Music Discovery and Taste Archaeology: Going Backwards to Go Forward

Related: how-your-taste-works Builds-on: conversations about music taste across 2025-2026 Informs: personal listening, discovery practice

Research question: How do you systematically discover music you don't know you'd love — by going backwards through influence chains, genre genealogy, and cultural history rather than forward through algorithmic recommendations?

Your Music DNA (Compiled from Conversations)

From ~15 conversations about music across 2025-2026, a clear profile emerges. This isn't guesswork — it's what you've said and responded to.

Core Identity

Groove-forward with harmonic sophistication. You prioritize pocket, instrumental interplay, and negative space over vocals as the primary hook. Drawn to "musician's music" — ensemble thinking, intentional arrangements, restraint. Pre-loudness-wars production aesthetic. You want to hear the room, the dynamics, the actual performance.

The Roots

• 90s R&B from FM North Wave (82.5 FM, Sapporo) — Boyz II Men, Mariah, TLC, Usher. Heard in the car, your aunt loved it. North Wave launched in 1993 as part of the JFL (Japan FM League, alongside J-Wave and FM 802) — the hipper alternative to JFN stations, heavy on Western R&B/pop. This wasn't a choice — it was the water you swam in. 90s-rb-in-sapporo-tracing-a-musical-memory
• Japanese R&B/pop — Misia, Utada Hikaru (CDs owned in high school in Kobe). The Japanese side of the same 90s R&B wave.
• Jazz foundation — John Coltrane's My Favorite Things is the foundational album. Modal exploration, rhythmic complexity, dynamic space, ensemble conversation.

The Current Map

What You've Described Loving

• Pocket — everyone locked in, bass and drums married
• Negative space — restraint, not filling every gap, silence as composition
• Harmonic vocabulary — jazz chords, tension/release, not basic pop progressions
• Recorded performance — hear the room, the attack and decay, instruments interacting
• Ensemble thinking — music as conversation, not individual showcase
• Craftsmanship over spontaneity — the Cornelius NPR set: "treating the live set like another production challenge to solve perfectly"

What You've Explored

• building-an-ai-dj-that-understands-music-taste — explored the idea of an LLM that understands music taste with cultural/geographic context. Concluded nobody has built the conversational cultural intelligence layer yet.
• chemical-brothers-generative-synth — built a generative synth inspired by Star Guitar / Further
• music-across-cultural-landscapes — explored how to translate bands across Japanese/Western contexts (Cornelius ↔ Washed Out, Sakanaction ↔ Phoenix, SAS ↔ DMB)
• music-streaming-app-alternatives — looked at Jellyfin, Navidrome, self-hosted options
• audiophile-cdbluetooth-source-component-options — looking for CD/BT/radio combo for an audiophile setup with existing amp/speakers

The Discovery Problem

Algorithms are good at giving you more of what you already like. You want the opposite: go backwards through the influence chain to find what you don't know you're missing.

The algorithm says: "You like FKJ → try Tom Misch." That's lateral. You want archaeological: "FKJ exists because of Herbie Hancock's Head Hunters → which exists because of Sly Stone → which exists because of James Brown's rhythm section innovations → which connects to the Afrobeat of Fela Kuti → which connects to the highlife of E.T. Mensah." Now you're somewhere you'd never arrive by algorithm.

Why Algorithms Fail for Deep Discovery

1. Recency bias — streaming platforms optimize for current catalog. They rarely surface a 1972 Ethiopian jazz record even if it's exactly your taste.
2. Popularity weighting — collaborative filtering favors well-listened tracks. The deep cut that 200 people love doesn't surface against the album track that 2 million streamed.
3. Genre silos — platforms categorize music into rigid genre buckets. Your taste crosses neo-soul, jazz, electronic, Japanese pop, and French touch. No single genre feed captures that.
4. No cultural context — Spotify doesn't know you grew up with FM North Wave (82.5 FM). It doesn't know that Hokkaido-to-Kobe-to-Seattle shaped how you hear music. It can't connect your taste to place and time.
5. Engagement optimization — they want you listening for hours. A system that says "stop and really sit with this one album" hurts their metrics.

The Discovery Toolkit: Going Backwards

Tier 1: Influence Tracing (The Archaeology)

WhoSampled — whosampled.com The most powerful backward-discovery tool. Input any artist you love, see who they sampled, who sampled them, and the connections between. 1.2M+ songs, 400K+ artists. Community-verified.

Your move: Start with artists whose samples you'd recognize — Nujabes, J Dilla, Robert Glasper, Madlib. Trace every sample back to its source. The source records are your discovery playlist. This is digital crate digging with a map.

The "Six Degrees of Musical Separation" feature connects any two artists through a chain of samples/covers/remixes. Try connecting Coltrane to FKJ. The path will surprise you.

AllMusic.com — influence chains Look up any artist. See "Influenced by" and "Followed by" sections. This creates a direct genealogy. Follow the "Influenced by" links backwards from your current favorites to find the roots you haven't heard.

Musicmap — musicmap.info Interactive genealogy of all popular music genres from 1870 to present. Click into a genre, see what it came from, what it became. Visual, exploratory, built for the kind of wandering you're describing.

Tier 2: Genre Exploration (The Map)

Every Noise at Once — everynoise.com Created by former Spotify data scientist Glenn McDonald. 6,000+ genres plotted on a scatter map. Click any genre, hear a sample. Down = organic, up = mechanical/electric. Left = dense/atmospheric, right = spiky/bouncy.

Your entry points based on your taste:

• Search "shibuya-kei" → click around nearby genres
• Search "neo-soul" → trace adjacent clusters
• Search "japanese jazz" → follow the neighborhood
• Search "french touch" → see what's adjacent that you've never named

Note: stopped updating after McDonald was laid off from Spotify in 2023, but the static map is still the best genre exploration tool that exists.

Rate Your Music (RYM) — rateyourmusic.com The gold standard for deep-cut discovery. Community-driven ratings, genre charts, user lists. The key feature: custom genre charts by decade and country.

Your move: Build custom charts for:

• Japanese jazz, 1970s-1980s (City Pop's jazz-funk roots)
• Brazilian jazz-funk, 1970s (Azymuth, Marcos Valle — groove-forward, harmonic, ensemble)
• British trip-hop, 1990s (the bridge between your downtempo taste and your R&B roots)
• French electronic, 2000s (Ed Banger, Kitsuné — Breakbot's world)
• Ethiopian jazz, 1960s-1970s (Mulatu Astatke — modal, groove-heavy, completely outside your current map)

RYM's user lists are the crate digger's bible. Search for themed lists like "groove-forward jazz," "Japanese electronic essential," "negative space in production."

Tier 3: Sample Archaeology (The Roots)

Discogs — discogs.com The world's largest music database. Every pressing, every release, every credit. Cross-reference producers and session musicians across albums. Find the bass player from your favorite Steely Dan track and see what else they played on.

Your move with Discogs: Follow the session musicians. The groove-forward sound you love was built by a relatively small community of players:

• Pino Palladino (D'Angelo's Voodoo, The Who, John Mayer) — follow his discography
• Thundercat (Kendrick Lamar's TPAB, his own solo work, Flying Lotus) — the modern equivalent
• Larry Graham (Sly & the Family Stone, Graham Central Station) — invented slap bass, the technique that feeds into everything you listen to

Bandcamp — bandcamp.com For CURRENT discovery, not backwards archaeology. But the tag system is powerful:

• Compound tag searches: "neo-soul + tokyo" or "jazz-funk + live"
• Location-based: explore specific city scenes
• Sort chronologically instead of by popularity to surface hidden gems
• Follow labels, not just artists (Brainfeeder, Stones Throw, Ed Banger)

Tier 4: Cultural Context (The Why)

BlackTape — music discovery with cultural context Lets you search by genre + country + decade and follow influence lines. Has a local AI layer for contextual searches like "electronic music from Southeast Asia with gamelan influences." Closest thing to the "AI DJ with cultural intelligence" you described wanting.

Radio Garden — radio.garden Listen to live radio stations worldwide on a 3D globe. Spin to Tokyo, Sapporo, São Paulo, Addis Ababa, Lagos, Paris. This is how you accidentally discover entire scenes you didn't know existed — the same way AFN Sapporo shaped your taste as a kid, except now you can choose the station.

NTS Radio — nts.live London-based community radio with shows covering everything from Japanese ambient to 70s Nigerian disco to Brazilian baile funk. The archive is massive. Each show is curated by a human with deep expertise in a specific niche. This is the closest thing to having a knowledgeable record store clerk.

The Discovery Practice: A System

Algorithms give you breadth. You want depth. Here's a practice for going backwards systematically:

The Weekly Dig (1-2 hours/week)

Step 1: Pick an anchor. One track from your current rotation that you love. This week's anchor.

Step 2: Trace backwards. Use WhoSampled + AllMusic to find 3 things:

• What this artist sampled or was influenced by
• Who played on the record (session musicians, producer)
• What genre/scene it came from

Step 3: Go one layer deeper. Take the most interesting influence you found and repeat. Now you're two generations back from your anchor. This is where discovery happens — the record from 1974 that sounds like nothing you've heard but explains everything you love.

Step 4: Listen properly. Pick one album from the dig. Listen front to back. On your audiophile setup. Not as background. This is the part algorithms can't do — the sit-with-it discovery that turns a data point into taste.

The Seasonal Deep Dive

Once per season, pick a cultural/geographic thread and go deep for a week:

These aren't random — each connects to something in your existing taste. Japanese jazz-funk IS your groove-forward DNA in its origin form. Brazilian groove is the harmonic sophistication + pocket you love, in Portuguese. British electronic bridges your downtempo and R&B roots. West African groove is where the rhythmic complexity you hear in Coltrane and D'Angelo comes from.

The Taste Genealogy: Your Specific Influence Tree

Working backwards from what you love now to what you haven't heard yet:

Your taste (2026)
├── Neo-soul / alt-R&B (The Internet, Mac Ayres, Glasper)
│   ├── D'Angelo's Voodoo (2000) ← YOU NEED THIS
│   │   ├── Prince (Dirty Mind, Sign o' the Times)
│   │   ├── Sly & the Family Stone (There's a Riot, Fresh)
│   │   └── Marvin Gaye (What's Going On, I Want You)
│   ├── Erykah Badu (Baduizm, Mama's Gun)
│   │   ├── Billie Holiday
│   │   └── Stevie Wonder (Innervisions, Songs in the Key)
│   └── J Dilla (Donuts) ← the production ancestor
│       ├── Ahmad Jamal (live recordings, space and pocket)
│       └── James Brown's rhythm section
│
├── Jazz-adjacent groove (FKJ, Tom Misch, Kiefer)
│   ├── Herbie Hancock (Head Hunters, Thrust) ← START HERE
│   │   ├── Miles Davis (Bitches Brew, On the Corner)
│   │   └── Sly Stone
│   ├── Roy Ayers (Everybody Loves the Sunshine)
│   ├── Azymuth (Light as a Feather) ← Brazilian jazz-funk
│   └── Ryo Fukui (Scenery) ← Japanese jazz, Sapporo-born
│
├── Japanese electronic (Mondo Grosso, Cornelius, D.A.N.)
│   ├── Yellow Magic Orchestra ← if you haven't gone deep, do it
│   │   ├── Kraftwerk
│   │   └── Haruomi Hosono solo work
│   ├── Shibuya-kei scene (Pizzicato Five, Fantastic Plastic Machine)
│   │   └── Burt Bacharach, Phil Spector (the Western sources Shibuya-kei sampled)
│   └── City Pop (Tatsuro Yamashita, Mariya Takeuchi, Taeko Ohnuki)
│       └── American AOR/yacht rock (the source material)
│
├── French touch (Phoenix, Breakbot, Daft Punk)
│   ├── Chic / Nile Rodgers ← the root of all French house
│   ├── Giorgio Moroder (Italo-disco)
│   └── Cerrone (Love in C Minor)
│
├── UK big beat / energy (Chemical Brothers, Prodigy)
│   ├── Rave culture (acid house, breakbeat)
│   │   ├── Detroit techno (Derrick May, Juan Atkins)
│   │   └── Chicago house (Frankie Knuckles)
│   └── Punk energy → post-punk → electronic crossover
│
└── Coltrane → modal jazz → groove foundation
    ├── McCoy Tyner solo work
    ├── Pharoah Sanders (Karma) ← spiritual jazz
    ├── Alice Coltrane (Journey in Satchidananda)
    └── Ethiopian jazz (Mulatu Astatke) ← modal + groove + totally different

The Priority Listens (What You Almost Certainly Haven't Heard But Will Love)

Sources

Discovery Tools

• WhoSampled: Discover music DNA
• Musicmap: Genealogy of popular music genres
• Every Noise at Once: Genre scatter plot
• Rate Your Music: Community discovery
• Bandcamp: Tag-based discovery
• NTS Radio: Curated community radio
• Radio Garden: Global radio on a globe
• AllMusic: Artist influence chains
• Discogs: Session musician cross-referencing

Conversation References

• building-an-ai-dj-that-understands-music-taste — Nov 2025, explored LLM-based music discovery with cultural context
• 90s-rb-in-sapporo-tracing-a-musical-memory — Nov 2025, traced 90s R&B roots from AFN Sapporo
• chemical-brothers-playful-production-style — Nov 2025, explored the energy/workout lane
• music-across-cultural-landscapes — Oct 2025, how to translate bands across Japanese/Western contexts
• breakbots-recent-music-and-current-activity — Nov 2025, French touch nostalgia
• chemical-brothers-generative-synth — Oct 2025, built a generative synth
• music-streaming-app-alternatives — Sep 2025, explored self-hosted options
• audiophile-cdbluetooth-source-component-options — Dec 2025, building an audiophile listening station