Raw audio recordings lack semantic searchable metadata about musical content

Evidence

However, as the databases of music nowadays contain millions of tracks, it would be useful to be able to make automatic queries on the database based on some higher-level specifications. For example, if only audio signals are available, it is not possible to make the request ‘I want to hear some sturdy bebop with no trumpet, a female vocalist singing in French, and with electric guitar’. A musically trained listener may be able to pick out such tracks using his or her memory and by listening to available tracks. A computer-based solution would be appealing, since such a task could potentially be automated and computers are much faster at such tasks than humans. — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 2577

Categorization

Category: Factual Claims

• CANDY: The ‘sturdy bebop with no trumpet, female vocalist in French, electric guitar’ query is a delightful concrete illustration of an abstract problem.
• XREF: Connects to the reader’s interest in how informational content hides in raw signals, paralleling signal-vs-noise themes in the soundscape work.

Related

• Music databases need automatic high-level content querying beyond raw audio signals. (similarity 0.89)
• Music information retrieval serves both casual listeners and music professionals needing structured database queries (similarity 0.78)
• MIR metadata is largely assigned by expert listeners, motivating automation as music creation accelerates (similarity 0.78)