Musical Instrument Digital Interface Files Research Articles

Acapella-based music generation with sequential models utilizing discrete cosine transform

Making musical instruments that accompany vocals in a song depends on the mood quality and the music composer’s creativity. The model created by other researchers has restrictions that include being limited to musical instrument digital interface files and relying on recurrent neural networks (RNN) or Transformers for the recursive generation of musical notes. This research offers the world’s first model capable of automatically generating musical instruments accompanying human vocal sounds. The model we created is divided into three types of sound input: short input, combed input, and frequency sound based on the discrete cosine transform (DCT). By combining the sequential models such as Autoencoder and gated recurrent unit (GRU) models, we will evaluate the performance of the resulting model in terms of loss and creativity. The best model has a performance evaluation that resulted in an average loss of 0.02993620155. The hearing test results from the sound output produced in the frequency range 0-1,600 Hertz can be heard clearly, and the tones are quite harmonious. The model has the potential to be further developed in future research in the field of sound processing.

The Billboard Melodic Music Dataset (BiMMuDa)

We introduce the Billboard Melodic Music Dataset (BiMMuDa), which contains the lead vocal melodies of the top five songs of each year from 1950 to 2022 according to the Billboard year-end singles charts. In this article, the dataset’s compilation process and attributes are described in detail. The melody from each of the 371 songs was transcribed manually in full to create 371 MIDI (musical instrument digital interface) files, and then melodies from the songs’ different sections (e.g., verses, choruses) were exported into separate files to create an additional 1,133 MIDI files of shorter melodies. Lyrics to the songs are provided separately from the melodic transcriptions. This report includes comprehensive descriptions and graphical representations of the available metadata per song and per melody. Analysis of verse and chorus melodies revealed structural differences between them: chorus melodies have significantly fewer notes and lower note density, but larger melodic intervals on average. Whether added to existing datasets or used as a complete dataset, BiMMuDa can serve as ground truth data for a variety of MIR tasks as well as provide insight into the study of Western pop melody.

Creating Zhudi and Guzheng sampler based on Nyquist

Nyquist is a strong and basic software for making music with algorithms. With the western instruments available as functions inside the software, this research aims to create sampler of traditional Chinese instruments that can read Musical Instrument Digital Interface (MIDI) files based on the present functions. This research mainly focuses on two instruments: Guzheng, a traditional plucked instrument in China and Zhudi, the oldest Chinese musical instrument found so far and the most representative woodwind in the Chinese musical instrument family. The methods used in this research is built on original functions, envelopes, filters in the software. After creating the samplers, the MIDI file users uploaded will be converted into scores inside Nyquist for the samplers to read. The separation of the MIDI file is done by if and else clauses, splitting the samplers to each focus on a certain part of the song. Based on the methods above, this research yields a sample song that has a MIDI file originally in the timbre of piano applied to both samplers at the same time. The outcome yielded applies two samplers separated at a certain pitch, forming a contrast between the main melody and the chords. This research successfully achieves its intended purpose, creating a bridge that connects western music culture with traditional eastern music culture.

GiantMIDI-Piano: A Large-Scale MIDI Dataset for Classical Piano Music

Symbolic music datasets are important for music information retrieval and musical analysis. However, there is a lack of large-scale symbolic datasets for classical piano music. In this article, we describe the creation of the GiantMIDI-Piano (GP) dataset containing 38,700,838 transcribed notes and 10,855 unique solo piano works composed by 2,786 composers. We extract the names of music works and the names of composers from the International Music Score Library Project (IMSLP). We search and download their corresponding audio recordings from the Internet. We further create a curated subset containing 7,236 works composed by 1,787 composers where the titles of downloaded audio recordings contain the surnames of composers. We apply a convolutional neural network to detect solo piano works. Then, we transcribe those solo piano recordings into Musical Instrument Digital Interface (MIDI) files using a high-resolution piano transcription system. Each transcribed MIDI file contains the onset, offset, pitch, and velocity attributes of piano notes and pedals. GiantMIDI-Piano includes 90% live performance MIDI files and 10% sequence input MIDI files. We analyse the statistics of GiantMIDI-Piano and show pitch class, interval, trichord, and tetrachord frequencies of six composers from different eras to show that GiantMIDI-Piano can be used for musical analysis. We evaluate the quality of GiantMIDI-Piano in terms of solo piano detection F1 scores, metadata accuracy, and transcription error rates. We release the source code for acquiring the GiantMIDI-Piano dataset at <a href="https://github.com/bytedance/GiantMIDI-Piano" target="_blank">https://github.com/bytedance/GiantMIDI-Piano</a>.

A Music Genre Classification Method Based on Deep Learning

Digital music resources have exploded in popularity since the dawn of the digital music age. The music genre is an important classification to use when describing music. The function of music labels in discovering and separating digital music resources is crucial. In the face of a huge music database, relying on manual annotation to classify will consume a lot of cost and time, which cannot meet the needs of the times. The following are the paper’s primary research findings and innovations: to better describe the music, this article will be divided into multiple local musical instrument digital interface (MIDI) music passages, playing style close by analyzing passages, passages feature extracting, and feature sequence of passages. Extraction of note feature matrix, extraction of topic and segment division based on note feature matrix, research and extraction of effective features based on segment theme, and composition of feature sequence are all part of the process. Because of the shallow structure of standard classification methods, it is difficult for classifiers to learn temporal and semantic information about music. This research investigates recurrent neural networks (RNN) and attention using the distinctive sequence of input MIDI segments. To create data sets and conduct music categorization tests, collect 1920 MIDI files with genre labels from the Internet. The method for music classification is validated when it is combined with the experimental accuracy of equal length segment categorization.

DBTMPE: Deep Bidirectional Transformers-Based Masked Predictive Encoder Approach for Music Genre Classification

Music is a type of time-series data. As the size of the data increases, it is a challenge to build robust music genre classification systems from massive amounts of music data. Robust systems require large amounts of labeled music data, which necessitates time- and labor-intensive data-labeling efforts and expert knowledge. This paper proposes a musical instrument digital interface (MIDI) preprocessing method, Pitch to Vector (Pitch2vec), and a deep bidirectional transformers-based masked predictive encoder (MPE) method for music genre classification. The MIDI files are considered as input. MIDI files are converted to the vector sequence by Pitch2vec before being input into the MPE. By unsupervised learning, the MPE based on deep bidirectional transformers is designed to extract bidirectional representations automatically, which are musicological insight. In contrast to other deep-learning models, such as recurrent neural network (RNN)-based models, the MPE method enables parallelization over time-steps, leading to faster training. To evaluate the performance of the proposed method, experiments were conducted on the Lakh MIDI music dataset. During MPE training, approximately 400,000 MIDI segments were utilized for the MPE, for which the recovery accuracy rate reached 97%. In the music genre classification task, the accuracy rate and other indicators of the proposed method were more than 94%. The experimental results indicate that the proposed method improves classification performance compared with state-of-the-art models.

Music Copyright Infringement Online Archive

Music copyright infringement suits often involve little-known musicians seeking compensation from financially successful songwriters and performers, for alleged misappropriation of the intellectual content of musical numbers that were never published, or that enjoyed limited circulation. Judicial opinions typically turn on the court's musical analyses of the works in dispute, but these analyses are not meaningful to the reader of the opinions without audio and visual representation of the musical numbers under scrutiny. These numbers are not, of course, presented along with the court's opinion, and complaining works in particular are generally accessible only from the court's case docket or from the files of the attorneys involved. Unless, then, one somehow already knows the disputed works, one cannot read critically the court's published opinion. To do so, one needs to see the score of the musical work, and/or see and hear a recorded performance of it -- the latter increasingly important as audio engineering and performance style become ever more vital to the commercial success of popular numbers. This can be surprisingly difficult; while everyone has heard of Andrew Lloyd-Webber and can obtain the sheet music and recording of Phantom, who has heard of Ray Repp, and his Till You, that he claims Webber's work plagiarized? Even commercially successful works of popular music have fleeting appeal and brief shelf lives; within a decade or so of publication, the recording and sheet music of once-popular numbers can often be found only in specialized archives, or occasionally, tattered and shopworn, in the collections of municipal libraries.To improve access to this information -- access particularly helpful to legal scholars writing from a historical perspective on music copyright issues -- Columbia Law School's on-line archive will present audio and graphic representations of musical works that are, or have been, the subject of adjudicated music plagiarism cases. Text files summarize and comment upon the opinions in light of the musical works in each case, and contain the full text of the court's opinion. Image files provide representations of pertinent segments of the musical works in question through standard music notation; sound files include both MIDI (Musical Instrument Digital Interface) files with non-stylized renditions of the purely musical content of relevant portions of the disputed numbers, and also streaming audio clips (and video clips when relevant) taken from analog and digital recordings of performances that were intended for public delectation.We believe that this on-line archive will be useful to copyright academics, attorneys practicing in this area, and to musicians seeking insight into an entertaining albeit Cabalistic topic. The project's Discussion and Questions module suggests new means of accessing and analyzing information that pertains to an area of law that now begs inquiry into the issue of how changes in the creation, content, dissemination and consumption of popular music in America in the 20th Century -- which have not, for the most part, been acknowledged by the copyright bar -- should inform music copyright infringement judges and litigants, who tend to approach their work with the paradigms developed for disputes involving Tin Pan Alley numbers in the early decades of this century.

The use of the Synthesia application to simplify Angklung learning

The majority of students in the department of pre-service elementary school teacher cannot read the notation correctly. However, they have a strong motivation to learn music composition, such as angklung music composition. This paper aims to analyze the development of students who are members of the music community in studying the composition of angklung music by using the Synthesia application. The research method was carried out in the following stages: (1) making angklung musical composition in the type of MIDI (Musical Instrument Digital Interface) files; (2) input MIDI files for the Synthesia application; (3) setting up the Synthesia application; (4) learning Angklung using the Synthesia application; and (5) evaluating the learning process. Research subjects were students who were members of the music community on campus, so that their age and musical abilities were very heterogeneous. The research subjects were 15 people, as many as one angklung player group. The results showed that angklung learning for students who had the low notation reading ability, became easier by using a touch of technology. Students can read, visual-based notation correctly. It can be concluded; that technology can provide the convenience in the process of learning music composition, especially in angklung music composition.

LSTM Based Music Generation System

Traditionally, music was treated as an analogue signal and was generated manually. In recent years, music is conspicuous to technology which can generate a suite of music automatically without any human intervention. To accomplish this task, we need to overcome some technical challenges which are discussed descriptively in this paper. A brief introduction about music and its components is provided in the paper along with the citation and analysis of related work accomplished by different authors in this domain. Main objective of this paper is to propose an algorithm which can be used to generate musical notes using Recurrent Neural Networks (RNN), principally Long Short-Term Memory (LSTM) networks. A model is designed to execute this algorithm where data is represented with the help of musical instrument digital interface (MIDI) file format for easier access and better understanding. Preprocessing of data before feeding it into the model, revealing methods to read, process and prepare MIDI files for input are also discussed. The model used in this paper is used to learn the sequences of polyphonic musical notes over a single-layered LSTM network. The model must have the potential to recall past details of a musical sequence and its structure for better learning. Description of layered architecture used in LSTM model and its intertwining connections to develop a neural network is presented in this work. This paper imparts a peek view of distributions of weights and biases in every layer of the model along with a precise representation of losses and accuracy at each step and batches. When the model was thoroughly analyzed, it produced stellar results in composing new melodies.

DRM을 지원하는 악기 연습용 안드로이드 MIDI 재생기 구현

The Musical Instrument Digital Interface (MIDI) file format is a typical format for dynamic music that can be used not only for playing, but also for supporting dynamic music activity, such as musical instrument practice. However, almost all of the existing MIDI applications focus only on the playing features. In this paper, therefore, we implement a MIDI file play application that supports various dynamic features such as part-play repeating certain sections that the user wants, tone and speed control while playing a MIDI file, downloading MIDI files from a server, and lastly Digital Rights Management (DRM). By supporting these features, we are confident that our MIDI application will help users who wish to practice musical instruments to do so conveniently, while effectively protecting the copyright of their purchased MIDI files.

A Novel Query-by-Singing/Humming Method by Estimating Matching Positions Based on Multi-layered Perceptron

The increase in the number of music files in smart phone and MP3 player makes it difficult to find the music files which people want. So, Query-by-Singing/Humming (QbSH) systems have been developed to retrieve music from a user’s humming or singing without having to know detailed information about the title or singer of song. Most previous researches on QbSH have been conducted using musical instrument digital interface (MIDI) files as reference songs. However, the production of MIDI files is a time-consuming process. In addition, more and more music files are newly published with the development of music market. Consequently, the method of using the more common MPEG-1 audio layer 3 (MP3) files for reference songs is considered as an alternative. However, there is little previous research on QbSH with MP3 files because an MP3 file has a different waveform due to background music and multiple (polyphonic) melodies compared to the humming/singing query. To overcome these problems, we propose a new QbSH method using MP3 files on mobile device. This research is novel in four ways. First, this is the first research on QbSH using MP3 files as reference songs. Second, the start and end positions on the MP3 file to be matched are estimated by using multi-layered perceptron (MLP) prior to performing the matching with humming/singing query file. Third, for more accurate results, four MLPs are used, which produce the start and end positions for dynamic time warping (DTW) matching algorithm, and those for chroma-based DTW algorithm, respectively. Fourth, two matching scores by the DTW and chroma-based DTW algorithms are combined by using PRODUCT rule, through which a higher matching accuracy is obtained. Experimental results with AFA MP3 database show that the accuracy (Top 1 accuracy of 98%, with an MRR of 0.989) of the proposed method is much higher than that of other methods. We also showed the effectiveness of the proposed system on consumer mobile device.

Development of a Portable Automated Piano Player CantaPlayer

This paper describes the development of a low cost, compact and portable automated piano player CantaPlayer. The system accepts digital MIDI (Musical Instrument Digital Interface) files as input and develops pushing actions against piano keys which in turn produces sounds of notes. CantaPlayer uses Pure Data, an audio processing software to parse MIDI files and serve as user interfaces. The parsed information will be sent to Arduino, an open source microcontroller platform, via serial communication. The Arduino I/O pins will be triggered based on the information from Pure Data of which connected transistors will be activated, acting as a switch to draw in larger power supply to power the solenoids. The solenoids will then push the respective piano keys and produce music. The performance of CantaPlayer is evaluated by examining the synchronousness of the note playing sequence for a source MIDI and the corresponding reproduced MIDI. Three types of MIDI playing sequence (scale, polyphonic and rapid note switching) were tested and the results were satisfactory.

Pattern Recognition Approach for Music Style Identification Using Shallow Statistical Descriptors

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> In the field of computer music, pattern recognition algorithms are very relevant for music information retrieval applications. One challenging task in this area is the automatic recognition of musical style, having a number of applications like indexing and selecting musical databases. From melodies symbolically represented as digital scores (standard musical instrument digital interface files), a number of melodic, harmonic, and rhythmic statistical descriptors are computed and their classification capability assessed in order to build effective description models. A framework for experimenting in this problem is presented, covering the feature extraction, feature selection, and classification stages, in such a way that new features and new musical styles can be easily incorporated and tested. Different classification methods, like Bayesian classifier, nearest neighbors, and self-organizing maps, are applied. The performance of such algorithms against different description models and parameters is analyzed for two particular musical styles, jazz and classical, used as an initial benchmark for our system. </para>

Method and apparatus for playing a digital music file based on resource availability

A method and corresponding equipment by which a synthesizer/ MIDI (musical instrument digital interface) device (10) is able to optimally perform a MIDI file (11) taking into account not the polyphony required by the MIDI file (11) as in SP-MIDI (scalable polyphony MIDI), but taking into account instead extended scalable polyphony (XSP) data 12b including the maximum number of instantaneous voices required by the MIDI file and the categories in which they occur for different channel masking, and also taking into account the architecture of the synthesizer/ MIDI device (10) in terms of a voice complexity coefficient table (12b) indicating the relative complexity (corresponding to a resource requirement) for voices in each category. The result is a total voice requirement table 12c-1 indicating typically less masking than would be required for the same synthesizer/ MIDI device to play the MIDI file according to SP-MIDI.