The defect only surfaces on certain genres, and your favorites might be one of them.
When seven expert listeners at Abbey Road Studios evaluated recycled vinyl pressings in a blind test, the 100% recycled variant finished last. Panelists reported more crackle, more clicks, more background noise.
And the defects weren’t evenly distributed across the test record. Side A, pressed with acoustic post-rock spanning a wide dynamic range and quiet passages, exposed pressing artifacts that Side B’s louder, compressed rock concealed.
Recycled vinyl might be inaudible on a pop album mastered to hit hard. On a record that breathes, the surface noise breathes with it.
Old Stock, New Discs, Blind Ears
The raw material came from a warehouse. Warner Music Group pulled roughly 10,000 obsolete records from European storage and shipped them to GZ Media, the Czech pressing plant that would shred, micronize, and re-press them into new discs at varying blend ratios.
These weren’t defective pressings but unsold stock, the kind that accumulates when demand forecasts overshoot and nobody has a plan for the surplus.
To answer whether recycled material could match virgin PVC on sound, the pilot pressed a bespoke test record in six variants: four recycled blends at 10%, 25%, 50%, and 100% recycled content, a virgin PVC control, and GZ’s own Eco Mix compound made from internal manufacturing scrap.
Both sides included 20-second silent sections designed to reveal surface noise.
Every pressing was anonymized with coded identifiers and handed to a seven-person panel at Abbey Road Studios, which evaluated them blind across two rounds.
What the Quiet Side Revealed
At first glance, the verdict looked reassuring. All variants fell within half a point on a ten-point scale, with the top three separated by just 0.2 points. The report concluded the differences were marginal, but marginal doesn’t mean random.
The blend that scored highest was the 25% recycled variant, which actually outperformed the virgin control. At the bottom was the 100% recycled variant, which drew increased reports of crackle, clicks, and background noise.
That ranking held even when the sensitivity analysis removed individual panelists one at a time, which means it wasn’t driven by a single outlier ear.
The quiet side is what separated them.
B-sides performed better than A-sides across several blends because the louder, compressed rock on Side B pushed the music signal above the noise floor, drowning out surface artifacts that Side A’s quiet, dynamic post-rock left in the open.
On vinyl, the noise is always there. However, the genre decides whether you notice.
So, Showell is right that the discs were consistent, but only inside a tightly controlled pilot that filtered out much of the messier real-world inventory.
From 10,000 Records to Three Usable Titles
Those 10,000 warehouse records didn’t all make it to the pressing plant. The pilot excluded all colored vinyl, all 7-inch and 10-inch formats, and all pre-EU-regulatory pressings that might contain lead stabilizers.
Of the five titles selected for the pilot, both colored vinyl pressings failed during processing, one from thermal degradation and the other from operational constraints. That left three black vinyl titles from different plants as the pilot’s entire usable feedstock.
All that filtering reveals something the headline numbers don’t. At real-world scale, a warehouse of unsold records would span decades of production across plants with different compound formulations and stabilizer chemistry.
The pilot’s screening criteria would eliminate a significant share of that inventory before a single record reached the shredder.
Why this matters shows up in the pilot’s own data.
GZ Media’s Eco Mix, produced from homogeneous internal scrap of known composition, ranked among the top-performing variants and caused zero operational issues, while the mixed-origin feedstock, drawn from records pressed at different plants with unknown compounds, scored lowest and proved hardest to stabilize.
In short, recycling works when you know what you’re recycling.
The process is different today, but the risk pattern has an older precedent. Once plants lose control over the source material, noise becomes harder to predict.
In the 1970s, pressing plants routinely used reground vinyl from returned unsold stock to cut costs, which produced records “much noisier and scratchier” than virgin pressings, with chunks of paper from old LP labels molded into the grooves of new ones.
Modern screening and micronization are better tools than 1970s regrinding, but the fundamental constraint is unchanged. Unknown compounds produce unpredictable results.
Seven Listeners, No Instruments
The report acknowledges its own limits more openly than most industry-funded research. Two of those limits matter more than the rest.
No instruments were used to verify the panel’s assessments. The evaluation relied on expert listening alone rather than comprehensive instrumental measurement of groove geometry or noise spectra across all variants.
Noise fingerprints were measured beforehand as a quality gate, but the report contains no spectral data, no signal-to-noise ratios, and no groove-level measurements that another lab could use to reproduce the results. The quality verdict is subjective consensus among seven listeners, not a dataset.
Those seven listeners all came from inside the project. Every panelist worked for one of the three organizations that co-authored the report, which left Abbey Road Studios, GZ Media, and WMG Studio Services as the only institutions at the table.
No independent mastering engineers, no competing plant operators, and no outside acousticians participated.
The blind methodology and sensitivity analysis guard against the most obvious forms of bias, since the ranking survived the removal of individual listeners.
But structural independence and procedural blindness are different things. A panel composed entirely of stakeholders evaluating a product they helped develop represents a gap the study’s own design cannot close.
Recycled on Paper, Not in the Groove
Even if the quality and measurement gaps were resolved, the pilot’s recommended path forward introduces a different kind of problem.
The report endorses plant-level mass balance accounting, where recycled content is tracked across total factory output rather than physically segregated by disc.
The recycled content is a statistical claim about the factory, not a physical fact about the disc in your hands.
Low-blend recycled vinyl can work. The 25% variant outperformed virgin PVC, and all six variants, including the 100% mixed-origin blend, spanned only about half a point on a ten-point scale.
But the framing has already shifted. In the report’s PDF, Showell assessed the variants as “remarkably consistent, with the level of recycled PVC not making much of a difference to the quality of the end product.”
On GZ Media’s promotional page for the same pilot, that hedge disappeared. The same engineer’s quote became declarative.
Same engineer, same pilot, two framings. The data showed a qualified success, but the marketing has already dropped the qualifications.