Loading Signal…
We believe that the music we listen to isn't just a preference—it's a direct reflection of how our brains process emotion, rhythm, and life itself.
Music has always been the ultimate unifier. Before language, there was rhythm. The frequencies we gravitate towards govern our physiological state and emotional baselines.
Our core thesis is simple: People whose brains crave the same sonic architectures are fundamentally compatible. Whether for deep friendships, romantic relationships, or creative collaborations, shared musical DNA is one of the strongest predictors of resonance between two human beings.
MusicDNAmatch wasn't built to just analyze data; it was built to bridge the gap between isolated listeners, using quantitative measurements of art to bring people together.
We map human musical taste into a mathematical space consisting of 12 distinct axes. When you provide public Spotify playlists, YouTube videos, or Last.fm profiles, our engine extracts the core metadata, genres, and audio features (such as acousticness, valence, danceability, tempo, and energy) of up to 50 tracks.
Each track is broken down into numerical values representing its mood, pacing, and instrumentation. We aggregate these features to find the mean frequencies and distributions of your listening habits.
We utilize natural language processing (NLP) to cluster the countless specific sub-genres associated with your tracks, combined with social tag data from Last.fm, into broader, comparable categories across our global pool.
These inputs are then combined with your explicit genre selections to calculate a single, unified 12D array: your Musical DNA Vector.
Once your node is established in the matrix, finding your "soulmates" becomes a geometry problem. We query the database for other active nodes and compute the Cosine Similarity and Euclidean Distance between your 12-dimensional vector and theirs.
Because the dimensions are normalized (ranging from 0.0 to 1.0), distance represents true structural discordance.
Beyond vector similarity, our engine performs a second pass to identify Direct Overlaps — finding specific shared tracks and common artists in your listening histories to reveal immediate conversational starting points.
Coherence is a mathematical measurement of how consistent or "focused" your musical taste is across our 12 DNA dimensions. We calculate the Weighted Variance of your track data.
A High Coherence (e.g., 85%+) means you have a highly specialized sound profile with sharp preferences (you strongly love specific traits and strongly dislike others). A Low Coherence means your tastes are eclectic and spread out across many different styles.
Every profile in our database is assigned a 12-dimensional array. When you click "Find Soulmates", our engine runs a Cosine Similarity Search inside the database using pgvector. It compares your 12 points against everyone else's 12 points to find the profiles geographically closest to you in that mathematical multi-dimensional space.
We don't just use "Rock" or "Pop" to match you. We use 12 distinct auditory axes ranging from Spectral Energy (intense soundscapes) and Rhythmic Drive (groove-forward music) to Melodic Warmth and Experimental Index. These dimensions capture exactly why you like the music you do.
While vector matching handles "vibes," direct overlap matching handles "facts." After we find someone with a similar DNA signature, we cross-reference your normalized track titles and artist names. This highlights specific common ground, like when two soulmates both have the same rare Tame Impala b-side or share a specific niche Jazz artist in their top 50 tracks.