selection, and emotion annotation processes. Moreover, we prototype use cases on clip-level music emotion classification and emotion-based symbolic music generation by training and evaluating corresponding models using the dataset. The result demonstrates the potential of EMOPIA for being used in future exploration on piano emotion-related MIR tasks.
We present an empirical study on embedding the lyrics of a song into a fixed-dimensional feature for the purpose of music tagging. Five methods of computing token-level and four methods of computing document-level representations are trained on an industrial-scale dataset of tens of millions of songs. We compare simple averaging of pretrained embeddings to modern recurrent and attention-based neural architectures. Evaluating on a wide range of tagging tasks such as genre classification, explicit content identification and era detection, we find that averaging word embeddings outperform more complex architectures in many downstream metrics.