Why Is An MRI Machine So Loud?

The short answer: An MRI is loud because its gradient coils physically vibrate. Pulsed electric current runs through those coils inside a very strong static magnetic field, and the resulting Lorentz force pushes the windings back and forth thousands of times per second. That motion pumps the surrounding air and you hear it as banging, knocking, and buzzing. Clinical scanners commonly reach 90 to 120 decibels, so earplugs or headphones are standard.

What actually makes the banging noise?

The big magnet in an MRI is silent. It sits at a constant field strength and does not move. The noise comes from a different set of coils inside the bore, called gradient coils.

MRI needs to know where in your body each signal came from. To do that it adds small, rapidly changing magnetic fields on top of the main field. Three gradient coils create those changes along the three axes. Every image line requires the current in one or more of these coils to switch on, ramp up, ramp down, and reverse.

Here is the physics. A wire carrying current inside a magnetic field feels a sideways push called the Lorentz force. A review in Frontiers in Physics describes it plainly: rapidly switched currents pass through the gradient coils in the presence of the static field, the conductors vibrate, and those vibrations radiate into the air as acoustic pressure waves.

In other words, the gradient coil works like the voice coil of a giant loudspeaker. The pulse sequence is the waveform. The scanner is not making noise as a side effect of imaging. The noise is the imaging, made audible.

The force scales with both the current and the strength of the static field. That is one reason higher field scanners tend to be louder. The National Institute of Biomedical Imaging and Bioengineering notes that patients hear clicking and beeping, and that sound intensity reaches up to 120 decibels in some MR scanners.

How loud is an MRI in decibels?

Loud enough to matter. Published measurements give a range, not a single figure, because the answer depends on the machine and the scan.

Researchers who measured clinical Philips scanners reported the results in Acoustic Noise Levels in High-field Magnetic Resonance Imaging Scanners. On the 3 tesla system, the time-averaged level averaged 91.4 dBA and the maximum averaged 96.5 dBA. On the 7 tesla research system, the same figures were 105.6 dBA and 114.0 dBA.

Fast sequences push higher. The Frontiers review reports that echo-planar imaging, used for functional and diffusion scans, produces sound pressure levels in the range of 110 to 120 dB.

For context, the NIOSH recommended exposure limit for workplace noise is 85 dBA averaged across an 8 hour day. NIOSH also uses a 3 dBA exchange rate: every 3 dBA increase halves the safe exposure time. A typical MRI sequence sits well above that line, which is why protection is not optional.

Why does the sound change from knocking to buzzing?

Because each pulse sequence switches the gradients on a different schedule. The repetition rate of the gradient pulses sets the pitch you hear. The amplitude and slew rate set the volume.

A slow, heavy sequence gives you a deep, spaced-out knock. A fast sequence gives you a high buzz or a machine-gun rattle. The same scanner can be dramatically quieter or louder depending on which series is running.

The numbers show it. In the high-field measurements above, a T1 spin echo series on the 3 tesla scanner produced a time-averaged 100.9 dBA, while a FLAIR series on the same machine produced 81.3 dBA. That is a wide spread on one magnet in one sitting.

This is also why the noise pattern tells you roughly where you are in the exam. Each change in the sound means the technologist started a new series. If you want a sense of how many series a study includes, see how long an MRI takes.

Is MRI noise actually regulated?

Yes. The FDA guidance on significant risk investigations of magnetic resonance diagnostic devices sets two acoustic thresholds. A scanner is treated as significant risk if it produces a peak unweighted sound pressure level greater than 140 dB, or an A-weighted rms sound pressure level greater than 99 dBA with hearing protection in place.

Read that second number carefully. The 99 dBA ceiling assumes you are already wearing protection. The raw level inside the bore can be far higher, and the protection is expected to bring it down. That is a design assumption, not a courtesy.

Can MRI noise hurt your hearing?

Unprotected, yes, it can. That is the whole reason for the limits above. With properly fitted protection, a diagnostic MRI is considered safe for hearing.

Some people notice muffled hearing or a brief ringing right after a long scan. Tell the technologist if that happens, and tell them during the scan if an earplug works loose. A plug that slides out partway through the exam loses most of its benefit.

Persistent ringing, new hearing loss, or ear pain after a scan is worth reporting to your doctor rather than waiting it out. If you are trying to sort out which symptoms need a visit, symptom.md covers that kind of triage question.

What hearing protection should you expect?

Every MRI center should offer it. Foam earplugs are the baseline. Many sites add MRI-safe headphones over the plugs, which also let the technologist talk to you and can play music.

The high-field measurement paper notes that well-fitting earplugs are a low-cost passive option that can attenuate noise by 10 to 30 dB. MedlinePlus describes the machine as producing loud thumping and humming noises and tells patients they can wear earplugs to reduce it.

Three practical points. Ask for both plugs and headphones rather than one or the other. Have the technologist check the plug seal before you go into the bore, because a poorly seated plug is the most common failure. And say something on the intercom if the fit changes, since the table can be paused between series. Noise is also a common trigger for anxiety in the bore, which overlaps with MRI claustrophobia.

Are there quiet or silent MRI scans?

There are, and they work. Zero echo time (ZTE) and ultrashort echo time (UTE) sequences keep the gradient nearly constant during readout instead of ramping it up and down. Less switching means less vibration.

A review of silent zero TE MR neuroimaging puts hard numbers on the gap. ZTE-based brain sequences at 3 tesla measured 53 dBA for RUFIS and 51 to 58 dBA for PETRA. Conventional comparisons in the same table ran 105 dBA for IR-SPGR and 103 to 114 dBA for gradient echo EPI. ZTE methods typically operate within about 5 dB of ambient room noise.

The catch is coverage. Quiet sequences are not yet available for every body part, every scanner, or every clinical question, and some quiet variants trade off scan time or contrast. Still, they are worth asking about, especially for brain imaging, for children, and for anyone with sound sensitivity. If a specific study is on your calendar, the general prep in what to expect at your first MRI is a good starting point.

The bottom line

The noise is the gradient coils flexing under Lorentz forces as their current switches inside the main magnet. It is unavoidable in conventional imaging, it commonly runs 90 to 120 decibels, and the FDA caps the A-weighted level at 99 dBA with hearing protection in place. Wear the plugs, ask for headphones too, and ask whether a quiet sequence is an option for your exam.

Last updated: September 2026. This article is for informational purposes only and does not constitute medical advice. Talk to your radiology team or your doctor about hearing protection and any hearing symptoms after a scan.