All posts by Emmanouil Benetos

On 5-9 November 2023, several AIM researchers will participate at the 24th International Society for Music Information Retrieval Conference (ISMIR 2023). ISMIR is the leading conference in the field of music informatics, and is currently the top-cited publication for Music & Musicology (source: Google Scholar). This year ISMIR will take place in Milan, Italy, and online.

This year, the AI and Music CDT is involved in the organisation of the conference, while also having a strong presence at the tutorials and scientific programme of the conference.

In the Scientific Programme, the following papers are authored/co-authored by AIM students:

• Real-time Percussive Technique Recognition and Embedding Learning for the Acoustic Guitar (Andrea Martelloni, Andrew McPherson, Mathieu Barthet)
• Symbolic Music Representations for Classification Tasks: A Systematic Evaluation (Huan Zhang, Emmanouil Karystinaios, Simon Dixon, Gerhard Widmer, Carlos Eduardo Cancino-Chacón)
• LyricWhiz: Robust Multilingual Lyrics Transcription by Whispering to ChatGPT (Le Zhuo, Ruibin Yuan, Jiahao Pan, Yinghao Ma, Yizhi Li, Ge Zhang, Si Liu, Roger B. Dannenberg, Jie Fu, Chenghua Lin, Emmanouil Benetos, Wenhu Chen, Wei Xue, Yike Guo)
• On the effectiveness of speech self-supervised learning for music (Yinghao Ma, Ruibin Yuan, Yizhi Li, Ge Zhang, Chenghua Lin, Xingran Chen, Anton Ragni, Hanzhi Yin, Emmanouil Benetos, Norbert Gyenge, Ruibo Liu, Gus Xia, Roger B. Dannenberg, Yike Guo, Jie Fu)
• Singing voice synthesis using differentiable LPC and glottal-flow inspired wavetables (Chin-Yun Yu, George Fazekas) – best paper award finalist
• FiloBass: A Dataset and Corpus Based Study of Jazz Basslines (Xavier Riley, Simon Dixon)
• Gender-coded sound: Analysing the gendering of music in toy commercials via multi-task learning (Luca Marinelli, George Fazekas, Charalampos Saitis)
• Modeling Harmonic Similarity for Jazz Using Co-occurrence Vectors and the Membrane Area (Carey Bunks, Simon Dixon, Tillman Weyde, Bruno Di Giorgi)

The following Tutorials will be presented by AIM students:

• Introduction to Differentiable Audio Synthesizer Programming (Ben Hayes, Jordie Shier, Chin-Yun Yu, David Südholt, Rodrigo Diaz)
• Transformer-based Symbolic Music Generation: Fundamentals to Advanced Concepts, Stylistic Considerations, Conditioning Mechanisms and Large Language Models (Berker Banar, Pedro Sarmento, Sara Adkins)
• Kymatio: Deep Learning meets Wavelet Theory for Music Signal Processing (Cyrus Vahidi, Christopher Mitcheltree, Vincent Lostanlen)

On Satellite Events, AIM PhD student Elona Shatri is general co-chair for the 5th International Workshop on Reading Music Systems (WoRMS 2023). Additionally, she will showcase two papers she collaborated on with master’s students: Towards Artificially Generated Handwritten Sheet Music Datasets (Pranjali Hande, Elona Shatri, Benjamin Timms, and George Fazekas) Improving Sheet Music Recognition using Data Augmentation and Image Enhancement (Zihui Zhang, Elona Shatri, and George Fazekas)

Finally, AIM PhD student Ilaria Manco is Sponsorship/Industry co-chair for ISMIR 2023.

See you at ISMIR!

Congratulations to two AIM PhD students, Andrea Martelloni and Max Graf, who have both won MIDI Innovation Awards 2023 in the hardware prototype and software prototype categories, respectively.

Two AIM PhD students have each won an award at this year’s MIDI Innovation Awards. Now in its 40th year, the MIDI Awards showcase products and projects that are using MIDI (Musical Instrument Digital Interface) in fresh and original ways, highlighting the role that MIDI technology has to play in enabling musical creativity.

Andrea Martelloni won the hardware prototype category for the HITar, an augmented percussive guitar which Andrea developed under the supervision of Dr Mathieu Barthet and Professor Andrew McPherson. Aimed at percussive fingerstyle guitarists, the HITar is a device that can be fitted to a regular acoustic guitar and can alter the way a player interacts with the instrument’s body. The unit employs sensors placed underneath the areas most commonly struck by players and uses an AI engine to determine which part of the hard is used for each percussive hit. The resulting MIDI output can then be used to integrate hardware or software instruments or samples into a performance, allowing guitarists to trigger drum and percussion samples or blend sample libraries and virtual instruments with their playing.

In the software category, Max Graf won for Netz, a mixed reality (MR) software instrument that blends the real and virtual worlds by displaying 3D virtual objects within a real environment. It’s a self-contained instrument that features an embedded sound engine, allowing users to produce sounds directly from the head-mounted display. The software’s interface appears as a network where nodes represent notes, which can be mapped to a tangible surface for tactile feedback or be positioned in the air. Performers’ hand poses and gestures are tracked in in real time, allowing Netz to translate subtle hand movements to expressive musical controls; gestures such as the opening and closing of fingers — or movements such as wrist rotation — are recognised and interpreted by the system and can be assigned to specific instrument parameters and MIDI controls.

Many congratulations from everyone in AIM to Andrea, Max and their supervisors for this huge achievement!

More information:

• MIDI Innovation Awards Sound on Sound – MIDI Innovation Awards 2023: Winners announced
• HITar Twitter/X and Instagram page

Two AIM PhD students have been given enrichment awards by the Alan Turing Institute, the UK’s national institute in artificial intelligence and data science, enabling them to join and interact with institute researchers and its community in the 2023/24 academic year.

Specifically, AIM PhD student Teresa Pelinski has been offered an Enrichment Placement Award for the project “Sensor mesh as performance interface” and AIM PhD student Shubhr Singh has been offered an Enrichment Placement Award for the project “Deep learning with graphs for multivariate time series data”.

Congratulations to both!