A5052209972- Music and Audio Processing
- Music Technology and Sound Studies
- Speech and Audio Processing
- Animal Vocal Communication and Behavior
- Diverse Musicological Studies
- Geochemistry and Geologic Mapping
- Musicology and Musical Analysis
- Speech Recognition and Synthesis
- Computational Physics and Python Applications
International Audio Laboratories Erlangen
2020-2022
The analysis of recorded audio material using computational methods has received increased attention in ethnomusicological research. We present a curated dataset traditional Georgian vocal music for musicology. corpus is based on historic tape recordings three-voice songs performed by the former master chanter Artem Erkomaishvili. In this article, we give detailed overview material, transcriptions, and annotations contained dataset. Beyond its importance research, carefully organized...
Choral singing is a central part of musical cultures across the world, yet many facets this widespread form polyphonic are still to be explored. Music information retrieval (MIR) research on choral benefits from multitrack recordings individual voices. However, there exist only few publicly available datasets singing. In paper, we present Dagstuhl ChoirSet (DCS), dataset cappella music designed support MIR The includes an amateur vocal ensemble performing two choir pieces in full and quartet...
In this study we examine the tonal organization of a series recordings liturgical chants, sung in 1966 by Georgian master singer Artem Erkomaishvili. This dataset is oldest corpus chants from which time synchronous F0-trajectories for all three voices have been reliably determined (Muller et al. 2017). It therefore outstanding importance understanding tuning principles traditional vocal music. The aim present to use various computational methods analyze what these can contribute ongoing...
Due to the rapid developments in machine learning and growing importance of opensource software, Python has become predominant computer programming language for research education many scientific fields.While engineering students on Master's level have skills different languages such as MATLAB, C/C++, or Java, they are often less experienced using associated software frameworks.The PCP notebooks contribute closing this gap by offering open-source educational material a Preparation Course...
In this study we examine the tonal organization of 2016 GVM dataset, a newly-created corpus high-quality multimedia field recordings traditional Georgian singing with focus on Svaneti. For purpose, developed new processing pipeline for computational analysis non-western polyphonic music which was subsequently applied to complete dataset. To evaluate under what conditions single tuning system is representative current Svan performance practice, examined stability obtained systems from an...
Over the last decades, various conceptually different approaches for fundamental frequency (F0) estimation in monophonic audio recordings have been developed. The algorithms' performances vary depending on acoustical and musical properties of input signal. A common strategy to assess reliability (correctness) an estimated F0-trajectory is evaluate against annotated reference. However, such annotations may not be available a particular collection are typically labor-intensive generate. In...
A central challenge for a cappella singers is to adjust their intonation and stay in tune relative fellow singers. During editing of recordings, one may want local individual or account global drifts over time. This requires applying time-varying pitch-shift the audio recording, which we refer as adaptive pitch-shifting. In this context, existing (semi-)automatic approaches are either labor-intensive face technical musical limitations. work, present automatic methods tools pitch-shifting...
Larynx microphones (LMs) make it possible to obtain practically crosstalk-free recordings of the human voice by picking up vibrations directly from throat. This can be useful in a multitude music information retrieval scenarios related singing, e.g., analysis individual voices recorded environments with lots interfering noise. However, LMs have limited frequency range and barely capture effects vocal tract, which makes signal unsuitable for downstream tasks that require high-quality...