Friday, May 13, 2005

Appropriate uses

High-definition MIDI is a means of encoding musical performance data with a high degree of accuracy. Currently used by Zenph Studios on grand pianos, the code can indicate very precise key pressures and pedal motions. This first part is not new, and not specific to Zenph Studios, though they are the only ones using Yamaha's HD MIDI specs. The Ohio State University Center for Music Cognition has a computer-controlled Bosendorfer grand piano that utilizes HD MIDI to record piano performances for analysis.

Where Zenph Studios has made strides is in the transcription of recorded sound to MIDI encoding. Ohio State's lab uses cameras and pressure sensors in the piano to encode performances, whereas Zenph has worked out a means of converting recorded WAV files, purely acoustic information, to MIDI files. Where this is particularly tricky is in polyphonic musical textures, with several notes playing at the same time (or very near the same time). Previous attempts at transcription software had only achieved a success rate of 80 - 90%, with up to 10% of the notes missing and another 10% transcribed wrongly. Zenph claims a 100% accuracy to its transcriptions, using proprietary technology.

This means of transcription can be immensely useful in performance research, looking at microvariations in pressure and timing that distinguish Horowitz from an amateur. It can also reveal important structural aspects of music that necessitate performance interpretations.

However, Zenph is also proposing two related uses of its technology that are inappropriate in my eyes (and ears). First, to celebrate their technological breakthroughs Zenph has arranged for a MIDI-generated performance from transcriptions of recordings by Glenn Gould and Alfred Cortot. Next Thursday, National Chopin Piano Competition winner Mei-Ting Sun is giving a recital in Fletcher Opera Theater, Raleigh, North Carolina. During this recital recreations of Gould's 1955 recording of the Goldberg Variations and Cortot's 1928 recording of a Chopin prelude will be performed. Related to this, Zenph proposes to use the technology to re-record old performances to provide a variety of improvements:
better piano (its timbre or richness)
better piano tuning (particularly individual out-of-tune strings)
better piano voicing (how the hammers hit the strings)
better room acoustics
less background noise – no interruptions from cars, coughs, airplanes, etc.
better microphones, more (or fewer) microphones
better microphone placement
better recording equipment
recorded at a better (higher) bit rate


I feel this use is out of step with current archival practices in the other arts. Preservation efforts for documents focus on maintaining current physical integrity, rather than improving the legibility. Preservation efforts in the visual arts attempt to restore the original appearance. In music, this is seen with efforts to denoise recordings, sometimes using very high tech efforts. But Zenph is proposing to "improve" the performances, creating sounds that had never existed before. My problem with this is that most of the fixes suggested - room acoustics, voicing, piano quality, and tuning - are things that the pianists had accounted for in their performances. Fats Waller might have chosen different voicings if his piano had been tuned differently. Arthur Rubenstein might have chosen to articulate a particular passage differently if the piano was different or the room had a different reverberation time. Zenph will rarely be able to account for all the variables that made up the original recorded performance, thus the archival possibilities of this technology is questionable at best. At worst, the new recordings will be the musical equivalent of John Wayne shilling Coors Light.

No comments: