Replacing the Orchestra? - The Discernibility of Sample Library and Live Orchestra Sounds

Area of Science:

• Music Technology
• Acoustic Perception
• Digital Audio Signal Processing

Background:

• Orchestra sample libraries (OSL) are increasingly used in music production.
• The ability of human listeners to differentiate between OSLs and live orchestras is largely unknown.
• This distinction is crucial for understanding the impact of OSLs on music authenticity.

Purpose of the Study:

• To investigate if listeners can reliably identify orchestral music produced using OSLs versus live recordings.
• To assess the performance of different listener expertise levels in distinguishing sound sources.
• To evaluate OSLs against Alan Turing's threshold for convincing simulated performances.

Main Methods:

• An internet-based experiment was conducted with 602 participants.
• Participants listened to a classic orchestral piece produced with a state-of-the-art OSL and a live orchestra recording.
• Listeners were asked to identify the sound source.

Main Results:

• Overall, listeners correctly identified the sound source 72.5% of the time, exceeding Turing's 70% threshold.
• Sound experts showed a higher accuracy rate in identifying the correct source.
• Non-expert listeners, representing the majority of music consumers, achieved only 68.6% accuracy, performing below the threshold.

Conclusions:

• While experts can discern OSLs, non-expert listeners struggle to differentiate them from live orchestras.
• The indistinguishability for most listeners suggests OSLs may become more prevalent due to economic factors.
• This has significant implications for the future of orchestral music production and consumption.