The Vienna 4x22 Piano Corpus

Abstract: In January and February 1999, twenty-two professional pianists from Vienna performed four excerpts from the Classic-Romantic repertoire on a SE290 computer-controlled Imperial grand piano by the Viennese piano manufacturer Bösendorfer. The pianists were advanced performance students and piano professors from the University of Music and Performing Arts Vienna. They were given the scores several days ahead of the recording session to allow to prepare their performances. The recording took place in an approximately 6-by-6-meters studio at the Bösendorfer piano company (Graf-Starhemberg-Straße 14, 1040 Vienna, building demolished in 2012). After a familiarization phase with the piano, the pianists were asked to perform complete takes of each of the four excerpts until they were satisfied and agreed upon a final version. The four piece excerpts were:

• Frédéric Chopin: Etude in E major Op. 10, No. 3, bars 1–21
• Frédéric Chopin: Ballade in F major Op. 38, 1–45
• Franz Schubert: Deutscher Tanz in f minor D.783 No. 15
• Wolfgang Amadeus Mozart: Sonata in A major K.331, first movement, bars 1–18.

The performances were recorded by the SE290 system into its proprietary and encrypted format, resulting in a file triple (.KB, .LP, .SP), converted afterwards to a standard MIDI format and a text representation using Bösendorfer's own SE software. The sound of the imperial grand piano was recorded onto digital audio tape (DAT) by a portable Tascam DA-P1 recorder using two AKG CK91 microphones placed in ORTF stereo technique towards the open lid of the piano. The MIDI files were converted to text and subsequently matched to the symbolic scores, resulting in match files, a text-based format created by Gerhard Widmer et al in 1998.

The data has been collected within Werner Goebl's master project (Goebl, 1999). To date, the data corpus has been used in several scientific studies by several authors. An incomplete list is provided here:

• Goebl, Werner (1999). Numerisch-klassifikatorische Interpretationsanalyse mit dem ‘Bösendorfer Computerflügel’.
  Diplomarbeit (Masters thesis), Universität Wien 1999, 2 Bände, 181 Seiten. (in German) ( 2.4MB)
  
  
• Goebl, W. (1999). Analysis of piano performance: Towards a common performance standard?
  Society of Music Perception and Cognition Conference (SMPC99), Evanston, USA, August 14–17, 1999.
  
  
• Goebl, Werner (2001). Melody lead in piano performance: Expressive device or artifact?
  Journal of the Acoustical Society of America, 110(1), 563–572, doi: 10.1121/1.1376133.
  
  
• Widmer, Gerhard (2001). Using AI and Machine Learning to Study Expressive Music Performance: Project Survey and First Report. AI Communications 14/3, 149–162
  
  
• Langner, J., and Goebl, W. (2003). Visualizing expressive performance in tempo-loudness space.
  Computer Music Journal, 27(4), 69–83, doi: 10.1162/014892603322730514.
  
  
• Pampalk, E., Widmer, G., & Chan, A. (2004). A new approach to hierarchical clustering and structuring of data with self-organizing maps. Intelligent Data Analysis Journal, 8(2), 131–149.
  
  
• Stamatatos, Statis and Widmer, Gerhard (2005). Artificial Intelligence 165/1, 37–56, doi: 10.1016/j.artint.2005.01.007.
  
  
• Dixon, Simon (2005): Live tracking of musical performances using on-line time warping. Proceedings of the 8th International Conference on Digital Audio Effects (DAFx’05), Madrid, Spain, September 20–22, pp 92–97
  
  
• Wang, S., Ewert, S., & Dixon, S. (2016). Robust and efficient joint alignment of multiple musical performances. IEEE/ACM Transaction on Audio, Speech, and Language Processing, 24(11), 2132–2145. doi:10.1109/TASLP.2016.2598318

Browse Dataset