US 5,616,876 · Granted 1997-04-01

The Patent That Let You Find Songs 'More Like' Your Favorite

Before Spotify's algorithm, there was a simpler idea: editors would describe what made songs similar (their vibe, energy, style), and then software would let you find new songs that matched a track you loved. You'd drag sliders to fine-tune how close the match had to be, and see what styles were actually playing in your playlist.

Curated and verified by the IsItPatented Editorial team

The plain-English version

What it protects

The claim covers an interactive music system where users can select a seed song and use a 'more like' function to discover similar tracks based on editorial descriptions of musical style. What's protected is the specific mechanism of using style tables (subjective categorizations prepared by human editors) to compute similarity, combined with a style slider that lets the user control how strict the match should be, and a style equalizer with eight faders that visualizes and lets users adjust the balance of song types in the playlist.

Why it matters

This patent represents an early attempt to solve the discovery problem in digital music—how do you find new songs you'll actually like? Rather than waiting for algorithmic machine learning, it proposed a human-editorial layer that made musical taste explicit and controllable. Microsoft filed this in 1995, before streaming as we know it existed, showing the company was thinking about music recommendation years before services like Pandora or Spotify launched. The approach of letting users see and adjust style preferences directly influenced how later services designed their recommendation interfaces.

Real-world use

When you use Spotify's 'Go to Song Radio' feature or adjust the vibe of a playlist with recommended songs, you're using a descendant of this idea—someone's taste profile meets editorial categories to suggest what to play next.

Original USPTO abstract

An interactive network provides music to subscribers. A "more like" function allows a subscriber to use a seed song to identify other songs that are similar to the seed song, and to add the new songs to the current playlist. The similarity between songs is based on the subjective content of the songs, as reflected in style tables prepared by editors. The subscriber may control the closeness of the match by adjusting a style slider provided by the user interface. A style equalizer employs eight faders that indicate the predominant styles of the songs in the playlist. A subscriber may use the style equalizer to see what types of songs are included in the playlist, and to adjust the mix of songs that are played from the playlist.

Patent details

Publication number: US 5,616,876
Filing date: 1995-04-19
Grant date: 1997-04-01
Assignee: Microsoft Corporation
Inventor(s): CLUTS; JONATHAN C.
CPC class: G09B5/14

View original record on Google Patents →

Want to file your own patent?

If you're building a music app or streaming feature that suggests songs based on style, search our database to see how competitors have protected similar discovery mechanisms.

Free patentability scan