CLR · TRANSMISSION 000 ARCHIVE ONLINE SIG / NOISE 47.2 dB TAPE TEMP 21.4° C 04 · 28 · MMXXVI
Center for Listening Research a Zabriskie division · Pittsburgh, PA
EST. MMXXVI · PITTSBURGH · ACTIVE RESEARCH

Working
notes
on the jam.

We're building tools to measure how a recorded jam moves through key, chord, drone, and texture. Early days. The corpus is small on purpose; the goal is to figure out what to measure and why, then say so out loud. Dispatches are case notes from that work. Not a guide to the “best” jams, not a leaderboard, not a catalog.

Read the Case Notes Meet the Researchers
01 // PERSONNEL

Resident Researchers

One PhD. One goose. Several thousand hours of tape.
Christopher Meiklejohn CLR-001
RESEARCHER · CLR-001

Christopher Meiklejohn

Director · Improvisation, Setlist Topology & Tape Archaeology

PhD, Software Engineering. Runs the Center single-handed (well, almost. See CLR-002). Corpus on a stack of second-hand drives in a closet next to the dehumidifier; methods on a whiteboard above it. Believes a setlist is just a graph that's been very rude to its labels. Also writes at christophermeiklejohn.com.

Software Engineering DSP Tapes since '94
CLR-002
LISTENER-IN-RESIDENCE · CLR-002

Goose

Quality Control · Drive Warmth · Critical Honking

A goose. Yes, that goose. Sleeps on the warm drives. Has never honked at a Type I. Has honked at, among others, the 1995-12-29 Worcester Bathtub Gin and the 2026-04-24 I Would Die 4 U. We are not making this up. The band has asked about the naming overlap. We have no comment.

Resident · est. 2024 Reliable on BTG Cannot be reached by email
02 // DISPATCHES

From the Notebook

Case notes from a small research lab. Method development, one performance at a time. Not a survey, not a ranking.
03 // PROCESSING

Active Studies

Open questions and works-in-progress. Some pan out, some don't. Both kinds get logged.
RUN ID WORKING TITLE HYPOTHESIS STATUS PROGRESS
CLR-RUN-010 Does a 30-minute jam evolve in place? When a band stays in one key with the same chord palette for half an hour, the music still moves. Self-novelty over the jam window should expose that motion, even when every other detector reports stillness. Active 15%
CLR-RUN-011 The algorithm vs. the jamcharts Run the 4-subtype taxonomy against the full Phish.net jamcharts (≈350 community-tagged Type II performances). Where does the machine agree with the fans? Where does it dissent? The dissents are the interesting cases. Collecting 10%
CLR-RUN-012 Listening for quotes Worcester '95 Bathtub Gin folds in The Who's "The Real Me." Memphis '95 Tweezer brushes against 2001, Gypsy Queen, and Slave. A chord-progression cross-correlation against a tease dictionary should automate the catch. Collecting 8%
CLR-RUN-013 Does Goose really sound like Phish-Lite? A persistent claim in two communities. Test it: cluster MERT embeddings across both bands' jams. If Goose lives nearer to Phish than to any other modern outfit, the claim has a measurable basis. If not, it doesn't. Active 28%
CLR-RUN-014 The bag-of-tricks detector Inside one 30-minute jam, the band turns over: funk → ambient → bluegrass → metal. Per-segment MERT should cluster those style segments. The number of clusters per jam ought to track how "type-II" a take feels in practice. Collecting 5%
04 // ON THE BENCH

A small sample, on purpose.

This isn't the full Phish or Goose catalog. It's a curated bench of performances we've hand-picked for method development: community-flagged Type IIs, songs with multiple comparable versions, and a rolling slice of recent shows. Phish has played 2,000+ shows; Goose, several hundred. We ingest selectively because every metric needs to be built and argued about one performance at a time.

124perfs
In the working sample · 1992 to 2026
24songs
Across 3 bands · for cross-version comparison
1PhD
+ 1 goose, in a basement, in Pittsburgh
10cases
Candidate matches · pending listening review

“Treat each dispatch as a worked example, not a verdict. The point is to develop the vocabulary (what a measurement means, where it fails, what it misses) by arguing about a small number of recordings in detail. The catalog comes later, if it comes at all.”

FOUNDING MEMO · CLR · MMXXVI