A Method for Studying Interactions between Music Performance and Rooms with Real-Time Virtual Acoustics

Elliot K. Canfield-Dafilou, Eoin F. Callery,
Jonathan S. Abel and Jonathan J. Berger

Center for Computer Research in Music and Acoustics, Stanford University

2019 AES Convention, Dublin Ireland

music in reverberant environments

  • Architectural spaces shape how music is written, performed, and heard
  • Composers likely had the acoustics of performance venues in mind when writing music
  • Musicians interact with and adapt to the acoustics of the performance venue
  • Some genres of music sound better in certain reverberant environments


We would like to study interactions between groups of performers and acoustics, and try to understand why certain acoustic spaces are well suited for certain genres of music

Prior Work

  • Gade (1989)
subjective role of orchestral platform acoustics
  • Ueno et al (2003)
role of early reflections and relative level of own instrument
  • Dalenbäck, et al (1992)
ability to differential room geometry
  • Lokki et al (2007)
Realtime auralization system
  • Woszczyk et al (2009)
Virtual Haydn project
  • Brereton et al (2011)
compared measured and modeled impulse responses using ambisonic system
  • Abel et al (2013)
Hagia Sophia emulation in concert setting using close mics

issues with prior work

  • listening only (no interactivity)
  • single musicians
  • "random" auralization signals
  • auralization presented over headphones
  • reliance on close mics
  • custom systems (not reproducible)


To do this we need an auralization system that

  • is real time
  • has adjustable acoustics
  • uses loudspeakers so performers can hear each other in the virtual acoustic space (without headphones)

We also need

  • methods for synthesizing reverberation from a specific acoustic with control over the various parameters of this acoustic (e.g., \(T_{60}\))
  • a reproducible approach


  1. Chiesa di Sant'Aniceto
  2. Impulse response measurement and processing
  3. Felice Anerio's music
  4. Auralization system
  5. Recording approach
  6. Results and analysis

Chiesa di Sant'Aniceto

  • Located in Palazzo Altemps, Rome Italy
  • Completed in the early 1600s
  • Relatively unmodified since its consecration
  • Has a corpus of music written/collected to be performed in the church

Chiesa di Sant'Aniceto

  • We made spatial impulse response measurements from likely performer/listener positions
  • Nominal \(T_{60} \approx 1.8\text{s}\)

impulse response processing

  • Generated statistically independent impulse responses from the Sant'Aniceto measurements
  • Synthesized IRs with longer and shorter reverberation times
    (0.5, 0.7, 1, 1.4, and 2 times the nominal reverb time)


  • The music came from the Altemps Codex
  • Several verses from Felice Anerio (1560–1614) Misericordia Domine
  • Anerio succeeded Palestrina as composer for the Papal Choir

Feedback Canceling Reverberation Architecture

  • We use a feedback canceling auralization system (Abel et al 2018)
  • It uses room microphones and loudspeakers to capture and reverberate sound sources
  • A filter designed from the measured transfer function between mic/speaker pairs minimizes feedback

Recording studio setup

  • Auralization system mics/speakers were positioned to avoid having occluded direct paths
  • Mics for recording were positioned like a typical session recording as well lavalier mics for each vocalist
  • The CCRMA studio has very short reverberation time (\(T_{60} \approx 0.2\text{s}\))

Recording approach

  • Two choirs recorded four verses of Anerio's music in each of six reverberant conditions twice
  • Recorded using a tonmeister approach of well-positioned room microphones
  • Also recorded with close-mics

example session video

Gloria Patri

Gloria Patri

Church Rock: interactive demo

Today in CCRMA recording studio (basement)

now–noon and 1pm–3pm


We showed:

  • A methodology for studying interactions between multiple musicians and variable room acoustics
  • A new method for recording in virtual acoustic environments
  • Preliminary results from a case study of Felice Anerio's music in Ciesa di Sant'Aniceto

Future work includes:

  • More substantial analysis of the recording
  • Studies involving other room acoustics parameters


  • Steve Sano
  • Stanford Chamber Chorale
  • Talya Berger
  • Alessandra Capodifiero
  • Elena Georgieva, Benjamin Josie, Mark Rau, and Ryan Smith
  • Funding from NAKFI ADSEM10