<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
        <title>Transactions of the International Society for Music Information Retrieval Feed</title>
        <link>https://reference-global.com/journal/UPJTISMIR</link>
        <description>Reference Global RSS Feed for Transactions of the International Society for Music Information Retrieval</description>
        <lastBuildDate>Fri, 23 Oct 2026 12:38:03 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>https://github.com/jpmonette/feed</generator>
        <image>
            <title>Transactions of the International Society for Music Information Retrieval Feed</title>
            <url>https://sciendo-parsed.s3.eu-central-1.amazonaws.com/ubiquity/68406d19da18af458e7684ee/cover-image.jpg</url>
            <link>https://reference-global.com/journal/UPJTISMIR</link>
        </image>
        <copyright>All rights reserved 2026, Ubiquity Press</copyright>
        <atom:link href="https://reference-global.com/2/rss/journal/UPJTISMIR" rel="self" type="application/rss+xml"/>
        <item>
            <title><![CDATA[MuVGM: The Multimodal Video Game Music Dataset]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.381</link>
            <guid isPermaLink="false">/article/10.5334/tismir.381</guid>
            <pubDate>Fri, 23 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[We introduce MuVGM, a large‑scale Multimodal dataset for Video Game Music research, combining symbolic transcriptions, reference audio, gameplay videos, box art imagery, textual descriptions, audio captions, and affective annotations across a broad range of titles and platforms. Video game music is uniquely suited to multimodal research because it is composed for specific, directly observable visual environments and narrative contexts, providing explicit functional grounding absent from most existing music datasets. The dataset comprises 7129 tracks from 1484 games across 39 consoles, spanning 37 years of game music history, and was constructed through a six‑stage pipeline combining web scraping, Wikipedia matching, cover song detection, audio fingerprinting, retrieval‑augmented text generation, and automatic music captioning. The dataset, including an interactive web‑based dataset explorer, is available at https://tylerhmc.com/muvgm/.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Stravinsky in Urbana: The Case for the Evolution of Digital Library Software Design to Support Musicological Scholarship]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.320</link>
            <guid isPermaLink="false">/article/10.5334/tismir.320</guid>
            <pubDate>Wed, 14 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[This article presents a case study that explores how Digital Library (DL) software can evolve to better support musicological scholarship. Drawing from a diverse set of ephemeral source materials—concert programs, oral histories, audio recordings, and correspondence—the project assembles an event‑focused digital archive documenting the musical presence of Igor Stravinsky in Urbana‑Champaign, Illinois. We describe how the Greenstone3 DL platform was used and extended to support the workflow of musicologists investigating this data‑rich collection. The article details both the technical implementation and the scholarly rationale behind these enhancements, presenting a model for how DLs can better serve research grounded in temporally structured, multimodal, and locally situated materials. From a technical standpoint, the pattern of use of the DL software is an iterative one, whereby a ‘rough‑cut’ cloud‑shared drive, developed by the scholars to store digitized versions of located documents, is transformed into a metadata‑enriched and increasingly refined DL collection. From a scholarly perspective, a key insight has been that, through this process, a musicologist obtains a deeper understanding of the material before them and so crafts new metadata structures that reflect their understanding; these in turn further shape how the DL is formed. Our findings highlight that more can be done in terms of the gearing of the software architectures provided by DL systems to better support the scholarly process, enabling more exploratory and inclusive forms of research and presentation, particularly in fields like musicology where diverse media and temporal narratives are central.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[The MPB Corpus: A Dataset of Melody, Rhythm, Harmony, and Melody&amp;#x2013;Harmony Relationships in Brazilian Popular Music]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.357</link>
            <guid isPermaLink="false">/article/10.5334/tismir.357</guid>
            <pubDate>Wed, 14 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[This paper presents the MPB Corpus, a dataset of Brazilian Popular Music (Música Popular Brasileira [MPB]) comprising derived symbolic annotations for 500 pieces across four musical parameters—melodic contour, melodic rhythm, harmony, and the melody–harmony relationship—making it the most comprehensive dataset to date for computational musicology of Brazilian music. It is released under a CC‑BY‑4.0 license in an open, machine‑readable tabular format, with documentation and reproducible analysis code. Because the annotations were derived from copyrighted musical scores, the corpus contains no audio recordings or score‑level representations such as MIDI or MusicXML; it nonetheless provides a reusable resource for MIR research on a culturally significant repertoire that remains underrepresented in existing datasets. To support the encoding process, we introduce analytical models for capturing rhythmic and melodic information, along with tailored visualizations and summary metrics. Finally, we present a brief exploratory analysis illustrating the dataset’s potential for formulating and addressing musicological questions about the genre.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[The Music Perception Toolbox: Analytical Methods for Pitch and Rhythm Similarity, Consonance, Complexity, and Structure]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.392</link>
            <guid isPermaLink="false">/article/10.5334/tismir.392</guid>
            <pubDate>Tue, 06 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[The Music Perception Toolbox is an open‑source package—available in both MATLAB and Python—for computing perceptually and cognitively motivated measures of pitch similarity, consonance, and scale and rhythmic structure. Inputs may be symbolic pitches or spectral peaks extracted from audio. The toolbox covers three families of measure: (i) similarity and complexity measures based on expectation tensors—continuous densities that embed weighted collections of pitches or time points into a unified framework applicable to both pitch and rhythm; (ii) consonance measures including spectral entropy, template harmonicity, tensor harmonicity, and sensory roughness; and (iii) structural features for scales and rhythms, including Fourier‑based balance and evenness, coherence, edge detection, projected centroid, mean offset, and Markov prediction. All three families operate on the same weighted multiset representation of the musical input. These measures have been validated as predictors of perceived tonal fit, consonance, affect, and rhythmic complexity across diverse empirical studies, including experiments with microtonal tunings unfamiliar to the listeners tested, and cross‑cultural fieldwork. Version 2.0 introduces a fully analytical framework for constructing, evaluating, and comparing expectation tensor densities, replacing the grid‑based discretization of earlier work with exact and substantially faster computation. This article presents the mathematical foundations—including the first publication of the analytical inner product underlying the expectation tensor cosine similarity (derived in the Supplementary Material)—describes the toolbox’s design, works through an application characterizing chordal probes by both their tonal fit and their intrinsic consonance, and surveys its prior empirical applications.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[AI and Music at a Crossroads: The Role of the Music Information Retrieval (MIR) Community in Shaping Trustworthy, Open, and Culturally Inclusive Music‑AI]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.372</link>
            <guid isPermaLink="false">/article/10.5334/tismir.372</guid>
            <pubDate>Fri, 25 Sep 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Artificial intelligence (AI) has become central to debates about the future of music, particularly with the rise of generative systems and their societal, economic, cultural, and legal implications. Public and policy discourse, however, often follows commercial narratives that narrow music‑AI to automated generation and treat music primarily as data or commodified output. This obscures both the broader range of AI applications in music and the long‑standing contributions and responsibilities of the music information retrieval (MIR) community. Addressed primarily to this community, the article argues that MIR is well positioned to contribute to current debates on trustworthy music‑AI but must also critically examine the limitations of its own research practices and incentives. We briefly situate MIR’s methodological development, applications, and shared infrastructures, emphasizing the culturally and socially situated character of music data. We then examine ethical, legal, and governance challenges in music‑AI and propose transparency, accountability, provenance, and sustainability as criteria for assessing current practice and guiding future work. Finally, we identify six interconnected priorities: mission‑oriented collaboration; sustainable open infrastructures; rigorous and context‑sensitive evaluation; cultural diversity and responsibility; engagement across research, industry, policy, and society; and interdisciplinary education and mutual capacity‑building. We conclude by identifying ways in which the MIR community can translate these priorities into practice through its research, infrastructures, evaluation and publication processes, educational activities, and engagement with musical, industry, policy, and civil society communities.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Data‑Driven Analysis of Musical Form and Harmonic Structure in AI‑Generated Popular Music: A Case Study with Suno and Udio]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.348</link>
            <guid isPermaLink="false">/article/10.5334/tismir.348</guid>
            <pubDate>Thu, 10 Sep 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[We present a large‑scale analysis of the harmonic and formal structures of music generated by users of two AI music‑generation platforms, Suno and Udio, compared with human‑composed popular music. We address three questions: (1) What considerations arise when applying music information retrieval tools to AI‑generated music? (2) How does AI‑generated music’s harmony relate to human‑composed popular music? (3) What structural similarities or differences exist? We analyze 20,002 Suno tracks, 20,002 Udio tracks, and 20,010 human‑composed recordings, performing harmonic analysis with a state‑of‑the‑art chord‑estimation model and structural analysis with a joint beat‑tracking and semantic section‑labeling model. We model harmonic progression and musical form using trigram and tetragram analyses of detected chords and labeled sections, quantifying their diversity across the three collections. Music in our Suno collection concentrates more strongly on Western pop harmonic patterns, such as I‑V‑vi‑IV, than music in either our Udio or human collection, and tends toward simple verse‑chorus repetitions, whereas Udio exhibits wider structural variety. These findings suggest that music generated by Suno (up to version 3.5) sits in a narrower harmonic and formal space than that of Udio and human‑composed music. Fully automated, open‑source music information retrieval tools now support corpus‑scale harmonic and structural analysis, although chord estimation remains an open challenge, particularly for inversions, extensions, and substituted dominants. We release our beat‑level chord annotations, metadata, section boundaries, self‑similarity matrices, audio links, and code under an MIT license.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Rhythm Guitar Tablature Continuation Through Picking Pattern Generation]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.368</link>
            <guid isPermaLink="false">/article/10.5334/tismir.368</guid>
            <pubDate>Mon, 10 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Rhythm guitar tablatures in Western popular music often rely on repetitive patterns that follow a chord progression and usually preserve textural and rhythmic properties. This paper introduces a novel approach for suggesting rhythm guitar tablature continuations, conditioned on chord sequences and textural metrics. To benefit from a modular approach, we propose to split rhythm guitar tablature continuation into two steps: choosing chord positions and generating picking patterns. This paper focuses on the latter step, for which we propose two models: 1) a rule‑based model mimicking copy‑and‑paste practices in tablature notation software and 2) a Transformer‑based model that can generate more varied continuations using conditioning signals like textural variations and structural information. We train and test our models on the rhythm guitar tracks from the DadaGP tablatures dataset. The performance of the models is evaluated through several quantitative metrics as well as an online user study. While the Transformer model demonstrates better quantitative results, the rule‑based model is preferred in subjective ratings for its consistency and usability. Consequently, both models show promise: the rule‑based model for assisting tablature notation, and the Transformer model for providing creative suggestions during composition.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[KRAISLER: A Multi‑Track Dataset of Piano and Violin Duet Recordings for Music Information Retrieval Research]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.338</link>
            <guid isPermaLink="false">/article/10.5334/tismir.338</guid>
            <pubDate>Thu, 06 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[The piano plays a central role in classical music, both as a solo instrument and in ensemble settings, often accompanying melodic instruments such as the violin, cello, and flute. Recordings of piano performances by human players in these musical arrangements have been valuable resources for music information retrieval (MIR) research, supporting tasks such as automatic music transcription, performance analysis, and automatic music accompaniment. However, most existing datasets focus on solo piano performances or provide only mixed audio in ensemble contexts, limiting the usage for various scenarios of music analysis. We present KRAISLER, a new multi‑track dataset of piano and violin duets featuring high‑quality audio from simultaneous recording of live studio ensemble performances by professional musicians in acoustically isolated rooms. The dataset comprises 20 excerpts from pieces spanning a wide range of historical periods and classical genres. Each excerpt is accompanied by directly captured piano MIDI data, musical scores with corresponding score‑alignment data, note annotations, and beat annotations. We evaluate the dataset across multiple MIR tasks using state‑of‑the‑art models, including piano or violin transcription, piano–violin source separation, score alignment, and beat tracking, and establish it as a benchmark. Beyond the tasks applied in the benchmark, we expect the dataset to be well‑suited for various expressive performance and ensemble research topics.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Beyond Ba(t)ch: A Multimodal Meta‑Corpus of Digital Chorale Transcriptions and Related Data]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.226</link>
            <guid isPermaLink="false">/article/10.5334/tismir.226</guid>
            <pubDate>Tue, 04 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Chorales (broadly speaking, hymns) have featured prominently in the musical life of Western cultures for hundreds of years, particularly in sacred, liturgical settings. They have also featured just as prominently in the theoretical and practical training of musicians and also (more recently) as a focus of attention in computational approaches to music. One strange quirk of that history is the dominance of J.S. Bach’s chorales: for all the many fine qualities of ‘the Bach Chorales’, that collection is neither representative of wider practice, nor (arguably) the most convenient vehicle for pedagogy, particularly at a basic level. This paper reports on a multi‑institutional effort to bring together a wider collection of chorales for musicians and research. This new meta‑collection is still centred on Germany and the 18th century, but features more collections from that regional‑temporal centre and also expands it. The dataset is openly available at https://github.com/Chorale‑Corpus, permissively licenced and indexed at https://doi.org/10.5281/zenodo.17229783.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Attention‑Enhanced Convolutional Neural Network for Music Genre Classification]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.268</link>
            <guid isPermaLink="false">/article/10.5334/tismir.268</guid>
            <pubDate>Tue, 04 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[The complex interplay of harmonic, rhythmic, and timbral elements within musical signals poses a challenge for accurate music genre classification. Conventional approaches have limitations in effectively capturing both localized features and long‑range contextual relationships within the audio data. To overcome this, we introduce an attention‑enhanced convolutional neural network (AECNN) architecture that employs convolutional layers to efficiently extract localized features and an attention mechanism to capture long‑range temporal dependencies within the input audio segment. This integrated approach empowers AECNN to comprehensively learn both local features and segment‑level contextual relationships, resulting in enhanced accuracy and robustness in classifying diverse music genres. The performance of the proposed model is evaluated on two datasets, the widely used GTZAN dataset and the novel IMG‑Amrita dataset, encompassing 12 distinct Indian music genres. Experiments conducted with varying audio sample durations (30, 15, 10, 6, and 3 s) demonstrate the impact of sample length on classification accuracy. Notably, AECNN achieves peak accuracies of 96.02% on the GTZAN dataset and 98.94% on the IMG‑Amrita dataset with 3‑s samples, underscoring its ability to capture intricate musical nuances with limited audio data. These findings highlight the significant potential of the AECNN for robust and versatile performance across a wide range of music classification tasks.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Beethoven Symphony Excerpt Dataset (BSED): An Evaluation Dataset for Orchestral Music Transcription]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.343</link>
            <guid isPermaLink="false">/article/10.5334/tismir.343</guid>
            <pubDate>Fri, 24 Jul 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Orchestral music poses significant challenges for automatic music transcription (AMT) due to its dense polyphony, diverse instrument timbres, and varied acoustic conditions. While AMT research for solo instruments and chamber music has advanced considerably, orchestral transcription remains underexplored, primarily because of the lack of high‑quality, time‑aligned datasets. In this work, we make three main contributions. First, we introduce the Beethoven Symphony Excerpt Dataset (BSED), a carefully curated, publicly available benchmark for evaluating orchestral AMT systems. BSED contains 20 short excerpts from Beethoven’s nine symphonies, each provided in five audio versions: four distinct concert recordings and one synthetic rendition, with an approximate total duration of 37 min. For each excerpt, we supply symbolic scores in multiple formats (PDF, MusicXML, Sibelius, MIDI, CSV) alongside high‑quality note‑level annotations obtained through robust, manually verified score–audio alignment, refined with transcription‑based onset features. Second, to facilitate model development, we release the Beethoven Symphony Dataset (BSD), a large‑scale orchestral training set comprising 62 h of public‑domain recordings with time‑aligned annotations derived from digital scores and structurally verified. While less controlled than BSED, BSD offers rich diversity across performances, conductors, and acoustic environments. Third, we establish a baseline for orchestral AMT by training an instrument‑agnostic note‑level transcription model on BSD and evaluating it on both BSED and the independent PHENICX dataset. Together, BSED and BSD address a critical gap in orchestral music information retrieval resources and provide a foundation for advancing AMT research toward more robust and generalizable systems.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Time to Align! Modelling Musical Timelines for Music Information Retrieval and Digital Musicology]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.296</link>
            <guid isPermaLink="false">/article/10.5334/tismir.296</guid>
            <pubDate>Thu, 23 Jul 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Aligning musical data – matching corresponding events across different representations of the same music – is central to Music Information Retrieval and Digital Musicology. Datasets span graphical representations (images, scans), logical encodings (scores, MIDI) and physical recordings (audio, video, sensor data), each imposing its own coordinate system for musical time. Integrating them requires costly alignment, but results are typically stored in ad hoc formats that are poorly documented and difficult to share and reuse. To address this, we present TimeToAlign!, a cross‑domain model for musical timelines and their alignments that organises time‑related data into continuous and discrete variants across three temporal domains – graphical, logical and physical – connected by typed conversion maps and explicit match claims. In addition to describing the model, we demonstrate it through five concrete use cases of increasing complexity: (1) aligning a historical piano roll with a digital score, (2) modelling flow control in a school song arrangement, (3) transferring annotations between graphical analyses, (4) propagating harmonic labels across hundreds of multimodal timelines of a Beethoven string quartet and (5) encoding the conceptual and temporal relationships in the genesis of a rock song. The model is implemented as timetoalign, an open‑source Python library with typed timelines, loaders for common formats (such as audio formats, MusicXML or CSV) and conversion maps. Runnable tutorial and how‑to notebooks, a documentation homepage and a collection of real‑world data from diverse repertoires accompany the library.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Art Official Intelligence: Music and Intellectual Property Philosophy in the New Age of Artificial Intelligence–Assisted Creativity]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.339</link>
            <guid isPermaLink="false">/article/10.5334/tismir.339</guid>
            <pubDate>Wed, 22 Jul 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[New forms of technological innovation often come with new concerns about impacts on artistry and creativity, including disruptions to markets, effects on rights relating to creative works, and artists’ changing relationships to their identities and creations. While previous work has begun to explore creatives’ experiences with the transformative technology of artificial intelligence (AI), further understanding is needed, particularly regarding how these tools could be designed to support individuals’ motivations to create and strengthen their connections to the AI‑assisted works they make. The philosophical traditions backing intellectual property law offer valuable insight into how to conceptualize the repercussions of AI on artists engaging in music creation. In this paper, we look into the philosophical foundations of intellectual property, specifically focusing on three primary approaches (incentives‑based utilitarian, Lockean, and personality‑based), and explicate how the consequences of AI tools on creativity can be understood from each approach—in particular, what may be viewed as significant challenges for future creation and innovation. We then provide design suggestions that could help mitigate those challenges in music AI tools.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Cross-Cultural Music Similarity: Bridging Human Perception, Signal Processing, and Foundation Models]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.341</link>
            <guid isPermaLink="false">/article/10.5334/tismir.341</guid>
            <pubDate>Tue, 21 Jul 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Understanding similarity across cultural boundaries remains a fundamental challenge in social sciences and artificial intelligence, with music serving as an ideal paradigm for cross‑cultural perception research. This work presents the first comprehensive evaluation of computational music similarity methods against human cross‑cultural music perception, spanning nine diverse musical traditions and encompassing both interpretable signal‑processing features and state‑of‑the‑art foundation models. We collected human annotations from 125 participants across diverse backgrounds, evaluating 1,130 unique audio pairs from Western, Mediterranean, Middle Eastern, Indian, and Chinese musical cultures. Each pair was assessed along three dimensions: overall musical similarity, cultural similarity, and recommendation‑level similarity. We systematically compare these human judgments against computational approaches, including signal‑processing features (rhythm, melody, harmony, timbre) and seven foundation models (MERT, CultureMERT, CLAP, Qwen2‑Audio variants). Results demonstrate that foundation models generally outperform signal‑processing features, with CLAP‑Music&amp;Speech achieving the highest alignment with human perception. Among signal‑processing features, melody consistently emerges as the most predictive of human similarity judgments. Cross‑cultural discrimination analysis reveals that, while humans show strong cultural awareness, computational methods achieve more modest discrimination, with Qwen2‑Audio demonstrating the best cultural boundary detection. Most significantly, ensemble methods combining both approaches achieve substantial improvements, demonstrating the complementary value of interpretable features and learned representations. This study establishes a comprehensive evaluation framework for cross‑cultural music similarity and provides crucial insights for developing culturally aware music technology systems.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Revisiting Expressive Timing in Piano Performance at Scale: A Distant Listening, Multi‑Corpus Study]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.317</link>
            <guid isPermaLink="false">/article/10.5334/tismir.317</guid>
            <pubDate>Fri, 17 Jul 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Empirical expressive performance studies are typically performed on small, curated datasets. In this study, we present a large‑scale analysis of expressive timing in classical piano performance using a ‘distant listening’ approach across six MIDI datasets totaling over 600 hours of music. We revisit three core aspects of expressive timing drawn from empirical performance research: (1) the relational invariance of timing patterns with respect to tempo; (2) the influence of score markings, register, and interval consonance on articulation; and (3) the role of asynchrony as a means of expressive accentuation. Our findings indicate that timing patterns do not scale proportionally with tempo; however, weak interactions between tempo and metrical position suggest that both factors jointly contribute to shaping timing patterns. Articulation analysis confirms associations with score markings and register, although the relationship with interval consonance is less consistent. Melody lead tendencies are consistently observed across datasets, though our findings indicate that those occurring on the beat level may primarily reflect dynamic accentuation. Overall, our results provide empirical validation for certain established theories of expressive performance while challenging others, illustrating how large‑scale computational analysis can complement traditional controlled studies by offering broader insights into performance timing.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Distorted Realities: Classifying Extreme Vocals Between Harmony and Noise – A Machine–Human Evaluation of Vocal Confusion Patterns]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.310</link>
            <guid isPermaLink="false">/article/10.5334/tismir.310</guid>
            <pubDate>Thu, 16 Jul 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Extreme vocal techniques like growling and shrieking are common in metal and related genres, giving the music a characteristic intensity. Although widely used, these techniques have received limited attention in Music Information Retrieval (MIR) and musicology, particularly in fine‑grained categorisation. Their niche status and marginalisation have led to limited representation in digital music libraries. Since current classification systems prioritise dominant subgenres due to a lack of detailed vocal categorisations, extreme vocal styles are often oversimplified or misrepresented, hindering tagging, retrieval, and analysis. Automation could enable consistent and nuanced vocal annotation, supporting the technical goals of MIR and analytical aims of musicology. Foregrounding vocal technique as a classification axis contributes to a richer understanding of vocal performance and genre formation. To assess machine learning as a reliable alternative to human judgment, this study developed a coarse frequency‑based qualitative taxonomy from interviews with 11 extreme vocal practitioners and compared human and machine performance using the Extreme Metal Vocals Dataset (EMVD), including a listening test with 158 expert participants. Perceptual results achieved 76.2% Unweighted Average Recall (UAR), while a Support Vector Machine (SVM) model trained on ComParE features achieved a UAR of 90% on average, using singer‑independent 3‑fold group cross‑validation. Feature group, ablation, and human prediction analyses suggest that the human‑machine gap reflects differences in cue prioritisation and classification performance. Results illustrate the feasibility of machine learning to annotate vocal styles, while showing that human confusion patterns remain informative for category boundaries. This highlights how MIR methods could facilitate analysis of underrepresented genres.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Semi‑Automatic Pipeline for the Transcription of Mensural Polyphony into Symbolic Interpreted Scores]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.292</link>
            <guid isPermaLink="false">/article/10.5334/tismir.292</guid>
            <pubDate>Wed, 15 Jul 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[This paper presents the first semi-automatic pipeline for transcribing early polyphonic music, written in separate parts, into symbolic scores. While current optical music recognition (OMR) software can recognize and encode musical symbols, it falls short of producing complete transcriptions from these historical sources. Polyphonic music from the late Middle Ages and Renaissance is written in mensural notation, where note duration cannot be inferred from graphical symbols alone. Accurate interpretation requires consideration of additional features, such as mensuration signs and the contextual placement of notes. This article introduces an integrated MIR pipeline that transcribes digital images of mensural sources into aligned symbolic scores by combining OMR, a mensural rhythmic interpretation expert system, and error-checking technologies—thereby reducing human intervention. A central aim of this work is to improve interoperability among existing specialized tools rather than develop new monolithic solutions and to rely on free, open-access software to enable adoption by institutions and scholars worldwide, including those in developing countries. The pipeline was successfully applied to the Guatemalan choirbook GCA-Gaha 1, demonstrating its practical utility and effectiveness.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Diplomatic Transcriptions Between Document and Text: Encoding Beethoven’s Sketchbooks]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.318</link>
            <guid isPermaLink="false">/article/10.5334/tismir.318</guid>
            <pubDate>Wed, 15 Jul 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Sketchbooks are a special type of music document. Typically, they are not intended for readers other than the composer, and the private notation used often does not follow regular rules of music notation. This means that the traditional categories of (music) philology, document and text, have a different relation in the context of sketchbooks. For printed editions, this led to the introduction of diplomatic editions, which try to increase accessibility of sketchbooks. This article introduces a Music Encoding Initiative–based data model for digital editions, which enables a broader view on sketchbooks and other complex music documents. It integrates document‑ and text‑centric perspectives and facilitates a holistic approach that eventually enables users of a digital edition to ‘liquify’ the formerly static editions. This innovative concept of fluid transcriptions will be illustrated using Notirungsbuch K, a sketchbook that Ludwig van Beethoven used in 1823. Besides documenting the concepts and data model necessary for this, the article will also introduce the Facsimile Explorer, a web application that is used to prepare data conformant to this model.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[An Audio-Perspective on the Divergent Paths of Techno in Germany and the United States]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.324</link>
            <guid isPermaLink="false">/article/10.5334/tismir.324</guid>
            <pubDate>Thu, 09 Jul 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Many documentaries on early house and techno music exist. Here, protagonists from the scenes describe key elements and events that affected the evolution of the music. In the research community, there is consensus that such descriptions have to be examined critically. Yet, there have not been attempts to validate such statements on the basis of audio analyses. In this study, over 9,000 early house and techno tracks from Germany and the United States of America are analyzed using recording studio features, machine learning, and inferential statistics. Three observations can be made: (1) German and US house/techno music are distinct; (2) US styles are much more alike; and (3) with respect to the recording studio features, US styles scarcely evolved over time compared to German house/techno. These findings are in agreement with documented statements and thus provide an audio-based perspective on why techno became a mass phenomenon in Germany but remained a fringe phenomenon in the United States. Observations like these can help the music industry estimate whether new trends will experience a breakthrough or disappear.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Publishing and Performing Digital Musicology on the Web: Signature Sound Vienna]]></title>
            <link>https://reference-global.com/article/10.5334/tismir.315</link>
            <guid isPermaLink="false">/article/10.5334/tismir.315</guid>
            <pubDate>Wed, 08 Jul 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Signature Sound Vienna is an interdisciplinary digital musicology investigation of the Vienna Philharmonic Orchestra’s New Year’s Concert series, conducted by reference to a collection of relevant concert recordings and scores assembled for this purpose. In this paper, we present an overview of the workflows and software tools developed co‑creatively between the project’s music scholars and technologists, aiming to facilitate the publication, analysis and dissemination of findings pertaining to musicological research data while adhering to FAIR principles of data management. These workflows and tools include automated conversion scripts to publish metadata of large collections of music media, retrieved and validated using the MusicBrainz Picard Tagger, as Linked Data adhering to the Music Ontology; mei‑friend, a browser‑based editor for music encodings; Listen Here!, an environment for tool‑assisted close listening to score‑aligned collections of performance recordings; and Primal, a platform for review and interaction with music annotation Linked Data. Building on Web standards and established ontologies, including the Solid platform for social Linked Data, the Web Annotation Data Model, the Music Annotation Ontology and the schema of the Music Encoding Initiative (MEI), our modular tools integrate on shared Linked Data structures. This supports complex musicological investigations through comparatively simple applications focusing on specific tasks and modalities, as well as enabling further extensions beyond the scope of our project. We present a case study of the use of this approach within our own scholarly investigations, showcasing findings from performance analyses of collections of the most frequently performed compositions of the New Year’s Concert series, and discuss the implications of our approach on broader digital musicology contexts.]]></description>
            <category>ARTICLE</category>
        </item>
    </channel>
</rss>