Low-Latency Audio Pitch Tracking: A Multi-Modal Sensor-Assisted Approach

Laurel Pardue, Dongjuan Nian, Christopher Harte, and Andrew McPherson

Proceedings of the International Conference on New Interfaces for Musical Expression

• Year: 2014
• Location: London, United Kingdom
• Pages: 54–59
• DOI: 10.5281/zenodo.1178899 (Link to paper)
• PDF link

Abstract:

This paper presents a multi-modal approach to musical instrument pitch tracking combining audio and position sensor data. Finger location on a violin fingerboard is measured using resistive sensors, allowing rapid detection of approximate pitch. The initial pitch estimate is then used to restrict the search space of an audio pitch tracking algorithm. Most audio-only pitch tracking algorithms face a fundamental tradeoff between accuracy and latency, with longer analysis windows producing better pitch estimates at the cost of noticeable lag in a live performance environment. Conversely, sensor-only strategies struggle to achieve the fine pitch accuracy a human listener would expect. By combining the two approaches, high accuracy and low latency can be simultaneously achieved.

Citation:

BibTeX Entry:

  @inproceedings{lpardue2014,
 abstract = {This paper presents a multi-modal approach to musical instrument pitch tracking combining audio and position sensor data. Finger location on a violin fingerboard is measured using resistive sensors, allowing rapid detection of approximate pitch. The initial pitch estimate is then used to restrict the search space of an audio pitch tracking algorithm. Most audio-only pitch tracking algorithms face a fundamental tradeoff between accuracy and latency, with longer analysis windows producing better pitch estimates at the cost of noticeable lag in a live performance environment. Conversely, sensor-only strategies struggle to achieve the fine pitch accuracy a human listener would expect. By combining the two approaches, high accuracy and low latency can be simultaneously achieved.},
 address = {London, United Kingdom},
 author = {Laurel Pardue and Dongjuan Nian and Christopher Harte and Andrew McPherson},
 booktitle = {Proceedings of the International Conference on New Interfaces for Musical Expression},
 doi = {10.5281/zenodo.1178899},
 issn = {2220-4806},
 month = {June},
 pages = {54--59},
 publisher = {Goldsmiths, University of London},
 title = {Low-Latency Audio Pitch Tracking: A Multi-Modal Sensor-Assisted Approach},
 url = {http://www.nime.org/proceedings/2014/nime2014_336.pdf},
 year = {2014}
}