Audio Brand Founder Spends $23,000 Testing 40 Subwoofers and Exposes Trusted Names Lying About Their Specs

Some didn’t even hit half of what their own spec sheets claimed.

For years, subwoofer brands have been able to publish Xmax figures under different assumptions, leaving buyers to compare numbers that were not always measured the same way.

Nick Apicella, founder of ResoNix Sound Solutions, spent about $23,000 commissioning an unnamed Klippel-trained third-party lab for a broader project involving roughly 40 subwoofers. The public data available now covers 21 drivers, all measured against one standard.

Some of those results landed far below the spec sheets. One driver delivered 44% of its advertised excursion, another triggered thermal protection at roughly half its rated power, and a flagship from a brand long respected among DIY enthusiasts hit just 62% of its published spec.

But the same leaderboard also ranks ResoNix’s own prototype subwoofers in the top three slots. Those models are now in pre-order at $725 to $1,150, which means the most detailed public Klippel dataset in car audio also promotes a product line developed through the same testing project.

Four Numbers, One Driver

The divergence between spec sheets and lab results starts with a deceptively simple problem. Xmax, the spec that every car audio buyer checks first, has never had a universally accepted measurement standard.

Adire Audio proved how badly this plays out by publishing three defensible Xmax numbers for one driver, the Shiva. Under IEC-62458 it measured 3.7mm, the 10% THD method yielded 14.7mm, and a 50% BL/compliance threshold pushed it past 16mm. Same subwoofer, same motor, and a four-to-one spread depending on which definition you chose.

That’s why Klippel’s Large Signal Identification module matters because it shifts the question from how far the cone can move to how far it can move before the driver stops behaving cleanly.

While traditional Thiele-Small parameters describe a driver near rest, where most behavior still looks linear, LSI instead maps what changes as the cone moves through real operating stroke, exposing the nonlinear limits that spec sheets often flatten into one number.

The test focuses on three main curves:

• BL(x) shows how motor force falls as the voice coil moves beyond the magnetic gap.
• Kms(x) shows how the suspension stiffens or relaxes across the driver’s stroke.
• Le(x) shows how inductance changes with cone position, which can create audible distortion before the motor-force limit is reached.

In plain terms, a driver can physically reach its advertised excursion and still stop producing clean output well before that point.

This standard isn’t new, though, and it isn’t Apicella’s. Wolfgang Klippel defined these limits in his foundational paper on voice coil peak displacement, and independent testing labs serving automotive OEMs, like Redrock Acoustics, have applied them to automotive drivers since the 1990s.

21 Drivers Under One Standard

The broader Klippel project covers roughly 40 subwoofers, but only 21 driver results are public on ResoNix’s leaderboard at the time of writing. Those published results are graded across 14 weighted categories totaling 1,250 points, alongside a separate 0-to-100 marketing accuracy score that measures how closely each brand’s published specs matched the Klippel measurements.

These scores ranged from more than 900 points to barely above 200, and nine of the 21 published drivers earned marketing accuracy marks below 50.

Built wrong for the published number

The largest misses were not just technical disagreements over how Xmax should be defined. In several cases, the measured behavior suggested drivers that were poorly matched to their published claims.

Wavtech’s thinPRO12 was the clearest example. It claimed 20mm of Xmax but delivered 8.85mm, giving it the largest raw excursion shortfall in the dataset and a marketing accuracy score of 15 out of 100.

The issue appeared structural rather than semantic, since the coil height was improperly sized for the driver’s split-gap geometry.

However, Stereo Integrity fared worse overall, as its BM-11 finished last in the composite ranking at 225 out of 1,250 and scored 5 out of 100 for marketing accuracy. It claimed 14mm and measured 9.31mm, while its suspension sat biased outward by about 6mm. At 17 volts, that offset showed up as roughly 45% total harmonic distortion at 25Hz.

Meanwhile, Adire Audio’s Brahma Mk2.5, a flagship from a brand that built its following in the DIY and SQ community, claimed 28mm and reached 17.46mm. Distortion ran at 12% across 20 to 100Hz, and marketing accuracy scored 35 out of 100.

Specs that pass, drivers that don’t

Not every mismatch came from an inflated BL-based Xmax claim. Some drivers met, or even exceeded, their published excursion number on motor force, then ran into a different limit that most spec sheets never mention.

That second limit was inductance.

The Le 17% variance threshold captures the point where voice coil inductance modulation pushes distortion past the ceiling for practical clean output, even if the cone has not reached its BL travel limit.

In other words, a driver can satisfy the advertised Xmax number and still become too nonlinear to use cleanly at that stroke.

Several tested drivers showed that split clearly:

• Audiofrog’s GB12D4 hit 87% of its 19mm BL claim, but practical clean stroke measured just 5.68mm, only 30% of the published number.
• Focal’s SUB 10WM Utopia M exceeded its BL spec at 20.50mm against a 17mm claim, yet its clean-output ceiling sat at 5.64mm, just 33% of the advertised figure.
• JL Audio’s 12W3v3-4 met its 13mm BL claim at 13.81mm, but clean stroke topped out at 5.38mm, just 41.5% of the published spec.
• JL Audio’s flagship 12W7AE-3 showed the same pattern at a higher tier, exceeding its 29mm BL claim at 34.1mm while being inductance-limited to 18.77mm of practical output, just 65% of the published spec.

Conservative claims, clean output

Other results cut the opposite way. A few manufacturers published Xmax claims that were conservative, accurate, or at least much closer to what the lab measured.

Three non-ResoNix examples stood out:

• Dynaudio’s Esotar2 1200 was the clearest under-promise. The company claimed 10.25mm and the lab measured 14.37mm, meaning the driver was under-rated by 40%. Distortion stayed below 5% from 20 to 120Hz, which helped it score 904 out of 1,250.
• Acoustic Elegance’s SBP 15, tested with the Apollo motor upgrade, paired an understated Xmax claim with unusually clean motor behavior. It measured 118% of its published claim with near-perfect BL symmetry. At 983 out of 1,250, it earned the highest non-ResoNix composite score in the database.
• Audiomobile’s Encore 4412 stood out for precision rather than understatement. It delivered a 19.11mm measurement against a 19.1mm claim, leaving essentially zero gap between marketing and reality. It was one of only two non-ResoNix drivers to achieve a perfect 100 on marketing accuracy.

But the dataset comes with a complication that no scoring formula can resolve.

The Tester Who Sells Subs

ResoNix’s conflict becomes unavoidable once its own prototypes enter the leaderboard. The GUS-15 prototype scored 1,077 out of 1,250, the GUS-12 hit 1,068, and the GUS-10 reached 996. The next independent driver, Acoustic Elegance’s SBP 15 at 983, trailed the lowest GUS score by 13 points.

All three prototypes are scheduled for delivery in early July 2026.

That tension also shows up in how ResoNix presents its own products compared with competitors. Competitor driver pages feature detailed editorial teardowns with marketing accuracy scores, and some include blunt recommendations against purchase. But the GUS prototype pages get different treatment, carrying a short disclaimer instead of a full write-up.

“This is just a pre-production prototype,” each GUS page reads. “Since this is not a production unit, I will not be doing a full write up.”

Measurement was handled by an anonymous lab, identified only as a “third-party, Klippel-trained engineering lab” kept unnamed “per their request to avoid being dragged into potential drama.” ResoNix frames that anonymity as protection from industry blowback. It also leaves readers unable to verify the testing arrangement outside ResoNix’s presentation of it.

A defense built on method

Apicella’s defense rests on process. He argues that a third-party lab performed every measurement, formula-based scoring and blind Photoshop layer-stacking minimize subjective judgment, and the same standard was applied to GUS prototypes and competitors. ResoNix has committed to publishing full Klippel data for every GUS production batch, and any tested manufacturer can request a retest.

Still, one central issue remains unresolved. The same company funding and publishing the leaderboard is also selling the products sitting at its top. Across 49 minutes of a video explicitly titled “Clearing Up Misconceptions,” Apicella addresses the reliability of the method and the motives of critics, but never directly confronts the GUS competing-product conflict.

His broader answer to criticism is that skeptics and manufacturers should publish equivalent data of their own. He argues that he spent his own money, that other companies have something to lose, and that critics should post their own data.

But they do not fully answer how readers should weigh a leaderboard controlled by a company whose own products rank first, second, and third.

Apicella’s challenge to critics reduces to a single demand. Post your own Klippel data. The industry’s response has been even simpler.

What the Industry Said Back

Searches for public responses from JL Audio, Sundown Audio, Audiofrog, Stereo Integrity, Adire Audio, Focal, and Wavtech returned zero on-record rebuttals as of May 2026. Brett Hanes, who spent nearly three decades at JL Audio, most recently as VP of Acoustic Research, before moving to Garmin, lists Klippel Testing as a LinkedIn specialty and has not responded publicly.

The industry’s best available defense predates the testing entirely.

Audiofrog founder Andy Wehmeyer has historically defended specs using the older “Voice Coil minus Gap Height plus 15%” method, which means brands can argue the legacy Xmax approach is legitimate even if Klippel shows different numbers. So, no tested brand has deployed this argument in response to ResoNix.

The most substantive external critique came from RAWCAtOr, a YouTube reviewer who owns the worst-scoring driver, the Stereo Integrity BM-11. He validates the measurements but disputes Apicella’s interpretation, arguing the scoring grades drivers against “an imaginary subwoofer” rather than accounting for real-world use cases.

On the other hand, forum critics on diyAudio turned the scrutiny back on ResoNix, arguing the company’s own GUS Qtc claims wouldn’t survive the same measurement standard.

“If they actually measured the frequency response the qtc would be more like 2,” one commenter wrote.

Rockford Fosgate already publishes “Klippel Verified” Xmax specs on its product pages, which proves no technical barrier prevents disclosure. It’s just that the brands tested by ResoNix have not followed with comparable public data.

The brands’ silence and the community’s measured skepticism leave one question open. What would this project look like without the conflict?

Measurement Without a Product Line

In home audio, independent Klippel-based measurement already has a model that avoids the conflict entirely. Erin Hardison runs Erin’s Audio Corner, owns a roughly $100,000 Klippel Near Field Scanner, discloses per review whether equipment was loaned or purchased, and does not sell speakers.

Because he has no speaker line to promote, his reviews do not ask readers to separate the measurement work from a competing product business.

Car audio doesn’t have its Erin yet. ResoNix published a third-party lab’s data, disclosed its scoring formulas, offered a retest protocol, and committed to GUS batch data, building the most detailed public Klippel dataset for car audio subwoofers in the process.

That makes the project valuable, but not structurally independent. The result is the most detailed public Klippel dataset for car audio subwoofers so far, produced by a company that now sells subwoofers of its own.

Independent measurement in car audio will reach that structural standard only when someone without subwoofers to sell funds the work, and so far no one has. Until then, ResoNix’s data and its conflict have to be read together.

The data exists, the conflict exists, and not one brand with the Klippel equipment to settle the question has posted a single measurement in response.