Deduplicating Performers Across Names, Pseudonyms and Roles

In short: Noctil's MatchIQ engine resolves the performer names found in sound recording line-ups into unique performer identities. It handles spelling variants, transliterations, stage names and pseudonyms, band memberships and role differences, even when no IPN or ISNI is present. ProcessEngine then merges matched performer records into one complete record under the organization's own rules, sending only uncertain cases to human review. In benchmark testing, performer matching produced fewer than 0.01% false positives or false negatives.

Know every performer once, under every name they record under. For neighbouring rights societies and performer CMOs that need one accurate record for each performer across thousands of sound recording line-ups.

The Situation

Every sound recording has a performer line-up: featured artists, session musicians, vocalists, conductors and other contributors, each with a role. Neighbouring rights societies pay performers based on these line-ups, which arrive from labels, performers and their agents, and partner societies abroad.

Across a repertoire of millions of recordings, the same performer appears thousands of times, often written differently each time. Before anyone can be paid correctly, every appearance has to be linked to one performer identity.

The Problem

Performer names are among the least consistent data in music:

• Name variants. One performer can appear with full names, initials, nicknames, misspellings, different name orders, or transliterations from other scripts.
• Stage names and pseudonyms. Artists record under stage names, aliases and side-project names that look nothing like their legal names.
• Same name, different people. Common names and shared stage names belong to different performers, and merging them pays the wrong person.
• Bands and members. A group name stands for several individual performers whose membership changes over time. Each member’s contribution has to be traced to the right recordings.
• Inconsistent roles. The same contribution can be labelled differently by each source: featured or non-featured, lead or backing vocals, instrument names or codes.
• Missing identifiers. Performer identifiers such as the IPN (International Performer Number) or ISNI are often absent from the line-up data.

Duplicate performer records split one artist’s royalties across several identities. Wrong merges pay the wrong person. Either way, performers lose trust in their society.

The Solution

The Noctil Platform combines MatchIQ for matching with ProcessEngine for workflow, to build one trusted identity for every performer.

1. Name intelligence. MatchIQ recognizes spelling variants, initials, name order, nicknames and transliterations across scripts, so variants of the same name resolve together.
2. Pseudonym and alias linking. Stage names, aliases and legal names are linked to one performer identity, while each name is kept for display and reporting.
3. Contextual disambiguation. To tell apart people who share a name, MatchIQ weighs the recordings they appear on, their roles and instruments, the labels and artists they work with, active years and nationality.
4. Group membership. Bands and ensembles are linked to their individual members over time, so each member’s contributions are traced to the right recordings.
5. Role normalization. Roles and instruments from different sources are mapped to one consistent vocabulary, such as featured and non-featured, and instrument codes.
6. Identifier use and enrichment. IPN, ISNI and IPI numbers are used wherever they exist. When they are missing, MatchIQ infers the identity, and newly confirmed identifiers enrich the performer record.
7. Workflow orchestration with ProcessEngine. ProcessEngine merges matched performer records under the organization’s rules. It sends uncertain matches, such as a possible pseudonym or a shared name, to human review, and records every decision for audit.

The Result

• Fewer than 0.01% false positives or false negatives in performer matching, based on benchmark testing
• One identity per performer, with every name, pseudonym and group membership linked
• Royalties paid to the right performer, without duplicate records splitting their earnings