xAI, Crypto Data Centers Emit Infrasound Tied to Illness, Tests Show

Datacenters Behaving Like Acoustic Weapons — Benn Jordan

Benn Jordan

Sub-20 Hz acoustic pressure waves from high-density computing facilities directly trigger vestibular and cardiovascular distress in humans while escaping conventional noise regulations.

AI-assisted summary of Datacenters Behaving Like Acoustic Weapons — original source. Check the original for full context.

Source published:

Key ideas

  • Field measurements reveal that sub-20 Hz infrasound emissions around facilities like xAI's Colossus and the Marathon Digital crypto mine register up to 10 dB louder than audible mechanical noise, propagating through ground and structures without attenuation. (7:15)
  • A double-blind psychoacoustic study of 74 vetted participants exposed to data center infrasound at reduced amplitudes showed dizziness jumping from 1.2 to 3.0 out of 10 and general discomfort quadrupling from 1.2 to 4.8. (23:16)
  • Industrial data center operations pose severe localized environmental threats, including unauthorized on-site methane turbines discharging hazardous nitrogen oxides in South Memphis and 70-90 dB continuous noise combined with sub-audible pressure in Granbury, Texas. (4:09)
  • Because over 3,000 new industrial data centers are planned across the United States, affected communities are encouraged to deploy continuous low-frequency monitoring tools like the Raspberry Shake 3D to establish immutable baseline data for regulatory and legal remediation. (25:42)

Detailed notes

## Background Published by Benn Jordan on YouTube on February 18, 2026, this investigative video explores the hidden physical and acoustic impacts of massive computing facilities. The description frames the project as "another depressing adventure in science and technology" documenting an ongoing experiment and research paper examining how data centers emit low-frequency sound.

## TL;DR Musician and researcher Benn Jordan investigates the physical toll of sub-audible infrasound emitted by massive computing infrastructure, traveling to Elon Musk's xAI Colossus data center in Memphis, a Bitcoin mine in Granbury, Texas, and the Permian Basin. Finding that low-frequency oscillations beneath 20 Hz can register significantly louder than audible mechanical noise, Jordan links these emissions to real-world medical crises reported by nearby residents, including nausea, dizziness, cardiovascular strain, and seizures. To validate these physiological effects, he conducts a deceptive, double-blind psychoacoustic trial exposing over 100 unknowing participants to simulated data center infrasound, documenting sharp, statistically significant spikes in dizziness and overall discomfort. Highlighting regulatory blind spots and the thousands of new facilities under construction, Jordan urges vulnerable communities to capture continuous baseline seismic and infrasonic recordings to build evidentiary trails for future legal battles.

## Atmosphere & style The presentation carries an eerie, investigative noir vibe, balancing deadpan dark humor with genuine scientific gravity and righteous indignation. Jordan moves between covert drone reconnaissance, spectral acoustic analysis, and deeply personal community interviews with an understated, dry demeanor that keeps the material grounded despite its surreal implications. Watching the video feels like listening to an audio engineer uncover a quiet public health conspiracy in plain sight, transitioning from technical sound design talk to chilling revelations about the physical and psychological toll exacted on ordinary neighborhoods.

## Bulletin - [07:15](https://www.youtube.com/watch?v=_bP80DEAbuo&t=435s) Field measurements reveal that sub-20 Hz infrasound emissions around facilities like xAI's Colossus and the Marathon Digital crypto mine register up to 10 dB louder than audible mechanical noise, propagating through ground and structures without attenuation. - [23:16](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1396s) A double-blind psychoacoustic study of 74 vetted participants exposed to data center infrasound at reduced amplitudes showed dizziness jumping from 1.2 to 3.0 out of 10 and general discomfort quadrupling from 1.2 to 4.8. - [04:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=249s) Industrial data center operations pose severe localized environmental threats, including unauthorized on-site methane turbines discharging hazardous nitrogen oxides in South Memphis and 70-90 dB continuous noise combined with sub-audible pressure in Granbury, Texas. - [25:42](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1542s) Because over 3,000 new industrial data centers are planned across the United States, affected communities are encouraged to deploy continuous low-frequency monitoring tools like the Raspberry Shake 3D to establish immutable baseline data for regulatory and legal remediation.

## Key takeaways - [00:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=0s) Infrasound emissions below 20 Hz induce genuine physiological pathology, including cortisol spikes, vestibular balance disruption, and long-term risks of vibroacoustic disease. - [04:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=249s) Elon Musk's xAI Colossus facility in South Memphis bypassed grid limitations by deploying unpermitted methane gas turbines that discharge toxic nitrogen oxides and volatile organic compounds into neighboring residential air. - [07:15](https://www.youtube.com/watch?v=_bP80DEAbuo&t=435s) Sub-audible infrasound around Colossus registered 10 dB louder than audible frequencies, carrying extensive physical energy capable of shaking homes up to a mile away. - [09:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=549s) Residents in Hood County, Texas, an area deeply rooted in oil and gas extraction, reported unprecedented health crises including severe nausea, dizziness, and seizures near the MARA Bitcoin mining facility. - [10:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=600s) Continuous field recordings at Granbury, Texas, revealed sharp sub-audible pressure spikes near 30 Hz that match sound signatures of jet engines preparing for takeoff when sped into the audible spectrum. - [15:13](https://www.youtube.com/watch?v=_bP80DEAbuo&t=913s) Recordings from the Permian Basin confirmed that heavy oil extraction and fracking operations produce substantially lower infrasonic amplitudes than those measured outside computing data centers. - [20:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1200s) A concealed infrasound trial conducted on synthesizer convention attendees demonstrated significant increases in eye irritation, restlessness, nausea, and disorientation among exposed subjects. - [23:16](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1396s) Analyzing absolute survey scores over sensational relative percentages showed that infrasound exposure primarily drives substantial baseline shifts in overall physical discomfort and vestibular dizziness. - [25:42](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1542s) Accurate detection of continuous industrial infrasound requires specialized equipment such as the Raspberry Shake 3D or lithium battery vibration sensors paired with seismic-to-Pascal mathematical conversion. - [25:42](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1542s) Modern municipal zoning and environmental agencies fundamentally fail by regulating infrasound as a minor acoustic nuisance rather than an ecological and public health emergency. - [25:42](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1542s) Establishing continuous pre-construction infrasound and seismic logging strips data center developers of the standard legal defense that low-frequency hums are pre-existing ambient noise.

## Detailed walkthrough ### [00:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=0s) The Data Center Paradox and Infrasound's Hidden Threat

Benn Jordan introduces the jarring irony underlying modern technology criticism: anyone watching his video is likely not a fan of data centers—or outright despises them—yet is actively streaming the video from one, while Jordan himself must upload his investigative research directly into that same infrastructure. Acknowledging that this hypocrisy is unavoidable unless one packs a bag to become a hunter-gatherer, he pivots to the alarming human toll that prompted his cross-country investigation. Residents living near massive computing facilities have reported devastating health crises, including a six-year-old girl who began suffering severe seizures, prompting her family to abandon their home and take out a second mortgage. Jordan sets out to document infrasound noise pollution at major installations, fly covert drone missions to inspect environmental non-compliance, and stage a double-blind human trial exposing over 100 unknowing participants to these exact recorded frequencies.

Moving past familiar debates about raw water consumption, strain on the electrical grid, and toxic chemical usage, Jordan focuses squarely on infrasound—acoustic pressure oscillations below 20 Hz, just beneath the threshold of human hearing. Elevated exposure to these low-frequency waves does not produce mere psychosomatic discomfort or anxiety-induced symptoms; it induces concrete, measurable pathology. Infrasound triggers sharp spikes in cortisol (the primary stress and hypertension hormone), induces vestibular disturbance leading to vertigo and nausea, and causes high-frequency hearing loss, shortness of breath, and cardiovascular strain as the heart works harder to circulate blood. Most alarmingly, chronic exposure is linked to vibroacoustic disease, a condition characterized by the abnormal proliferation of extracellular matrices that thickens blood vessel walls and restricts cerebral blood flow long after the sound itself has ceased. As Jordan emphasizes regarding the biological reality of this phenomenon: “So it's not like this anxiety-based or obsessive compulsive thing where an unpleasant feeling or an unusual situation makes you manifest symptoms.”

- Physiological effects of infrasound: Cortisol spikes, vestibular imbalance, vertigo, nausea, and shortness of breath. - Long-term clinical risks: Thickening of blood vessels from vibroacoustic disease, permanent reduction of cerebral blood flow, and high-frequency hearing loss.

```mermaid flowchart TD accTitle: Physiological Impacts of Infrasound Exposure accDescr: Flowchart showing how sub-20 Hz acoustic pressure waves lead to immediate and long-term health consequences.

subgraph Source["Acoustic Source"] A["Industrial Infrasound (<20 Hz)"] end

subgraph Acute["Acute Physiological Responses"] B["Cortisol Spike & Hypertension"] C["Vestibular Imbalance, Nausea & Vertigo"] D["Increased Cardiovascular Workload"] end

subgraph Chronic["Long-Term Pathology"] E["Vibroacoustic Disease"] F["Extracellular Matrix Growth"] G["Vascular Thickening & Reduced Cerebral Blood Flow"] end

A --> B A --> C A --> D A --> E E --> F F --> G ```

[03:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=180s) Mapping the Measurement Sites

To gather an unimpeachable dataset, Jordan targeted locations that represent distinct acoustic environments across the United States. He needed industrial sites of massive scale where neighboring communities were actively reporting classic symptoms of infrasound illness, contrasted against standard baseline environments and an absolute natural control devoid of human activity.

He established five distinct recording targets across the country: - **Colossus Data Center (Memphis, Tennessee):** The massive xAI training facility housing hundreds of thousands of GPUs. - **Marathon Digital (MARA) Bitcoin Mine (Hood County, Texas):** A cryptocurrency mining facility embedded close to a rural residential community. - **The Permian Basin (West Texas):** The epicenter of extensive oil and gas drilling operations. - **National Ambient Average:** A randomized aggregate of typical everyday American infrasound baselines collected inside hotel rooms, private residences, restaurants, and standard wilderness areas. - **Death Valley (California):** An extreme negative baseline located deep in the desert, hundreds of miles away from any major industrial infrastructure.

[04:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=249s) Colossus and xAI: Environmental Fallout in Memphis

Jordan travels to South Memphis to investigate Colossus, the facility constructed by Elon Musk's xAI to train the large language model Grok. Jordan notes the grim timing of his visit: while standing across the street, news was breaking that Grok's 550,000 GPUs had been generating non-consensual deepfake imagery, including over 23,000 depictions of child sexual abuse material. Running this scale of compute requires immense power: the entire city of Memphis has an electrical capacity of roughly 3,500 megawatts, of which Colossus demands 13%. However, the local grid could only supply around 5%, leading the facility to circumvent constraints by burning methane gas in unpermitted mobile turbines on site.

To assess the site, Jordan flew a drone overhead to capture visual evidence for journalists and environmental regulators, documenting rows of methane turbines venting emissions straight into the surrounding community. The turbines exhaust carbon monoxide, sulfur oxides, volatile organic compounds, and dangerous levels of nitrogen oxides (NOx)—pollutants with well-documented links to severe cardiovascular and respiratory diseases. Jordan contrasts the leniency shown in Tennessee with international standards, citing India’s 2019 fine against Volkswagen of approximately $71 million USD for falsifying NOx emissions. In this predominantly Black neighborhood in South Memphis, these emissions are released directly into residential air, creating a foul stench resembling a mixture of natural gas, chemical runoff, and manure that makes it physically difficult to breathe even half a mile away. Musk's planned expansion, dubbed "MacroHard," aims to expand power consumption to 1.5 gigawatts—representing half of the entire peak grid capacity of Memphis—subsidized by the local population whose drinking water table is simultaneously depleted by industrial cooling towers venting steam into the atmosphere. Confronting the absurdity of the infrastructure, Jordan remarks: “Elon Musk, never wanting to turn down that cringe knob, has named the new expansion to the facility 'MacroHard.'”

- Colossus power footprint: Draws 5% of Memphis's 3,500 MW capacity from the grid, with planned expansions scaling to 1.5 GW (50% of the city's peak capacity). - On-site generation: Unauthorized methane gas turbines discharging hazardous levels of nitrogen oxide (NOx), sulfur oxides, and volatile organic compounds (VOCs). - Local environmental impact: Air pollution causing immediate respiratory irritation half a mile away, alongside millions of gallons of water evaporated into steam instead of returning to the local aquifer.

[07:15](https://www.youtube.com/watch?v=_bP80DEAbuo&t=435s) Acoustic and Infrasonic Signatures of Colossus

Setting up audio gear roughly 300 feet from the main Colossus structures, Jordan captured the audible and sub-audible acoustic spectrum to demonstrate the physical assault of the installation. Using standard spectral visualization, he highlights that audible mechanical whines and industrial clatter hovered around -30 dB, with low mechanical rumbles reaching -28 dB. The inaudible infrasound beneath 20 Hz, however, registered an astonishing 10 decibels louder than the audible noise, packing tremendous acoustic energy that went entirely undetected by ear.

Jordan moved his equipment roughly a mile away to a public park adjacent to a local elementary school to evaluate how the sound propagates through a neighborhood. Higher-frequency audio attenuates over distance, but low-frequency rumbles and sub-audible infrasound carry through the atmosphere and physical ground unhindered. The amplitude of the sub-20 Hz oscillations was strong enough to push his microphone preamps close to clipping despite substantial headroom, generating energy capable of shaking the structural framing of surrounding houses. To make the silent phenomenon tangible to the audience, Jordan filtered out every frequency above 20 Hz, isolated the sub-audible recording, and sped up the sample rate by a factor of 16 (four times faster, then four times faster again) to shift the deep industrial infrasound squarely into the human audible range. Assessing the devastating acoustic environment imposed on the community, Jordan observes: “This rumbling is no good and would make your life utterly miserable.”

- Distance measurements: 300 feet directly across the street from the primary data hall, and 1 mile away at an elementary school playground. - Amplitude disparity: Infrasound frequencies below 20 Hz registered roughly 10 dB louder than the audible mechanical noise floor. - Audio demonstration: A 16x speed conversion was required to shift the deep sub-audible building and turbine oscillations into the audible human hearing range.

[09:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=549s) Hood County, Texas and Industry Bias

Shifting from Tennessee to Texas, Jordan examines the Marathon Digital (MARA) Bitcoin mining facility situated in Hood County. When evaluating public pollution complaints per capita, Jordan explains that researchers must account for regional cultural and industrial bias. An oil refinery erected in environmentally conscious Berkeley, California, will inevitably register far more air quality complaints per 100,000 residents than the identical facility placed in Oklahoma City.

Hood County sits directly atop the fossil-fuel-rich Barnett Shale, where local identity and employment are deeply entwined with heavy energy extraction. The county boasts 174 active oil and gas companies, and its voting demographic is solidly conservative, making it culturally resistant to environmental hand-wringing. Criticizing extraction infrastructure in a local establishment, Jordan jokes, would likely earn a visitor a trip to the nearest emergency room. The fact that residents in this industrial, pro-energy community began filing vocal complaints about physical suffering from the crypto mine signaled that the acoustic pollution was exceptionally severe. As Jordan reflects on the local sentiment: “There are currently 174 oil and gas companies with operations in Hood County alone, and I'd imagine that criticizing this industry in one of the local bars might be a good way to get a free tour of one of the local emergency health care” facilities.

[10:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=600s) Investigating the Granbury Bitcoin Mine

Benn Jordan turns his investigation toward Granbury, Texas, examining the severe impact of a local Bitcoin mining facility (operated by Marathon Digital Holdings, or MARA). While sustained audible noise from the mine has been characterized by residents as a 70 to 90 decibel jet engine-like roar—prompting the company to construct sound-dampening walls that did little to solve the issue—Jordan realized that the reported physiological reactions did not match typical audible noise pollution. Instead, complaints of debilitating health problems in humans, alongside adverse effects in local plants and animals, strongly correlated with known symptoms of low-frequency infrasound exposure.

Jordan initially attempted to remain detached, noting, “I didn't want pesky things like empathy or local politics clouding my research.” However, the pervasive nature of the crisis became immediately clear when merely bringing up the subject of noise pollution to a friend at a local gas station spontaneously drew a crowd of patrons who all wanted to share their personal experiences.

```mermaid flowchart TD accTitle: Infrasound Impact on Granbury Community accDescr: Flowchart showing the cause and reported effects of crypto mine infrasound emissions on local residents.

subgraph Source["Pollution Source"] A["MARA Crypto Mine<br>(70-90 dB Audible Noise + Infrasound)"] end

subgraph Symptoms["Reported Physical Impacts"] B["Debilitating Nausea & Dizziness"] C["GI Issues & Ear Ringing"] D["Nervousness, Anxiety & Seizures"] end

subgraph Fallout["Community Consequences"] E["Residents Forced to Relocate"] F["Visiting County Officials Fall Ill"] G["Ongoing Civil Lawsuits & Remote Logging Setup"] end

A --> B A --> C A --> D B & C & D --> E B & C & D --> F B & C & D --> G ```

Recording from a secluded farm roughly three-quarters of a mile away from the mine on two separate occasions, Jordan captured a spectrum of audible drones—which he wryly describes as a “choir of... crypto bullshit.” Dropping into sub-audible frequencies revealed an intense, brutal spike around 30 Hz. Though sitting below the human auditory threshold, the pressure was physically palpable:

- When sped up into the audible spectrum, the sub-audible frequencies sound indistinguishable from sitting right next to an airplane preparing for takeoff. - The severe low-frequency hum completely destroys the tranquil atmosphere of the rural countryside. - Temporary anomalous low-frequency spikes were observed during long ambient captures, though Jordan excluded these temporary peaks from his baseline datasets.

[12:35](https://www.youtube.com/watch?v=_bP80DEAbuo&t=755s) First-Hand Accounts: Cheryl Shadden's Story

Jordan sits down with Cheryl Shadden, a resident living directly across the street from the crypto mine. He describes her as both exceptionally likable and someone who is rightfully furious about the facility's intrusion onto her land. Keeping his infrasound hypotheses secret so as not to bias her answers, Jordan asked her to describe the direct physical symptoms she experiences on her property:

> “Motion sickness, dizziness, nausea, GI side effects. My ears ring and buzz. Now, nervous, anxious, worried.” — *Cheryl Shadden*

These symptoms consistently match the medical footprint of acute infrasound exposure. Shadden explains that neighbors frequently text one another when sudden waves of sickness occur simultaneously, describing instances where working outdoors at 3:00 PM on a Saturday resulted in sudden, horrific nausea to the point of nearly passing out, only to learn neighbors down the road were experiencing identical episodes at the exact same moment.

The health consequences have led to drastic choices in the community: - One neighboring family completely walked away from their home and took out a second mortgage to move south to Somervell County after their six-year-old daughter began experiencing seizures in the summer of 2024; her symptoms disappeared entirely once relocated. - During a meeting at the power plant involving executives from Constellation Energy, the mayor of Glen Rose, and a Somervell County commissioner, the commissioner became actively ill from the crypto mine's noise and can no longer stay in the area for more than an hour without feeling sick. - To build a foundation for legal and regulatory challenges, Jordan provided Shadden with a monitoring unit to record and stream continuous 24/7 infrasound data from her property to a remote server, allowing them to map acoustic levels directly against power plant maintenance shutdowns.

[15:13](https://www.youtube.com/watch?v=_bP80DEAbuo&t=913s) The Permian Basin: Extreme Industry Benchmark

To establish a baseline comparison, Jordan traveled to the Permian Basin, a sprawling industrial energy hub spanning northwest Texas and eastern New Mexico. The region generates over 6 million barrels of crude oil per day—more than 40% of total U.S. oil production—and accounts for 15% of the nation's natural gas extraction. Infrastructure constraints frequently bottleneck this massive natural gas output, causing operators to vent raw methane or engage in extensive flaring.

Because vast high-pressure water injection from fracking operations has transformed this geologically quiet area into an active earthquake hotspot logging over 2,000 earthquakes per year, Jordan expected it to be the single most infrasonically active location in North America outside of a data center. During his field recording, a nearby wellbore turned on and began emitting an unsettling drone as the drill string plunged in and out of the well. However, the spectral data revealed an unexpected result:

- While infrasound amplitudes in the Permian Basin were loud and closely matched the energy levels of the audible audio spectrum, they were noticeably lower than the infrasound measured around the data centers in Memphis and Granbury. - The infrasound output near the data centers exceeded ambient baseline noise by orders of magnitude so vast that linear percentage comparisons fail, requiring logarithmic regression models to represent accurately. - Noting that presenting mathematical logarithmic regression models would cause YouTube audience retention to plummet to zero, Jordan pivoted toward an empirical psychoacoustic experiment.

[18:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1080s) Human Testing at Buchla and Friends

Determined to safely assess the physical and psychological toll of data center infrasound, Jordan designed a double-blind experiment reproducing the infrasonic waves at a reduced 25% to 30% of their real-world data center amplitudes. The central hurdle was finding a testing environment where human subjects could be exposed to acoustic testing without knowing infrasound was the primary variable.

Jordan reflected on past sonic weapon experiments he had built—such as a speech-jamming acoustic gun that derailed coherent talking—and decided on the ideal testing demographic:

> “Unless your wallet contains a CIA badge, the only place in the world where you could get away with testing sonic weapons on people is a deeply nerdy synthesizer convention.” — *Benn Jordan*

Securing comprehensive liability insurance, Jordan convinced the organizers of the Buchla and Friends synthesizer convention in downtown Los Angeles to let him rent out a booth and a private, enclosed room. His experimental testing framework featured: - Groups of five convention attendees pulled into an insulated room equipped with five chairs, five pairs of studio headphones, and a single oil painting of owls. - Every participant signed liability waivers and was told they were evaluating a three-minute musical piece while closely observing the owl painting and logging their emotional and physical reactions. - To introduce a randomized bias variable and keep himself entertained, Jordan improvised a different fictional backstory for every group explaining why the owl painting was supposedly cursed or haunted, attributing it variously to post-WWII North Dakota in the 1940s and '50s, 1970s occult lore, late-1980s Soviet collapse artifacts, or early-1990s Japan.

[20:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1200s) The Deceptive Infrasound Experiment and Initial Results

To test how inaudible, low-frequency sound affects human subjects, Benn Jordan orchestrated a deceptive psychoacoustic experiment under the guise of an art-and-music evaluation. Volunteers were brought into an experimental space and asked to focus on examining a painting for ominous details while ambient music played in the background. Unbeknownst to the participants, a specialized, massive subwoofer was concealed in the room, pumping out infrasound in exactly 50% of the sessions. Jordan reveals that the setup was so subtle that even he often could not tell if the infrasound generator was actively running unless he looked for physical evidence, such as water visibly vibrating inside a bottle. After the exposure sessions concluded, participants were required to complete an extensive symptom survey and sworn to secrecy for 48 hours before being debriefed on the true nature of the test. Jordan recounts confessing to them: “Everything that I told you is a lie. Yeah, I got this on eBay for 30 bucks. The music, there's nothing substantial about it whatsoever. 50% of the people who are participating in this study are being subject to a fraction of the infrasound that's coming out of data centers across the US.”

Jordan candidly admits the precarious business reality of funding independent scientific experiments via a YouTube channel: “I don't know why, but I feel like it's important that you know that the existence of this video depended on these survey results. There was really no point in making a video about infrasound levels that don't affect anyone. I've found that this is the big drawback about funding research through a YouTube channel.” While over 100 people participated, rigorous filtering was applied to ensure the integrity of the data. Anyone who skipped a question, indicated suspicion, claimed they consciously felt a physical vibration, or participated in a session where Jordan himself knew whether the infrasound was active was thrown out. This narrowed the field to a clean, double-blind dataset of 74 subjects.

The initial percentage comparisons between the control and infrasound-exposed groups revealed significant physiological and emotional divergences: - Exposed participants were 25% more likely to experience tingling sensations or restless legs, and the few individuals who experienced heart palpitations were entirely in the infrasound group. - The exposed group saw a 12% increase in tiredness, a 33% increase in nausea, and a 150% spike in dizziness. Jordan emphasizes this vestibular disturbance as a direct physical reaction to low-frequency pressure waves and notes it is the most frequent symptom cited by citizens in Granbury, Texas, near crypto-mining operations. - Psychologically, participants subject to infrasound were 11% less likely to feel "creeped out" and 43% less likely to feel spiritual—a drop Jordan theorizes stems from nausea and vestibular disorientation distracting subjects from transcendent or reflective thoughts. - Infrasound exposure generated a 20% increase in chills, a 33% rise in irritability or depression, a 55% surge in anxiety, and a 100% increase (twice as likely) in feelings of sadness. - Subjects exposed to infrasound were three times as likely to report eye irritation, aligning with the historical "ghost in the machine" research. - While an 80% increase was noted for "lethargy," Jordan chose to strike the metric entirely from the study, explaining that its distinction from the 12% "tiredness" increase was too semantically ambiguous for participants. - Most glaringly, the infrasound group was three times as likely to report a generalized sense of discomfort.

[23:16](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1396s) YouTube Data vs. Researcher Data

Jordan steps back to unpack the tension between communicating science to a broad internet audience versus presenting it with academic rigor. In internet media, catchy headline statistics like "blah blah is eight times more likely to blah blah" are effective hooks that get casual viewers down educational rabbit holes, but relative percentage increases can easily distort real-world magnitude. For instance, while infrasound subjects were statistically 33% more likely to report nausea, a casual viewer might visualize 33 out of 100 people becoming physically sick. In reality, the average self-reported nausea score on a scale from 0 to 10 rose from a baseline of 0.9 in the control group to just 1.2 in the infrasound group—an increase that is mathematically true, but so low in absolute terms that it barely warrants serious clinical concern on its own.

```mermaid flowchart TD accTitle: Raw Survey Scores Comparison accDescr: Bar chart representation comparing control versus infrasound scores across key metrics on a 0 to 10 scale.

subgraph Discomfort ["Discomfort Score (0-10)"] D_Ctrl["Control: 1.2"] D_Inf["Infrasound: 4.8"] D_Ctrl -.->|"4x absolute jump"| D_Inf end

subgraph Dizziness ["Dizziness Score (0-10)"] Z_Ctrl["Control: 1.2"] Z_Inf["Infrasound: 3.0"] Z_Ctrl -.->|"2.5x absolute jump"| Z_Inf end

subgraph Spirituality ["Spiritual Feeling (0-10)"] S_Ctrl["Control: 4.2"] S_Inf["Infrasound: 2.4"] S_Ctrl -.->|"Significant drop"| S_Inf end

subgraph Nausea ["Nausea Score (0-10)"] N_Ctrl["Control: 0.9"] N_Inf["Infrasound: 1.2"] N_Ctrl -.->|"Minor shift (33% relative)"| N_Inf end ```

When evaluated as raw, absolute scores rather than relative percentages, only a select cluster of metrics exhibited profound shifts. The score for feeling spiritual dropped from an average of 4.2 among control subjects to 2.4 under infrasound. Dizziness jumped from a baseline score of 1.2 up to an average of 3.0. The most dramatic divergence occurred in the measurement of overall physical and mental discomfort: control subjects averaged a negligible 1.2, whereas participants in the infrasound group surged to an average score of 4.8. Reflecting on this unmistakable baseline shift, Jordan concludes: “When reviewing that data, it's hard for even the most skeptical person to come to any other conclusion then. The presence of elevated levels of infrasound is likely to lower your quality of life.” Jordan emphasizes that while an hour could easily be spent picking apart flaws in the pilot study, having 74 vetted subjects provides a strong foundation to petition for expanded research funding and test multiple distinct frequencies with greater repeatability.

[25:42](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1542s) Measurement Techniques, Regulatory Failure, and Community Action

Capturing infrasound accurately requires specialized hardware that goes far beyond standard acoustic microphones. Jordan outlines the three distinct capture tools he utilizes in the field: 1. **Specialized Measurement Microphones:** Convenient to use because they handle like traditional microphones, offering a flat frequency response down to 9 Hz and capturing signals down to 3 Hz before low-end roll-off occurs. 2. **Raspberry Shake 3D:** A high-speed, relatively affordable vibration detection system capable of monitoring full-spectrum, scientific-grade infrasound below 3 Hz without frequency fall-off. Although it demands significant initial calibration, it has the distinct advantage of recording seismic and acoustic phenomena continuously and indefinitely. 3. **Lithium Battery Vibration Sensors / Accelerometers:** Highly accurate, industrial-grade sensors traditionally used to detect minute structural abnormalities and prevent catastrophic failure in large lithium-ion battery banks. While extraordinarily accurate, they are significantly more expensive.

Because vibration sensors record raw seismic and acceleration units, the data must be mathematically converted into Pascals before it can be analyzed or auditioned like an audio file. Jordan notes that after years of painstakingly calculating these conversions manually inside Microsoft Excel, he wrote an automated Python script capable of batch-processing hundreds of sensor logs simultaneously, which he has released publicly on his GitHub repository.

The ultimate impetus behind Jordan’s investigation is the ceaseless flood of messages he receives from people worldwide suffering from mysterious structural vibrations, low-frequency hums, and unexplained health declines. He argues that modern environmental zoning is fundamentally failing because regulators, architects, and municipal bodies treat noise pollution as a nuisance rather than an ecological and public health hazard: “Prior to this experiment, I feel like we have seen more than enough data to make infrasound monitoring and regulation just as important as air or water quality, especially when you look at the studies pertaining to the cardiovascular and nervous system.”

With over 3,000 new industrial data centers currently planned or under active construction across the United States alone, Jordan cautions against waiting for sluggish regulatory agencies: “And while I suspect that having a universally recognized safety threshold for infrasound might be an overly optimistic goal with a government that can't even enforce their own rules with Elon Musk's methane turbines, it can go a long way in both existing and future lawsuits that seek remediation from problematic noise issues caused by data centers and infrastructure that is actively decreasing both the quality of life and health for those with the misfortune of living near them.”

To counter the standard legal defense used by data center operators—who routinely claim that sub-audible hums and ground vibrations were pre-existing environmental background noise—Jordan urges residents facing nearby development to take immediate action: “If you have a data center being built nearby your residents and the infrasound research in this video scares the shit out of you, now would be a great time to set up something like a raspberry shake and start storing a daily log of seismic and infrasonic activity.” Establishing an immutable baseline log prior to facility operation leaves data center operators and their legal teams with zero plausible deniability when class-action attorneys step in.

## Notable quotes > [25:42](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1542s) Prior to this experiment, I feel like we have seen more than enough data to make infrasound monitoring and regulation just as important as air or water quality, especially when you look at the studies pertaining to the cardiovascular and nervous system. — Benn Jordan

> [23:16](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1396s) The presence of elevated levels of infrasound is likely to lower your quality of life. — Benn Jordan

## Mentioned Benn Jordan, Cheryl Shadden, Elon Musk, xAI, Colossus, Grok, MacroHard, Marathon Digital Holdings (MARA), Granbury, Hood County, Barnett Shale, Somervell County, Glen Rose, Constellation Energy, Permian Basin, Death Valley, Buchla and Friends, Raspberry Shake 3D, Volkswagen, Memphis, Tennessee, Texas, YouTube, GitHub, Microsoft Excel, Python

## Video description Join me in another depressing adventure in science and technology. [Research paper from the experiment is a WIP]

💗 Support this channel and join an amazing community: http://www.patreon.com/bennjordan

👀 Stalk me on social media for more frequent updates: https://linktr.ee/BennJordan

🔴 Subscribe To My Streaming Channel. I stream weekly! https://www.youtube.com/watch?v=geaIEWGhgsI

⚡Those lovely custom acoustic panels in the background: https://psyacoustics.com

Timestamps: Intro - 0:00 Infrasound - 2:14 xAI/Colossus - 4:09 MARA Crypto Mine - 9:09 The Permian Basin - 15:13 Experimenting On Humans - 18:00 YouTube Data vs Research Data - 23:16 Recording Techniques - 25:42

[Source](https://www.youtube.com/watch?v=_bP80DEAbuo)

中文简报

测试显示xAI与加密数据中心排放次声波引发疾病 高密度计算设施排放的低于20 Hz的声压波可直接引发人体前庭系统和心血管不适，同时逃避了常规噪声监管。 音乐家兼研究员 Benn Jordan 调查了大型计算基础设施排放的人耳不可听次声波对人体造成的生理伤害，他实地走访了埃隆·马斯克位于孟菲斯的 xAI Colossus 数据中心、得克萨斯州格兰伯里的比特币矿场以及二叠纪盆地。Jordan 发现低于 20 Hz 的低频振动可能明显比可听机械噪声更强，并将这些排放与附近居民报告的恶心、头晕、心血管压力和癫痫等现实健康危机联系起来。为验证这些生理影响，他开展了一项欺骗性双盲心理声学试验，让 100 多名在不知情下的受试者暴露在模拟数据中心次声波环境中，记录到头晕和整体不适感出现显著且具有统计学意义的激增。针对监管盲区以及数以千计在建的新设施，Jordan 呼吁易受影响的社区进行持续的基线地震与次声记录，为未来的法律抗争积累证据链。 实地测量显示，在 xAI 的 Colossus 和 Marathon Digital 加密矿场等设施周边，低于 20 Hz 的次声波排放比可听机械噪声高出多达 10 分贝，且能在地面和建筑结构中毫无衰减地传播。 (7:15) 一项针对 74 名通过审核的受试者开展的双盲心理声学研究表明，在较低振幅的数据中心次声波暴露下，头晕评分（满分10分）从 1.2 跃升至 3.0，总体不适感则翻了两番，从 1.2 升至 4.8。 (23:16) 工业级数据中心的运营带来了严重的局部环境威胁，包括南孟菲斯未经许可的现场甲烷涡轮机排放有害的氮氧化物，以及得克萨斯州格兰伯里持续 70 至 90 分贝的噪声与次可听压力的共同作用。 (4:09) 鉴于全美计划新建超过 3000 座工业数据中心，建议受影响社区部署如 Raspberry Shake 3D 等持续低频监测工具，以建立用于监管和法律救济的不可篡改的基线数据。 (25:42)

Reading summary

Investigator Benn Jordan reports that large-scale computing facilities and crypto mines emit intense, inaudible infrasound that causes measurable physiological distress and long-term health risks for neighboring residents. Based on field measurements and a double-blind human exposure trial, Jordan argues that modern noise regulations fail to account for sub-audible acoustic pressure. Biological Mechanisms of Infrasound Exposure (0:00) Acoustic oscillations below 20 Hz operate beneath conscious human hearing but induce direct, measurable pathologies rather than psychosomatic reactions. Infrasound triggers acute cortisol spikes, vestibular disturbances resulting in vertigo and nausea, cardiovascular strain, and high-frequency hearing loss. Chronic exposure is clinically linked to vibroacoustic disease, in which the abnormal proliferation of extracellular matrices thickens blood vessel walls and permanently restricts cerebral blood flow. Field Signatures Across Industrial Installations (4:09) Acoustic surveys across several North American sites revealed that data centers produce infrasonic amplitudes far exceeding ambient baselines and heavy extraction infrastructure. At xAI's Colossus facility in Memphis, sub-20 Hz frequencies registered roughly 10 dB louder than audible mechanical noise at 300 feet and maintained enough energy one mile away to push microphone preamps near clipping. The facility also relied on unauthorized on-site methane turbines emitting dangerous levels of nitrogen oxides, sulfur oxides, and volatile organic compounds into an adjacent residential area. At the Marathon Digital Bitcoin mine in Granbury, Texas, acoustic monitoring detected an intense low-frequency spike around 30 Hz. Residents reported severe motion sickness, ear ringing, and sudden concurrent nausea, prompting one family to abandon their home after their child developed seizures that ceased after relocation. Infrasonic energy measured near these computing facilities surpassed even the heavy industrial baseline of the Permian Basin, where fracking operations drive over 2,000 earthquakes annually. Double-Blind Human Exposure Testing (20:00) To evaluate the physiological effects in a controlled setting, Jordan conducted a deceptive double-blind trial on 74 vetted participants at an Los Angeles synthesizer convention. Subjects were exposed to hidden subwoofers reproducing data center infrasound at only 25% to 30% of real-world amplitudes under the guise of evaluating art and music. Participants were unaware that low-frequency acoustic pressure was being tested. While relative percentage increases included a 150% rise in dizziness, a 55% jump in anxiety, and a 33% increase in nausea, Jordan emphasized that raw score shifts provide clearer clinical context. On a 0-to-10 self-reported scale, average dizziness rose from 1.2 in controls to 3.0 in exposed subjects, while feelings of spirituality fell from 4.2 to 2.4. Most notably, the raw metric for general physical and mental discomfort jumped from 1.2 among controls to 4.8 in the infrasound group. Regulatory Blind Spots and Baseline Logging (25:42) Municipal zoning frameworks and regulatory bodies currently classify industrial noise as an aesthetic nuisance rather than an environmental health hazard, leaving infrasound entirely unmonitored. Capturing these waves requires specialized hardware, including measurement microphones calibrated down to 3 Hz, continuous Raspberry Shake 3D vibration monitors, or lithium battery bank accelerometers, with raw seismic acceleration converted into Pascals. With more than 3,000 industrial data centers planned or under construction across the United States, Jordan warns that official safety thresholds will lag behind commercial expansion. Because facility operators frequently defend against lawsuits by alleging that low-frequency rumbles were pre-existing background noise, Jordan advises residents near planned sites to deploy continuous sensors and log baseline acoustic and seismic data before operations begin. Jordan acknowledges that his 74-subject study serves as a preliminary pilot requiring academic replication across distinct frequencies. However, he contends the recorded data demonstrates that unchecked infrasound degrades quality of life, urging citizens to establish preemptive baseline logs as legal evidence against industrial operators.

精炼阅读摘要

调查员 Benn Jordan 指出，大型数据中心与加密货币矿场产生的 20 Hz 以下次声波并非单纯的听觉滋扰，而是会诱发头晕、恶心与心血管压力的隐形病理污染源；随着美国各地机房激增，现行环保监管与声学标准正面临严重缺位。 次声波的隐性病理与生理危害 (0:00) 次声波指频率低于 20 Hz、处于人类听觉阈值之下的声学压力振荡。Jordan 强调，高强度次声波暴露所引起的反应并非心理暗示或焦虑催生的假想症状，而是具有明确生理病理基础的客观损伤。在短期内，此类低频波会刺激皮质醇激增导致高血压，扰乱前庭系统引发眩晕与恶心，并带来心血管负荷加剧、呼吸急促及高频听力损失。从长期病理机制看，长期暴露与振动声学疾病（vibroacoustic disease）直接相关，该病症会导致细胞外基质异常增生、血管壁增厚并永久性降低脑血流量，即使在声源停止后影响依然持续。 大型算力基建的声学与环境实测 (4:09) 为了对比真实环境数据，Jordan 调查了田纳西州孟菲斯 xAI 的 Colossus 数据中心与德州胡德县的 Marathon Digital（MARA）比特币矿场。在配置 55 万块 GPU 的 Colossus 设施外约 300 英尺处测量显示，20 Hz 以下不可见次声波振幅比听得见的中高频机械噪音高出约 10 分贝；即便移至一英里外的小学，低频能量依然能穿透空气与建筑结构。由于电力缺口，该工厂私自燃烧甲烷轮机供电，向周边社区排放一氧化碳、氮氧化物等有害废气并大量蒸发地下水。在德州胡德县，MARA 矿场除了发出 70 至 90 分贝的可闻轰鸣外，30 Hz 附近的低频压力峰值同样显著。周边居民长期遭受严重恶心、眩晕、耳鸣与肠胃失调，甚至出现儿童癫痫发作导致家庭被迫举债搬迁的情况。在重工业基准地二叠纪盆地（Permian Basin），尽管拥有 174 家油气企业且水力压裂引发频繁地震，其实测次声波能量却显著低于上述两处算力中心。 双盲声学实验与人身反应量化 (20:00) 为验证次声波对人体的实际影响，Jordan 在洛杉矶举办的 Buchla and Friends 模块合成器展会上开展了一场双盲实验。测试在一间挂有猫头鹰油画的密闭隔音室内进行，100 多名签署免责声明的受试者被告知在背景音乐下观察画作并记录身心感受，而隐蔽的低音炮仅在 50% 的场次中以实际机房强度 25% 至 30% 的振幅播放次声波。剔除意识到震动或存疑的样本后，研究留存了 74 名合格受试者的纯净数据。结果显示，暴露组出现头晕的概率高出 150%，恶心高出 33%，焦虑感上升 55%，眼部刺激报告高出三倍；在 0 至 10 分的绝对评分上，整体身心不适感均值从对照组的 1.2 跃升至次声波组的 4.8，头晕均值从 1.2 升至 3.0，而精神愉悦感则从 4.2 降至 2.4。Jordan 坦承这种网络自主科研的局限，但他强调即便在较低能量下，数据也证明次声波显著降低了生活质量。 监管盲区与社区自救监测策略 (25:42) Jordan 指出，传统市政声学规划将次声波仅当作噪音滋扰而非公共卫生危害，监管政策在技术与执行层面均严重滞后。在全美规划或在建的工业级数据中心超过 3,000 座的背景下，机房运营方通常会在诉讼中辩称低频震动属于既有的环境背景噪音。为打破运营方的抗辩，Jordan 建议居民在数据中心破土动工前，利用特制测量麦克风、锂电池振动加速度计或性价比适中的 Raspberry Shake 3D 设备提前布设 24 小时连续监测节点。通过将地震与震动原始信号转换为帕斯卡声压单位并留存防篡改日志，社区居民能够在集体诉讼中建立先于机房投产的声学基线，为环境维权与司法索赔锁定铁证。 该项目指出了当前 AI 与加密基建扩张背后被严重低估的次声波公共卫生威胁。虽然 Jordan 主导的 74 人双盲实验为探索性试点，尚需专业科研机构争取更多资金来针对特定单一频率进行大规模重复验证，但其测量方案为受算力中心困扰的社区提供了一套技术取证范式。

Detailed notes — zh

## Background 该视频由音乐制作人兼科技研究者 Benn Jordan 于 2026 年 2 月 18 日发布在 YouTube 频道上。简介中他将这段内容形容为又一场“令人沮丧的科技与科学探险”，并表示相关的实验研究论文仍在撰写完善中。 ## TL;DR Benn Jordan 针对全美各地数据中心急剧扩张所引发的居民健康危机展开了深入的实地调查，重点揭露了低频次声波（Infrasound）对人体造成的潜伏性损伤。他实地走访了马斯克的 xAI Colossus 孟菲斯算力中心以及得克萨斯州的加密货币矿场，通过专业声学仪器测得远超环境底噪的次声波能量，并记录了周边居民从恶心、耳鸣到儿童癫痫发作等惨痛遭遇。为了验证因果关系，他借洛杉矶模块合成器展会之机开展了一场伪装成“凝视闹鬼油画”的双盲对照人体实验，证实次声波会直接诱发头晕、焦虑与显著的躯体不适。最后，他开源了自研的数据处理脚本并公开监测方案，呼吁将次声波纳入法定环境监管，并敦促社区居民提前记录声学基线以应对未来的法律诉讼。 ## Atmosphere & style 全片弥漫着一种兼具硬核极客探究精神与冷峻批判意味的氛围，既有对现代科技共生矛盾的辛辣自嘲，也有面对受害居民时的深切同情。Benn Jordan 的叙述风格沉着严密且不失幽默，穿插着无人机夜探工业烟囱、荒野录音对比以及展会上荒诞却周密的“闹鬼画作”心理实验，叙事节奏张弛有度。观看时能清晰感受到一种从隐秘未知的生理不适逐渐剥茧抽丝、最终直面科技巨头庞大机器轰鸣时的压抑与震撼。 ## Bulletin - [00:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=0s) Benn Jordan 针对全美多处数据中心展开实地声学调查，发现其释放的低于 20 Hz 次声波振幅远超常规可听噪声，甚至足以诱发严重前庭失衡与血管壁增厚等振动声学疾病。 - [04:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=249s) 马斯克旗下位于孟菲斯的 xAI Colossus 算力集群因自建数十台甲烷燃气轮机违规自发自用，不仅造成极难呼吸的毒气排放，其不可见的次声波振幅比可听工业噪声还要高出 10 dB。 - [15:13](https://www.youtube.com/watch?v=_bP80DEAbuo&t=913s) 油气重工业区帕米亚盆地虽年均发生 2000 多次人造地震，但其水力压裂现场的次声波恶劣程度依然远低于孟菲斯与胡德县的两座数据中心。 - [20:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1200s) 在洛杉矶合成器展上进行的 74 人严密双盲次声波测试证实，暴露于模拟数据中心次声波下的受试者，头晕发生率暴增 150%，整体躯体不适感评分飙升至 3 倍。 - [25:42](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1542s) 全美目前有超过 3,000 座新建数据中心正在规划或建设，作者呼吁周边居民立即部署 Raspberry Shake 等设备记录背景基线，以防科技巨头在集体诉讼中推卸责任。 ## Key takeaways - [00:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=0s) 数据中心产生的人耳不可见的次声波（低于 20 Hz）会直接导致皮质醇激增，长期暴露可引发振动声学疾病，导致血管壁异常增厚及脑部血流持久下降。 - [04:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=249s) 孟菲斯 xAI Colossus 数据中心规划耗电占全市电网 13%，现场动用多台未完全合规的甲烷涡轮发电机自备供电，排放大量刺鼻废气并诱发附近居民呼吸困难。 - [04:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=249s) 在 Colossus 数据中心一英里外的居民区实测显示，高频噪声衰减明显，但次声波能量几乎无阻拦穿透并剧烈震动住宅框架，振幅比可闻工业噪声高出整整 10 dB。 - [09:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=549s) 得克萨斯州胡德县拥有 174 家油气企业，当地居民极度习惯重工业作业，但 MARA 比特币矿场数据中心激发的异常低频与次声波仍彻底冲破了居民生理承受极限。 - [12:35](https://www.youtube.com/watch?v=_bP80DEAbuo&t=755s) 比特币矿场竖立的实体声屏障无法阻挡穿透力极强的低频及次声波，周边农场动植物出现病变，矿场对门女童甚至因持续低频震荡诱发频繁癫痫被迫举家搬迁。 - [15:13](https://www.youtube.com/watch?v=_bP80DEAbuo&t=913s) 油气水力压裂核心区帕米亚盆地的次声波振幅并未与其全频段脱节，而数据中心泄漏的纯次声波能量高到只能用对数回归描述，两者环境危害性质截然不同。 - [18:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1080s) 为绕过声学武器测试的法规红线并规避心理偏置，作者在 Buchla 合成器展会上借用“闹鬼油画”作为心理掩护，对参会者完成了模拟数据中心次声波的人体双盲实验。 - [20:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1200s) 双盲实验清洗出的 74 份干净数据显示，次声波暴露使受试者头晕概率暴增 150%，眼睛受刺激上升至 3 倍，躯体整体不适感从 1.2 分飙升至 4.8 分。 - [23:16](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1396s) 在科学传播中必须警惕自媒体常用的“相对百分比”增幅陷阱，应当结合绝对量化评分来筛选决定性生理表征，实验确凿证明次声波显著降低了生活质量。 - [25:42](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1542s) 捕捉次声波可选用专用测量麦克风、Raspberry Shake 3D 或工业级加速度计，作者开发了开源 Python 程序用于将原始振动地动数据批量换算为声压帕斯卡单位。 - [25:42](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1542s) 面对全美 3,000 座规划中的数据中心，居民必须在施工前提前布设地动和次声波传感器建立基线日志，以击破科技巨头声称次声波为原有环境底噪的免责辩护。 ## Detailed walkthrough ### [00:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=0s) 次声波的隐秘威胁与全国调查计划 Benn Jordan 开门见山地指出了公众面对现代科技时不可避免的荒谬感与伪善处境：推荐观看其视频的受众通常对数据中心（Data Center）毫无好感甚至极度厌恶，然而大家却正是通过数据中心串流观看着这段视频，而他自己也是在揭露数据中心危害健康的新证据后，把成片上传回数据中心。除非有人打算收拾行囊重返原始的狩猎采集生活，否则在可预见的未来，这种与数据中心的共生矛盾根本无法回避。 随着各地数据中心爆发式扩张，Benn Jordan 开始注意到周边居民出现了一系列神秘而严重的健康异样。在走访中，有居民心碎地提及“他们的女儿只有六岁，却开始频繁癫痫发作（seizures）”，甚至有人为了逃离这种环境，不得不舍弃居所并背负第二套房贷。为了厘清事实真相，他决定展开一场横跨全美的实地声学调查，重点监测数据中心产生的次声波（Infrasound）污染。团队不仅携带专业声学仪器赴多地采集原始低频震荡，还设计了一项双盲实验，让超过 100 名不知情的受试者暴露在相同频率的次声波环境下，以评估身体反应。 针对数据中心的建设现状，Benn Jordan 明确指出： > “To put it simply, regardless of where your ethical, environmental, political, or economic interests are, if a data center is being built nearby your home, you're generally kind of f***ed.” - **次声波的生理危害**：次声波指频率低于 20 Hz、超出人类可听范围的声音。暴露在高强度次声波下会导致皮质醇（压力与高血压相关荷尔蒙）激增，诱发前庭平衡障碍、眩晕、恶心与呼吸急促；长远来看还会带来高频听力损失、焦虑、抑郁，并加重心脏泵血负担。 - **振动声学疾病（Vibroacoustic Disease）**：该病理机制表明，次声波能促使细胞外基质异常增生，致使血管壁增厚。即便脱离次声波暴露环境，脑部血流量依然会长期受到抑制，这证明相关症状绝非心理暗示或焦虑症自显，而是可测量的真实病理损害。 - **五处对比监测选址**： 1. 位于田纳西州孟菲斯的埃隆·马斯克旗下 xAI Colossus 数据中心； 2. 位于得克萨斯州胡德县（Hood County）的 Marathon Digital（MARA）比特币挖矿与数据中心园区； 3. 西得克萨斯州二叠纪盆地（Permian Basin）油气田中心腹地； 4. 遍布全美各地的环境背景次声波随机均值（涵盖酒店、民宅、餐馆至荒野）； 5. 远离一切人造设施与大型基础设施的加州死亡谷（Death Valley）。 ```mermaid flowchart TD accTitle: 数据中心次声波病理影响与实地采样路径 accDescr: 展示次声波对人体生理健康的直接损伤机制，以及全美五个对比采样监测点。 subgraph Hazards ["次声波生理损害机制 (< 20 Hz)"] H1["皮质醇激增 & 血压升高"] H2["前庭失衡、眩晕与恶心"] H3["振动声学疾病: 血管壁增厚 & 脑供血下降"] H4["心脏负荷增加 & 呼吸困难"] end subgraph Sites ["全美五个对比采集点"] S1["孟菲斯: xAI Colossus"] S2["得州胡德县: MARA 比特币矿场"] S3["西得州: 二叠纪盆地油气田"] S4["全美随机背景噪音均值"] S5["内华达/加州: 死亡谷腹地"] end Hazards -.->|"现场测量与双盲验证"| Sites ``` --- ### [04:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=249s) 孟菲斯 Colossus 数据中心的污染现状与频谱真相 首个监测点是位于田纳西州孟菲斯的 xAI Colossus 数据中心。这里运行着埃隆·马斯克用于训练大语言模型 Grok 的庞大算力集群。Benn Jordan 站在街对面拍摄时，该设施正由 55 万块 GPU 维持运转。这一庞然大物的能耗极其惊人：孟菲斯整座城市电网容量仅约 3500 兆瓦（MW），而 Colossus 规划耗电就占了约 13%；眼下该设施仅从当地市政电网抽取约 5% 的电力，其余缺口全靠现场布置的数十台甲烷燃气轮机违规自发自用。 为了看清现场真实布局，Benn Jordan 操纵无人机潜入厂区上方取证。画面清晰捕捉到成排的甲烷涡轮发电机，这些设备直接向空气中排放大量一氧化碳、硫氧化物以及挥发性有机化合物（VOCs）。尤其是高浓度的一氧化氮和二氧化氮，早已被医学证实会直接诱发心血管与呼吸系统重症。Benn Jordan 实地探访时切身感受到了窒息感： > “And I can tell you from having the unfortunate experience of spending an afternoon there, even half a mile away from the facility, it is literally difficult to breathe.” 现场空气充斥着天然气、工业废液与宛如肥料发酵般的刺鼻恶臭；与此同时，大量用于水冷冷却的地下水未回灌水网，而是直接化为白汽蒸发。马斯克还将该园区的新扩建工程命名为充满恶搞意味的“MacroHard”。完全建成后，其耗电量将攀升至 1.5 吉瓦（GW）——占据整座孟菲斯市用电高峰容量的 50%，而巨额电网升级账单却由呼吸着有毒废气的当地居民买单。 现场声学频谱分析彻底打破了人们的直觉： - **可听工业噪声**：在距离主楼 300 英尺的街对面，车辆在麦克风旁大力关门的撞击声作为参照；数据中心冷却系统和电机的工业杂音稳定徘徊在 -30 dB 左右，低频轰鸣的峰值约在 -28 dB。 - **次声波的隐匿峰值**：当分析仪下潜至 20 Hz 以下的人耳不可见区域时，次声波的声压振幅竟然比可听噪音还要高出整整 10 dB，几乎突破了录音电平的动态余量（Headroom）。 - **居民区与小学实测**：在一英里外靠近小学的居民区公园，高频噪音衰减较快，但次声波不仅毫无阻拦地穿透空气，其振幅甚至足以剧烈震动整栋住宅的建筑框架。 - **加速试听处理**：为了让人耳能够感知这些隐秘波形，Benn Jordan 在软件中滤除了所有 20 Hz 以上的声音，并将采样率连续两次提高 4 倍（即加速 16 倍播放），将深沉的次声波提升至人耳可听频段，呈现出令人不安且极度沉闷的工业引擎咆哮声。 --- ### [09:09](https://www.youtube.com/watch?v=_bP80DEAbuo&t=549s) 得州胡德县的背景差异与民意基底 在离开孟菲斯后，团队前往第二个监测点——得克萨斯州胡德县的 Marathon Digital（MARA）大型比特币矿场兼数据中心。在切入数据分析前，Benn Jordan 强调了在科学研究中审视“人均污染投诉数据”时必须剔除的样本偏差（Bias）：如果将炼油厂建在加州伯克利这种环保主义者密集的区域，每十万人收到的空气投诉量显然会居高不下；但若将其安置在偏好赛车文化的俄克拉荷马城，民意反馈则完全不同。 得克萨斯州胡德县地处巴尼特页岩（Barnett Shale）能源走廊，整个县域深深植根于化石能源与电力开采工业，其选民政治光谱呈现出极度坚定的保守“深红”属性。 > “There are currently 174 oil and gas companies with operations in Hood County alone, and I'd imagine that criticizing this industry in one of the local bars might be a good way to get a free tour of one of the local emergency health care [facilities]...” 仅在胡德县境内，就有多达 174 家油气企业同时开展作业。当地民众长久以来对工业开发极为包容，甚至视工业噪音与烟囱为繁荣的象征。然而，正是这样一个对重工业耐受度极高、绝不可能无病呻吟的地区，如今却因为比特币矿场与数据中心的噪音爆发了空前强烈的集体健康抗议，这预示着其声学环境的恶化程度早已冲破了人类生理机能所能耐受的极限。 ### [10:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=600s) 比特币矿场的低频轰鸣与小镇居民的集体梦魇 在这类特定地区，当大量居民开始集体投诉数据中心的噪声污染时，绝非空穴来风或夸大其词。Benn Jordan 在过去 18 个月里阅读了大量关于比特币矿场数据中心产生噪声的报道，但现实情况远比普通的“扰民”恶劣得多。现场的噪声污染通常被描述为 70 到 90 分贝、如同喷气机引擎般持续轰鸣的巨响，而大量居民报告的身体症状更是触目惊心。这引发了多起民事诉讼；矿场运营方虽然修建了隔音墙试图吸收声音，但对于周围居民而言几乎没有任何改善。更有甚者，周边大量植物与动物出现的健康恶化，也与矿场的运行高度吻合。长时间暴露在持续的可闻噪声中固然是对生命的摧残，但居民表现出的许多症状已经超出了常规噪声污染的范畴，反而完全契合已知的次声波暴露症状。 为了保持客观中立，Benn 在进入小镇前原本打算避免被同理心或当地政治纷争干扰研究。然而，他在一家加油站仅仅向朋友随口提了一句“噪声污染”，瞬间就引来了一小群人围拢过来；每一个推门进来的新顾客在听到对话后，都忍不住插上一嘴并分享自己的痛苦遭遇。整起事件的背后交织着诉讼、腐败、漠视以及致残性的健康问题。Benn 前往当地距离比特币矿场约四分之三英里的一处民居进行了两次实地录音。在分析初次采样的频谱时，他发现了大量诡异的可闻嗡鸣与持续低频，“就像一首由加密货币狗屁勾当组成的小型合唱团”。 而当他把分析切入更低频段时，真实情况令人震惊。在接近 30 Hz 的位置——一个人类耳朵几乎听不见、纯靠身体肉感感知的频段——震荡极为凶猛。加速播放后，声音听起来就像紧挨着一架正准备起飞的喷气客机，但这些能量其实全部深埋在人类听力阈值之下。两周后他在同一地点采集了较长时间的环境底噪样本。那处房产本是一座与世隔绝、风景秀丽的农场，如今这种轰鸣彻底摧毁了原有的宁静氛围，低频能量的持续存在毋庸置疑。 - **持续噪声级**：现场可闻噪声在 70 至 90 分贝之间，类似喷气发动机在地面持续试车。 - **环境异常**：除人类出现严重反应外，农场动植物也出现了与矿场开机周期同步的病变现象。 - **隔音措施失效**：矿场竖立的实体声屏障完全无法阻隔穿透力极强的低频与次声波。 --- ### [12:35](https://www.youtube.com/watch?v=_bP80DEAbuo&t=755s) 对话受害者 Cheryl Shadden：无法逃离的肉体折磨 Benn Jordan 拜访了住在比特币矿场正对面的居民 Cheryl Shadden。Benn 将她形容为一个既极具亲和力、相处极其愉快，但同时又绝不可轻易招惹的人，而眼下她正对家门口的数据中心出离愤怒。在采访中，Cheryl 逐一列举了自数据中心落户自家对门以来身体承受的痛苦折磨：持续的晕动症、眩晕、恶心、胃肠道紊乱、耳鸣以及耳朵内持续的嗡嗡声，伴随而来的还有长期的神经紧张、焦虑与恐慌。在采访前，Cheryl 以为 Benn 只是在做一期普通的噪声污染专题，Benn 刻意隐瞒了自己对次声波的研究，然而 Cheryl 所陈述的每一个症状，恰好全都是人体过度暴露于次声波环境下的典型临床表现。 Cheryl 描述了邻里之间经常出现的痛苦共鸣：“往往是周六下午三点，我在外面干活，突然恶心到整个人快要昏倒。我进到屋里，路下方的邻居就发来消息说他们也是同样的时间、完全一样的感觉，根本没法在户外工作，简直太可怕了。”当 Benn 询问是否有人被迫搬家时，Cheryl 给出了一个令人心碎的案例：隔壁一家人的六岁小女儿在 2024 年夏天突然开始癫痫发作，父母最终不得不直接放弃这里的房产，背上第二套房贷搬到了南边的萨默维尔县（Somervell County），只为让女儿远离这个地狱；而小女孩在搬离后，癫痫症状便彻底消失了。更有甚者，当地官员在埃森哲/星座能源（Constellation Energy）与格伦罗斯（Glen Rose）市长及县专员开会时，其中一位专员在发电厂开会期间直接被加密矿场的噪声震到发病，只要在该区域停留超过一小时就会呕吐不适。 面对这种不可持续的生存困境，Benn 当场提出为 Cheryl 提供一套 24 小时运行的次声波监测设备，长期隐蔽地将实时数据回传至远程服务器进行记录。这套系统旨在将次声波振幅变化与数据中心停机维护的时间线进行精确对照，为后续的法律诉讼与立法听证提供无懈可击的技术铁证，甚至为其他遭受次声波污染困扰的地区树立诉讼模板。 > “在所有向外界表达观点的陈述中，Benn Jordan 总结了这次行动的最终法律目标：‘That way, further legal and legislative action can be based on clear evidence and maybe even provide an operational legal template for other regions with infrasound pollution.’” --- ### [15:13](https://www.youtube.com/watch?v=_bP80DEAbuo&t=913s) 帕米亚盆地实测：天然气压裂区与数据中心的颠覆性对比 为了探寻人造次声波的极限峰值，Benn 将目光投向了为全美乃至本次视频播放提供海量能源的帕米亚盆地（Permian Basin）。该盆地横跨德克萨斯州西北部和新墨西哥州东部，供应了全美超过 40% 的原油（日产逾 600 万桶原油）以及全美 15% 的天然气。这里的天然气产量极高，以至于外输管道经常陷入运力瓶颈，导致许多开采商只能直接在水力压裂现场进行天然气燃烧排放（flaring），甚至直接将甲烷 venting 排放到大气中。这里本就地广人稀，居民区多为石油工人的临时组合屋。 水力压裂需要向地下注入海量的高压水流，这硬生生将一个远离任何主要断裂带的地质稳定区，改造成了一个每年发生超过 2000 次地震的活跃震中。Benn 原本坚信，这里既然集齐了大型输气管线、石油钻机和人工诱发地震，必然是除数据中心外地球上次声波最活跃的人类工程现场，绝对会诞生他所记录到的最强次声波极值。录音期间，附近的一口钻井启动运行，钻柱在井筒内上下穿梭，发出一阵令人毛骨悚然的声音。然而，仪器监测结果却彻底打破了他的预判：帕米亚盆地的次声波振幅虽然可观，但与其可闻频段的振幅大体处于同一水平。 令人不寒而栗的结论由此诞生：帕米亚盆地的高压油气压裂现场虽然远谈不上理想，但其产生的次声波恶劣程度居然远远比不上孟菲斯（Memphis）的 xAI Colossus 或格兰伯里（Granbury）的数据中心。这两个数据中心周边的次声波水平高出环境底噪太多，以至于用百分比计算都失去了数学意义，必须引入对数回归才能进行有效对比。 ```mermaid flowchart TD accTitle: 帕米亚盆地与数据中心次声波能量及环境对比 accDescr: 对比帕米亚盆地水力压裂油气区与大型数据中心在人口密度、次声波特征及峰值能量上的差异 subgraph Permian["帕米亚盆地 (油气水力压裂区)"] P1["高压注水与钻井作业<br/>年均 2000+ 次人造地震"] P2["极低人口密度<br/>以工人临时宿舍为主"] P3["次声波振幅与全频段相当<br/>未出现异常脱节突起"] end subgraph Datacenters["数据中心 (Memphis xAI / Granbury MARA)"] D1["巨型水冷机组与排风矩阵<br/>持续性纯次声波能量泄漏"] D2["紧邻居民区与农场<br/>社区常住人口密集"] D3["次声波远超环境底噪<br/>能量高到只能用对数回归描述"] end P3 -.->|"峰值远低于"| D3 ``` --- ### [18:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1080s) 洛杉矶 Buchla 展会：用“闹鬼油画”做掩护的人体声学盲测 既然数据中心的次声波能量高到无法用常规算术线性衡量，Benn Jordan 决定退回实验室逻辑：如果将数据中心现场录得的次声波声波，以原振幅或能量的 25% 至 30% 在受控环境中还原，并让不知情的受试者置身其中，人类究竟会产生何种生理与心理反应？受此前研发致使他人无法正常言语的定向声学武器启发，Benn 深知在公共场合拿人类测试声学武器极易触犯法律红线。正如他调侃所言：“除非你的钱包里塞着一张中情局的警徽，否则在这个世界上，你唯一能够合法拿声学武器在人类身上做实验的地方，就只有极度极客的模块合成器展会了。” 全球最硬核极客的合成器聚会，当属由先锋合成器名厂 Buchla 在洛杉矶市中心举办的“Buchla and Friends”。Benn 向主办方提出了一个极为离奇的方案。在购买了巨额责任保险后，主办方爽快地同意租给他一个展位以及一间完全封闭的隔音房。Benn 的团队在展会现场招募参会者，签署知情同意免责书，将其带入房间进行测试。 整个实验装置极其诡异且经过周密设计： 1. **测试环境**：隔音室内摆放着五张椅子、五副耳机，以及正前方一幅描绘着几只猫头鹰的古旧油画。 2. **心理偏置消除（心理诱导）**：为了掩盖真正的测试目标（次声波测试），Benn 对每批由 5 名受试者组成的小组编造了一套完全随机的荒诞故事，声称那幅猫头鹰油画是“受诅咒的闹鬼之物”——一会儿说它源自 90 年代初的日本，一会儿说它来自 70 年代末或 80 年代末苏联解体时期，甚至吹嘘是二战后 40 至 50 年代从北达科他州流出的。 3. **实验流程**：受试者被告知他们将戴上耳机，在封闭房间内一边凝视这幅“闹鬼油画”，一边聆听一段时长为 3 分钟的音频，期间必须密切觉察并记录自己身体与情绪上的所有细微变化。 这种荒诞的故事设定不仅让 Benn 整个周末都沉浸在扮演故事大王的社交挑战乐趣中，更关键的是在科学统计层面打散了受试者的心理偏置变量，使他们误以为自己在参与一场关于视觉暗示与玄学氛围的感知测试，从而将身体对低频/次声波振动的真实生物物理反应纯粹剥离出来。随着一轮轮测试展开，参试者在听完音乐后给出的生理与情绪反馈，展现出了极具启发性且令人不适的真实数据。 > “关于在公开场合拿人做声学测试的严苛限制，Benn Jordan 戏谑地评论道：‘Unless your wallet contains a CIA badge, the only place in the world where you could get away with testing sonic weapons on people is a deeply nerdy synthesizer convention.’” ### [20:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1200s) 双盲次声波人体实验与症状统计 Benn Jordan 详细解释了他对志愿者隐瞒真相的具体实验过程。在实验中，受试者以为自己只是在欣赏一幅画并寻找其中的怪异细节，然而房间内实际安装了一台巨型专用超低音扬声器（Subwoofer），在 50% 的测试场次中会开启次声波（Infrasound）。由于次声波频率极低，连 Jordan 本人如果不借助水瓶中的水面振动，往往也无法肉眼或凭耳朵分辨设备是否处于工作状态。实验结束后，他要求受试者填写体验问卷，并要求他们严格保密 48 小时，随后才向他们坦白真相：“Everything that I told you is a lie.”（我之前告诉你的一切全都是谎话。）他直言那幅画只是自己在 eBay 上花 30 美元买来的，音乐本身没有任何特殊深意，而半数参与者实际上暴露在了模拟全美各地数据中心所释放的次声波环境下。 这项研究最核心的部分是一份症状感受量表，其指标均来自以往真实次声波暴露的相关经典研究。Jordan 坦率承认，通过 YouTube 频道自筹资金进行科学研究存在巨大的现实矛盾：“Finding the truth is both expensive and risky.”（探寻真相既昂贵又充满风险。）如果实验数据显示次声波对人没有任何影响，那么这个耗资巨大的视频项目根本无法成立；幸运的是，实验最终取得了极其显著且有统计价值的成果。在超过 100 名参与者中，Jordan 严格剔除了漏填问卷、对测试产生怀疑、声称察觉到物理振动，或者他本人提前知晓开闭状态的样本，最终获得了来自 74 名受试者的干净双盲数据（Double-blind Data）。 ```mermaid flowchart TD accTitle: 双盲次声波实验流程图 accDescr: 展示受试者筛选、双盲分组测试以及数据清洗的完整实验步骤 subgraph 实验实施阶段 A["招募 100+ 名志愿者"] --> B["虚构艺术/音频测试背景"] B --> C{"双盲随机分组"} C -->|"50% 场次"| D["隐蔽超低音开启次声波"] C -->|"50% 场次"| E["对照组（静音无刺激）"] end subgraph 数据清洗与分析 D & E --> F["填写加权症状感受问卷"] F --> G["48 小时保密协议后告知真相"] G --> H{"数据清洗剔除异常"} H -->|"剔除漏填/察觉振动者"| I["获得 74 份干净双盲有效数据"] end ``` 根据这 74 份严密筛选后的数据，暴露在次声波下的受试者展现出一系列明显的生理与情绪波动： - 出现刺痛感（Tingling sensation）或下肢不安综合征（Restless legs）的概率增加了 25%。 - 疲劳感（Tiredness）增加了 12%，恶心感（Nausea）增加了 33%。 - 头晕（Dizziness）发生率暴增了惊人的 150%。Jordan 强调前庭系统紊乱（Vestibular disturbance）是低频压力波最直接的生理反应，而这恰恰与德克萨斯州格兰伯里（Granbury, Texas）数据中心周边居民最常投诉的健康问题完全吻合。 - 感到发冷（Chills）的概率上升了 20%，烦躁或抑郁（Irritability or depression）上升了 33%。 - 眼睛受到刺激（Eye irritation）的概率飙升至 3 倍，这一现象与关于“机器中的幽灵”（Ghost in the machine）的经典声学研究论文高度吻合。 - 焦虑感（Anxious）增加了 55%，悲伤感（Sadness）翻倍（增加 200%），而最核心的整体不适感（Discomfort）更是飙升至 3 倍。 - 令人意外的负向指标：次声波组感到“毛骨悚然”（Creeped out）的概率降低了 11%，产生“灵性体验”（Spiritual）的概率降低了 43%。Jordan 推测头晕等强烈的躯体不适直接压制了大脑去感受超自然或灵性维度的空间。此外，“嗜睡”（Lethargy）虽然暴增 80%，但因词义与疲劳重叠模糊，被他主动从最终判定指标中剔除。 --- ### [23:16](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1396s) 数据呈现的两面性：自媒体传播 vs 严谨科研标准 Jordan 针对研究数据的沟通方式展开了深入反思。他指出，在他这种跨界工作中存在两种完全不同的数据表达模式：一种是“YouTuber 模式”，追求通俗化与传播力；另一种是“研究人员模式”，强调严谨的绝对量化与可复现性。大众通常正是通过信息流中“某某现象增加了百分之几百”这种耸动标题接触到前沿科学概念的，但百分比的相对增幅很容易掩盖基数本身的微小。 为了向科学共同体负责，Jordan 剖析了实验中的具体绝对评分（量表区间为 0 到 10 分）。例如，恶心感虽然相对增幅达到 33%，但在 0 至 10 分的绝对自评中，次声波组的平均分仅为 1.2 分，对照组为 0.9 分，两者的绝对数值都非常低，在统计学上并不足以单独作为决定性证据。同理，眼睛刺激的绝对得分也很微弱，亟需更大样本量的测试才能得出定论。 然而，在剔除这些边缘指标后，有三项关键指标在绝对分值上拉开了无法忽视的巨大鸿沟： 1. **灵性感受（Spiritual）**：对照组平均分高达 4.2 分，而次声波暴露组直接跌落至 2.4 分。 2. **头晕症状（Dizziness）**：次声波组平均分达到 3.0 分，而对照组仅为 1.2 分（净增幅超过 150%）。 3. **整体躯体不适（Discomfort）**：这是整个实验最具破坏性的核心数据，对照组的平均不适分仅为 1.2 分，而次声波暴露组飙升到了 4.8 分。 面对这组触目惊心的数据对比，Jordan 直言不讳地下了定论：“The presence of elevated levels of infrasound is likely to lower your quality of life.”（高水平次声波的存在极有可能大幅降低你的生活质量。） --- ### [25:42](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1542s) 次声波监测设备、法律维权与基础设施监管倡议 在讨论如何推进该领域的研究时，Jordan 强调当前科学界最缺乏的是“可重复性”（Repeatability）。为此，他公开了自己目前用于捕获次声波的三种不同技术路径，并客观比较了各自的优劣： | 监测设备 / 技术方案 | 优势与频响特性 | 缺点与局限性 | |:--- |:--- |:--- | | **专用测量麦克风** (Measurement Mic) | 平直频响下限可达 9 Hz，极限可拾取至 3 Hz；操作便携，使用逻辑与普通麦克风无异。 | 低于 3 Hz 的频段衰减严重，无法实现全频谱物理级精确测量。 | | **Raspberry Shake 3D 系统** | 价格相对亲民且测量精准；离开声学范畴采用高速振动检测，能够实现无衰减的全频谱次声波长期连续监测。 | 部署过程极其繁琐，需要大量的初始校准与环境搭建。 | | **工业级加速度振动传感器** | 常用于大型锂电池异常检测与安全预警的高精度微型传感器；测量精度达到工业/科学级别。 | 价格极其昂贵（“Not cheap, it is expensive”），难以大规模普及。 | 由于 Raspberry Shake 和工业振动仪输出的是地动或加速度原始信号，若要转换为声学通用的可听或可视压强单位（帕斯卡，Pascals），原本需要在 Excel 中进行极其痛苦的公式换算。Jordan 表示他专门编写了一套开源 Python 程序，可批量处理数百个数据文件，并将代码上传到了个人的 GitHub 仓库。 驱动 Jordan 耗费心血投入这项调查的初衷，是他多年来收到的数千封来自全球各地的求助邮件与留言——无数普通人正在默默遭受类似“大地的嗡嗡声”（The Hum）这种低频噪音的摧残，却求告无门。他严肃指出，全美目前有超过 3,000 座新建数据中心正处于规划或施工阶段，必须将次声波和低频噪声监测提升到与水污染、空气质量同等严格的法定监管高度。 在谈及监管现实时，Jordan 讽刺地指出：“And while I suspect that having a universally recognized safety threshold for infrasound might be an overly optimistic goal with a government that can't even enforce their own rules with Elon Musk's methane turbines, it can go a long way in both existing and future lawsuits that seek remediation from problematic noise issues caused by data centers and infrastructure that is actively decreasing both the quality of life and health for those with the misfortune of living near them.”（尽管我认为寄希望于建立全球公认的次声波安全阈值可能过于乐观——毕竟连针对埃隆·马斯克那批甲烷涡轮机的现行法规，政府都根本无法严格执法——但这些数据对于当前和未来的诉讼而言仍至关重要，能帮助那些不幸居住在数据中心和基础设施附近、身心健康与生活品质正遭受实质侵蚀的受害者寻求赔偿与整治。） 针对家园周边正在或即将开建数据中心的普通居民，Jordan 给出了关键维权建议：立即自行配置 Raspberry Shake 等地震与次声波监测设备，建立起当地环境声学与地动基线的每日日志。因为在面对居民投诉与诉讼时，科技巨头最常用的免责辩护就是声称“无法证明这些次声波在施工前并不存在”。在缺乏政府作为的情况下，先期保存的客观数据将成为集体诉讼律师（Class action attorneys）发起法律反击的决定性武器。最后，Jordan 呼吁大家关注科学同行评审进程，并感谢 Patreon 社群对独立科研项目的资助。 ## Notable quotes > [00:00](https://www.youtube.com/watch?v=_bP80DEAbuo&t=0s) To put it simply, regardless of where your ethical, environmental, political, or economic interests are, if a data center is being built nearby your home, you're generally kind of f***ed. — Benn Jordan > [23:16](https://www.youtube.com/watch?v=_bP80DEAbuo&t=1396s) The presence of elevated levels of infrasound is likely to lower your quality of life. — Benn Jordan ## Mentioned Benn Jordan, Cheryl Shadden, Elon Musk, xAI, Colossus, Grok, MacroHard, Marathon Digital, MARA, Buchla, Buchla and Friends, Raspberry Shake 3D, Constellation Energy, Permian Basin, Hood County, Memphis, Granbury, Somervell County, Death Valley, YouTube, GitHub, Patreon ## Video description Join me in another depressing adventure in science and technology. [Research paper from the experiment is a WIP] 💗 Support this channel and join an amazing community: http://www.patreon.com/bennjordan 👀 Stalk me on social media for more frequent updates: https://linktr.ee/BennJordan 🔴 Subscribe To My Streaming Channel. I stream weekly! https://www.youtube.com/watch?v=geaIEWGhgsI ⚡Those lovely custom acoustic panels in the background: https://psyacoustics.com Timestamps: Intro - 0:00 Infrasound - 2:14 xAI/Colossus - 4:09 MARA Crypto Mine - 9:09 The Permian Basin - 15:13 Experimenting On Humans - 18:00 YouTube Data vs Research Data - 23:16 Recording Techniques - 25:42 [Source](https://www.youtube.com/watch?v=_bP80DEAbuo)

Source quotes

[0:00] Benn Jordan: To put it simply, regardless of where your ethical, environmental, political, or economic interests are, if a data center is being built nearby your home, you're generally kind of f***ed. [23:16] Benn Jordan: The presence of elevated levels of infrasound is likely to lower your quality of life.

Source quotes

[25:42] Benn Jordan: Prior to this experiment, I feel like we have seen more than enough data to make infrasound monitoring and regulation just as important as air or water quality, especially when you look at the studies pertaining to the cardiovascular and nervous system. [23:16] Benn Jordan: The presence of elevated levels of infrasound is likely to lower your quality of life.

Original transcript

[0:00] If you're in the demographic of people who generally get recommended my videos, then chances are that you are not a fan of data centers. You may even hate data centers, yet here you are, watching a video that's being streamed directly from one. And here I am, making a video showing you my research and evidence on covering a completely new way that data centers are harming people's health. And then I'll be uploading this video to a data center. But unless this is what you're watching while you pack a bag to live the rest of your life as a hunter and gatherer, the hypocrisy in our relationship with data centers is unavoidable for the foreseeable future. troubling correlation. Their daughter was six years old. She started having seizures. Walked away from their home here, took out a second mortgage. In this video, we're going to be traveling all across the country collecting infrasound noise pollution recordings from data centers. And then we're going to conduct a double blind research study that exposes over 100 unknowing human participants to those same frequencies to see how it makes them feel. Spoiler alert. The results are terrifying. And along the way, we might as well take an objective, practical, deep dive into how and why the explosion of data centers are so problematic. And we're even going to covertly fly drones over some of the most controversial ones to see how badly they're breaking environmental regulations. Let's go! Data centers are pretty controversial right now, unrelated to the new big issues that I'm going to be talking about in this video. To put it simply, regardless of where your ethical, environmental, political, or economic interests are, if a data center is being built nearby your home, you're generally kind of f***ed. But this video isn't about the power and water consumption or the toxic chemicals or the unsettling economic circumstances. Last year I started hearing about some mysterious and troubling symptoms from people who live nearby readings from around the country, and there was enough correlation for me to dig pretty deep into an area of research that I already had some experience in. Infrasound is any sound below 20 hertz, just below the human hearing range, and I've done a massive video project on it if you're interested in learning more, but for the scope of this video, let's just talk about how it might harm you. Being in the presence of elevated levels of infrasound has been shown to show spikes in cortisol levels, which is the hormone associated with stress and hypertension. It can cause vestibular balance, vertigo, and nausea. Then there's something called vibroacoustic disease, which suggests that infrasound can cause abnormal growth of extracellular matrices, which thickens blood vessels and reduces blood flow to the brain even long after the infrasonic exposure has ended. It's been shown to cause high-frequency hearing loss, shortness of breath, anxiety, depression, and it even makes your heart have to work harder to pump blood. So it's not like this anxiety-based or obsessive compulsive thing where an unpleasant feeling or an unusual situation makes you manifest symptoms. infrasound can tangibly harm your health in a very real and measurable way. And there are potentially millions of people being subjected to these low oscillations of pressure waves and not realizing it or understanding the cause of some of the aforementioned symptoms that they might be experiencing. So the first thing I needed was a whole lot of data. And, no pun intended, data centers were the perfect vector of enormous, unnatural industrial sites frequently surrounded by people reporting symptoms commonly associated with infrasound exposure. I chose five sites to take infrasonic measurements. The first would be the Colossus Data Center in Memphis. The second would be the Marathon Digital Bitcoin Mining and Data Center facility in Hood County, Texas. The third would be right in the middle of the vast oil and gas fields in the Permian Basin in West Texas. The fourth would be a randomized average of ambient infrasound levels all over the country from hotel rooms to houses to restaurants to remote wilderness. And the fifth would be far away from most sources of human-made infrasound, right in the middle of Death Valley, hundreds of miles away from any sort of major infrastructure. Okay, let's measure some infrasound! This is Colossus in Memphis, Tennessee. It's the data center that Elon Musk built to train Grok, his LLM and anti-woke chatbot that has referred to itself as "Mecha-Hitler." It was here, coincidentally, while we were standing across the street, when its 550,000 GPUs were being monitored images of people without their consent. And among those, over 23,000 of them depicted children. But when Grok, and subsequently Colossus, is not generating child abuse material, it's also really, really, really funny. Do you think that all the gold is in Fort Knox? Do I think all the gold is in Fort Knox? Yes. A f***ing conspiracy theorist? Yes. Yes. What are you, a f***ing conspiracy theorist? What do I like? I'm a f***er. [5:00] AI with a penchant for chaos and I'm stuck talking to you Does it give you actual answers to and talk shit? Or is it mostly just talk shit? We're tuning it. As you can imagine, supplying this type of vital resource to society uses a whole lot of electricity. The city of Memphis, Tennessee, can support about 3,500 megawatts and this facility uses about 13 percent of that. But right now it's actually only drawing about five percent from the local power grid and illegally offsetting the rest of it with methane gas turbines. I sent a drone over the Colossus facility to get a better look and then to share the footage with journalists and regulators. These methane turbines that you see emit monoxide sulfur oxides and volatile organic compounds into the air. And the health effects of these chemicals aren't like some sort of exaggerated woke controversy. Compounds like nitrogen oxide have long been proven to directly cause a wide range of respiratory and cardiovascular diseases. They're commonly regulated or even banned in developing countries. For example, in 2019, India fined Volkswagen the equivalent of $71 million USD for hiding nitrogen oxide in their emissions tests. But here, in the predominantly black residential neighborhood Nitrous oxide flows freely. And I can tell you from having the unfortunate experience of spending an afternoon there, even half a mile away from the facility, it is literally difficult to breathe. This place smells so f***ing bad. It's a mixture of, like, natural gas, what I can only assume is chemical runoff, and then, like, I don't know, manure? turned into steam rather than returning to the area's water table. And it's growing rapidly. Elon Musk, never wanting to turn down that cringe knob, has named the new expansion to the facility "MacroHard." But if all goes to plan, the data center will use 1.5 gigawatts of power. That's 50% of Memphis's total peak power capacity. And of course, the cost for this infrastructure will be subsidized by the same people breathing in the nitrogen oxide. I could spend the rest of the day talking about the absurdity of this structure, but I was there for the infrasound, and I found bunch of it. My first data sample comes from just across the street, about 300 feet away from the largest of the colossus buildings. The audible noise pollution was about as bad as the air pollution and scenery was. I'm going to show you what this looks like in a way that's pretty easy to digest and we don't need seismology software or anything like that. I'm amplifying this by a little bit so we could see a little bit more down here in the spectrum, and to give you an idea of how loud the data center is, this is a car door slamming very close to the microphone. of what a human being can hear in the audio spectrum. And so all of these whines and general industrial noises are hanging out around negative 30 dB in this recording. And so like the lowest of rumbles is peaking out at like negative 28. And then if we go down to infrasound, the infrasound that you can't hear is like 10 decibels louder than that. Another sample I took was from a park near an elementary school in a residential area about a mile away. This in particular is a great more than lower frequency or infrasonic ones. This video is playing through a sound system with subwoofers or even if you're wearing some good headphones, you will very quickly realize how frankly uninhabitable this noise makes the entire neighborhood. This rumbling is no good and would make your life utterly miserable. And then this rumble down here is happening at an amplitude that's loud enough to shake the frame of your house. So humans can't hear this, but look how loud that is. It's actually almost peaking the recording, and I gave this microphone a whole lot of headroom. Okay, so I'm gonna delete everything above 20 Hertz here. So all audible frequencies, and now we're only left with the infrasound. Now I'm going to interpret the sample rate four times faster, and I'm gonna make it even four times faster than that. This way all of those infrasonic frequencies will be audible to you. Next up! When looking at data through the lens of research, when reading through hundreds of public reports of types of pollution per capita, you have to consider bias. For example, if you've built an oil refinery near the high number of environmentalists living in Berkeley, California, you'd probably have a lot more air quality complaints per 100,000 people than you would if you built one in Oklahoma City, which is famous for its go-kart races. And from a data perspective, shale, this region is supported by a solid backbone of fossil fuel and energy production, and their voting record is as reliably red as one can imagine. There are currently 174 oil and gas companies with operations in Hood County alone, and I'd imagine that criticizing this industry in one of the local bars might be a good way to get a free tour of one of the local emergency health care [10:00] facilities. Well, it wouldn't be free. So when a whole lot of people in this particular region are complaining about data center noise pollution, you know it's not likely to be hyperbole. Over the last 18 months or so, I have read a lot of articles about a noise coming from a data center that mines bitcoin, but it seemed like it was a lot more than just a nuisance. While the noise pollution is described as a 70 to 90 decibel jet engine-like sound, a large number of reports cite symptoms This turned into multiple civil suits, and the company built walls to try and dampen the sound, but it resolved little to nothing for those living nearby. There are even a number of reports of plants and animals suffering from health effects that are closely correlated with the mine's operations. Sustained audible noise pollution can be a blight on any living thing nearby, but a lot of these symptoms don't line up with the symptoms of just noise pollution exposure. What they do line up with is some of the known symptoms of infrasound exposure. to anyone in the city. I didn't want pesky things like empathy or local politics clouding my research. But merely mentioning noise pollution to my friend in a gas station immediately drew a small crowd, and it was as if every new customer that walked in overheard the conversation and then had something new to contribute to it. And this whole situation is actually a wild series of events that involves lawsuits, corruption, neglect, debilitating health issues, but I went there for one thing, I visited a residence about maybe three quarters of a mile away from the Bitcoin mine on two different occasions, and this is the first. And we could see a whole bunch of weird, audible drones and sustained noises, like a little choir of... crypto bullshit. Now of course let's go down to those lower frequencies and Oh, that's brutal. It's almost as if the worst of it's like right here at 30 hertz, which is too low to really hear. You just feel it. Fed up a whole bunch. It sounds like you're next to a plane that's about to take off, but all these frequencies are actually below your range of hearing. This is from the exact same property about two weeks later, and I got a nice long ambient sample here. Keep in mind that this property is literally a nice secluded farm, and this noise absolutely destroys the atmosphere. I don't really know what's causing these peaks down here, but they're temporary, so I'm not including them in my data. But it's pretty undeniable that this sound here So I met with someone who lives right across the street from the Bitcoin mine, Cheryl Shadden. And I would describe Cheryl as the rare combination of someone who is extremely likable and an absolute pleasure to be around, but also someone who you do not want to piss off, who is actively extremely pissed off about the data center near her land. Like what physical symptoms have you felt just from your property being directly across the street from the data center? Motion sickness, dizziness, nausea, GI side effects. My ears ring and buzz. Now, nervous, anxious, worried. By the way, prior to this interview, Cheryl thought that I was just doing a story on the noise pollution as I was keeping my infrasound research close to my sleeve. Yet every single one of these symptoms is commonly associated with excessive infrasound exposure. Several of us will reach out to one another. So it's three o'clock on a Saturday afternoon. I'm outside, I'm trying to work, I am just horrifically nauseated to the point where I feel like I'm going to pass out. I go in the house, a neighbor messages down the road, same thing, same time. They're trying to do outside work. They can't do it. It's just horrific. Yeah. Do you know of anybody who's like moved? Who's just left? Who's just been like, I can't handle this anymore? Yes, literally one family, their daughter was six years old. She started having seizures in the summer of 24. They basically walked away from their home here, took out a second mortgage, moved, to Somerville County, south of us, to get their daughter away from this. If you know, has her symptoms kind of gotten better since she's left? Right. They went away? Yep. Wow. When we met with the executives at Constellation Energy, the mayor and Glen Rose, and then another commissioner in Somerville County, one commissioner, he... The noise from the crypto mine actually made him ill when he was at the power plant meeting with the executives. He can't be in this area for more than an hour and he starts getting sick. I asked Cheryl if she would be okay with me giving her a device that could record infrasound from her property 24 hours a day over a long period that would be discreetly streaming it to a server that I could log remotely. This would at least allow me to be able to better correlate these [15:00] timeline of infrasound levels with the timeline of the data center being powered down for maintenance. That way, further legal and legislative action can be based on clear evidence and maybe even provide an operational legal template for other regions with infrasound pollution. For many of you watching or listening to me, a whole lot of energy required to host, stream, and play back this video came from right here: the Permian Basin spanning from northwest Texas to eastern New Mexico. This area is where over 40% of the oil produced in the entire United States comes from, and it accounts for over 6 million barrels of crude oil per day. And this is also the extraction point for 15% of America's natural gas. And as you would natural gas output is so high that it gets bottlenecked by the pipelines. Because of this, many of the operators resort to flaring, or just burning the fuel at the fracking site. Many operators also just vent the methane directly into the open air. The good news is that this area is pretty empty by design. It has a very low population density. And when you do find some residential areas near the oil fields, they're typically temporary or modular housing units limited to housing oil workers. I am here because it's quite possibly the most infrasonically active place on that isn't associated with the data center. We already know that major pipelines and oil rigs can cause infrasound, but the massive amounts of high-pressure water injection required for fracking has literally turned this geologically sedentary region, far from any major fault lines, into an active earthquake epicenter with over 2,000 earthquakes per year. So I personally can't think of a better place to get a sample of some man-made infrasound. While I was recording this, a nearby wellbore turned on. *Wooow* *laughs* And then just started playing this creepy song as the drill string went in and out of the well. At least that's what I'm assuming I'm hearing. but the amplitude is comparable to the rest of the stuff in the audio spectrum. I personally expected the Permian Basin to be the winner here, like the loudest infrasound that I had ever recorded, the peak levels that all infrasound polluters would aspire to reach someday. But while the infrasound levels were nowhere close to ideal, they still weren't as bad as they were near the data centers in Memphis or Granbury. The infrasound levels at these two data centers were so much higher than the ambient values that it doesn't even make practical sense percent higher. It makes much more sense to compare these samples using a logarithmic regression. And when you do that, a really interesting thing happens. The line representing the YouTube viewer retention for this video sinks to zero. I have a better idea. I needed to figure out a way to reproduce the infrasound waves at 25 to 30 percent of the volume or amplitude near the data centers, and then get people to sit in a room while exposed to it without involved in any way. Now some of my viewers may remember this: a sonic weapon that I created that prevents people from speaking coherently. I think it was June or... elections, damn it. That's hard. Are you are where... Unless your wallet contains a CIA badge, the only place in the world where you could get away with testing sonic weapons on people is a deeply nerdy synthesizer convention. And the coolest and nerdiest one, in my opinion, is in downtown Los Angeles, organized by the company that makes the most deeply nerdy synthesizers - Buchla. I proposed a very weird idea to the organizers of Buchla and Friends, and after promising to come equipped with a whole lot of liability insurance, they were open-minded and let me rent out a booth and an entire enclosed room where I could have a team pull people in from the convention for a unique experiment, have them agree to some waivers, and then help me get the data I so desperately needed. I had an insulated room with five chairs, five pairs of headphones, and an oil painting of some owls. I had five people per research And for every single one of those research groups, I improvised a new story of why the owl painting was known to be haunted. It comes from early 90s Japan, comes from the late 1970s, late 1980s, when the Soviet Union collapsed in the 40s and 50s after the second world war in North Dakota. This not only kept me entertained and socially challenged throughout the weekend, but it also randomized a bias variable in the data. Then I told them that the audio recording that they would closed room while studying the painting and listening to the music for three minutes while paying attention to how they felt physically and emotionally. And it was actually really fun and interesting to hear about how the music made them feel and even some insightful [20:00] ominous details they noticed in the painting. What the volunteers didn't know was that there was a gigantic specialized subwoofer in the room that created infrasound in 50% of the sessions, and unless I saw water vibrating in a bottle or something, I often didn't know if the infrasound was on or off myself. It's not terribly loud in here. Let's turn the infrasound on. It's kind of scary. After the sessions, I had them fill out a survey about their experience and then swore them to secrecy for the next 48 hours before explaining to them that they had been lied to. Everything that I told you is a lie. Yeah, I got this on eBay for 30 bucks. The music, there's nothing substantial about it whatsoever. 50% of the people who are participating in this study are being subject to a fraction of the infrasound that's coming out of data centers across the US. The most important part of the study was a list of feelings or symptoms with a weighted were associated with real-time infrasound exposure symptoms from some older studies. I don't know why, but I feel like it's important that you know that the existence of this video depended on these survey results. There was really no point in making a video about infrasound levels that don't affect anyone. I've found that this is the big drawback about funding research through a YouTube channel. Some of the projects I'm involved with aren't really all that interesting, and you don't really even hear about them. Finding the truth is both expensive and risky. And if you want to see more content like this, consider joining But anyway, fortunately for the entertainment side of this business, the results were notable enough for me to finish the video and sit here and tell you about it. The experiment had over a hundred participants, but I disqualified anyone who missed a survey question, or someone who seemed suspicious or said that they felt a vibration, or where I was privy to whether the infrasound was on or off. So this is clean, double-blind data from 74 subjects. The people exposed to infrasound were 25% more likely to feel a tingling sensation or restless legs. The very few people who experienced infrasound group. Infrasound caused a 12% increase in tiredness, a 33% increase in nausea, and an alarming 150% increase in dizziness. That one is huge to me as vestibular disturbance is an obvious physiological effect of low frequency pressure waves, but it's also the most common complaint from residents in Granbury, Texas. Interestingly, the infrasound group was 11% less likely to be creeped out and 43% less likely to feel spiritual. One theory of mine is that things like dizziness someone from feeling spiritual or exploring the possibility of some type of extraordinary event. The infrasound also appeared to cause a 20% increase in chills and a 33% increase in irritability or depression. The infrasound group was three times more likely to experience eye irritation, which very much correlates with the famous ghost in the machine paper I mentioned in my previous infrasound video. The group also had an 80% increase in lethargy, which is interestingly pretty drastically different from their 12% increase in tiredness. That could be my mistake, definition of lethargy might just be a little bit too ambiguous. So I'm casting that from the data and we'll just stick with the tiredness metric. Team Infrasound was 55% more likely to feel anxious, two times more likely to experience sadness, and the big one: three times more likely to experience discomfort. Let's have a quick chat about data. I feel like in my very unique line of work here there are two different ways to communicate data. There's the youtuber way, which is useful And then there's the researcher way, which is useful for interpreting or building on research. And this isn't anything new. I would argue that it's not even a bad thing, because it's generally how the general public learns about science. You scroll through your news feed, you see blah blah is eight times more likely to blah blah. You click it and that opens up a little rabbit hole where you learn about things or behaviors that you never knew existed. So admittedly, in the previous segment, you've got a heaping dose of data that's optimized for a YouTube video. it to other researchers. So for example, the 33% more likely to experience nausea metric, the viewer may imagine a room with 100 people in it and then 33 of them experiencing nausea, which is very much not the case. The average nausea feeling score from 0 to 10 that each user put in for infrasound is 1.2. For the control group, it's 0.9. That does indeed reflect a 33% increase over the control group, but the numbers are just so low that it doesn't really warrant in further research. Another metric that had very low numbers but is worth further research is eye irritation. But this is just yearning for a much larger sample size to verify and figure out what's going on. In my opinion, the only profound metrics here are the spiritual one with the control group score averaging 4.2 and the infrasound one averaging 2.4. Then dizziness with the infrasound group averaging a score of 3 and the control group averaging a score of 1.2. And by far the most profound metric of all. And if this gets further research, this is the metric that will probably [25:00] be the focus of it, the average score in the control group for discomfort was 1.2. The average score for discomfort in the infrasound group was 4.8. When reviewing that data, it's hard for even the most skeptical person to come to any other conclusion then. The presence of elevated levels of infrasound is likely to lower your quality of life. I could probably spend the next hour criticizing my own research and rattling off ways that it could be improved, but a sample size of what's needed to get funding for further research which could improve the research conditions, increase the sample size, and test for different frequencies. What this research really needs is something that most research lacks: repeatability. I currently have three different ways of capturing infrasound, each with their own advantages and disadvantages. The first is a special measurement microphone that has a flat frequency response that goes down to 9 hertz, but it captures all the way down to 3 hertz. So it's very convenient. It operates just like a microphone, Now if we want full-spectrum infrasonic readings with no fall-off, or if we want scientific-grade accuracy, or if we want to record below 3 Hz, we need to leave the microphone world and use high-speed vibration detectors. One of these that is relatively affordable and accurate is the Raspberry Shake 3D system. It does require a whole lot of setup, but it can record indefinitely. The other is a small vibration sensor and accelerometer that's traditionally used to detect abnormalities and predict safety problems in large lithium batteries. cheap, it is expensive. If we want to treat the data from the Raspberry Shake or the vibration meter like a playable audio file, we need to convert the data into Pascals. I used to excruciatingly do this in Microsoft Excel, but now I've made a Python program that can do this with hundreds of files at a time. I'll put it on my GitHub. The majority of the motivation from this project comes from the literally thousands of messages, emails, and comments that I've received over the years from people seeking help or some form of relief from low frequency noise like the hum is for regulators, legislators, or even architects to start taking the environmental effects of sound into consideration when approving or designing a new data center or a new factory or infrastructure component or even a residential building or an office. Prior to this experiment, I feel like we have seen more than enough data to make infrasound monitoring and regulation just as important as air or water quality, especially when you look at the studies pertaining to the cardiovascular And while I suspect that having a universally recognized safety threshold for infrasound might be an overly optimistic goal with a government that can't even enforce their own rules with Elon Musk's methane turbines, it can go a long way in both existing and future lawsuits that seek remediation from problematic noise issues caused by data centers and infrastructure that is actively decreasing both the quality of life and health for those with the misfortune of living near them. I've already been in touch with peers to review and formalize this data, if not expand there are over 3,000 new data centers that are planned or already under construction in the U.S. alone. And right now is the time to have your local community and legislators set a standard for what's acceptable levels of power use, water use, air pollution, and noise and infrasound pollution. If you have a data center being built nearby your residents and the infrasound research in this video scares the shit out of you, now would be a great time to set up something like a raspberry shake and start storing a daily log of seismic and infrasonic activity. Because the to complaints and lawsuits is that there is no way of proving whether the infrasound wasn't already present before construction. And even if your legislators have failed you and have no motivation to help your community, you can bet that class action attorneys certainly do. And once again, if you want to help make more research and content like this possible, or join an amazing community, or have access to a pool of music samples and field recordings, and a whole bunch more, then my Patreon is for you, and you can join for as little as one dollar. Thanks for watching. Keep creating. Bye.