Model 3 Road Noise Evaluation

Model 3 Road Noise Evaluation

Full Disclosure: I am a long-time automotive noise and vibration engineer and live in the Bay Area working at one of the many EV startups (not Tesla). But I did spend nearly 5 years at Tesla developing the NVH behavior of the Model S (and beginning the Model X), so I know well the noise behaviors of these cars. I am happy to say that I just received my own Model 3 (VIN17xxx) and naturally performed some noise benchmarking tests on it. I have access to a Model S P100D (VIN160XXX) and a Mercedes S-550. Below are results of my road noise testing. I test these cars on a moderately coarse surface at 45 MPH (to minimize wind and engine noise), at both the driver's outboard ear and right rear passenger outboard ear locations.

Driver's Ear:
Model 3 - 67 dBA
Model S - 65 dBA
MB S-550 - 61 dBA

Rear Passenger:
Model 3 - 69 dBA
Model S - 67 dBA
MB S-550 - 62 dBA

As you can see, the Model 3 performs well (in general, anything lower than 70 dBA on this road is good), and while not quite as quiet as the Model S, it nonetheless is very comfortable. Of course, the MB S-550 is the champion here and nothing I have ever tested comes close. You'll also notice that the rear seat is consistently louder than the front, but that the S-550 does a better job of balancing that. The Model 3 rear seat does sound a little "open" due to noise coming from the trunk area, but it is not objectionable.

Kudos to the Tesla NVH Team as this is very good performance for a car in this class. I will also mention that wind noise is actually better in my Model 3 than in the Model S I have access to. The door sealing system (especially near the mirror base) is much improved in the Model 3 and it makes a huge difference. Nice job!

Kathy Applebaum | April 18, 2018

Thanks for posting actual data! Much appreciated.

sroh | April 18, 2018

I'll echo Kathy's thanks. I have heard some people mention the 3 seemed louder than the S. It's good to have actual data.

Williamticknor | April 18, 2018

This is a terrific post. Thank you for this very helpful information.

Just_Ted | April 18, 2018

Nice review, and good info. Thanks.

What are your thoughts on the benefits of adding sound dampening to the Model 3?

I usually do at least the doors (Dynamat or equiv., thin closed cell foam to acoustically de-couple, then 1 lb/sf mass loaded vinyl). My Jetta and my Prius have benefited greatly from this.

In the Jetta and Prius, the floor was also a problem area (rang like a drum) and treatment also brought a reduction in noise. Would that be necessary in the M3 with that dense battery under the floor?

What about just exposing the wheel wells (peel back carpet) and damping up by the feed and in the trunk? In your opinion does M3 need that or would it be diminishing returns at that point?

Thanks again!

Just_Ted | April 18, 2018


Teslaguy | April 18, 2018

Two things that improve noise from the rear.
I put the carpet from my MS in the rear trunk. Does not fit perfect but is ok.
I put a piece of cardboard in the “window” on the rear deck and put a couple small holes in it to allow some airflow to the trunk.
You could go a step further and use sound material behind the carpet over the rear wheel wells.

H271 | April 18, 2018

Nvharmony, do you have any opinions on the wind reduction kit available from rpmtesla?

ReD eXiLe ms us | April 18, 2018

I wonder how many hundreds of pounds of sound battening material are needed for the Mercedes-Benz to achieve its result, and whether the current C-Class comes close?

NVHarmony | April 18, 2018

I find the Model 3 is very good overall for noise and even though I am a "noise engineer" with very high standards, I have no plans to improve the NVH performance of my car. I like it the way it is.

Just_Ted - Ignoring my comment above, you might find that the Dynamat, foam decoupler and heavy layer makes a difference but it somewhat depends on what Tesla has already done behind the door trim panel. It also depends on how strong that noise path is in the first place. I am not yet brave enough to take mine apart and have a look. As for the floor, the battery pack under the floor pretty much wipes out any noise contribution from the floor in general. This is one of the HUGE noise advantages Tesla has over any car without a large battery pack under the floor. Speaking about "Dynamat", keep in mind that this only provides damping for structural vibrations of the sheet metal, and while this can (and does) contribute to noise, it is only a part of what is needed. Think of it as a necessary but not sufficient part of the solution.

H271 - My Model 3 (VIN 17xxx) has no wind noise issue related to the windshield gap, so I have to believe this is a build variation problem. If you car experiences this, then the gap seal will work great. Any easy test is to simply tape off the gap on your car and drive around and see if you can tell a difference.

ReD eXiLe ms us - the S-550 has lots of sound damping materials, probably about 100 lbs worth, and so does the Model S (although perhaps just a tad less). Most cars in that class will have similar sound packages (A8, 750i, LS460, etc.).

NVHarmony | April 18, 2018

I should also mention that my Model 3 has the 19" wheels with "ContiSilent" tires. In general, lower profile tires are noisier (so the 19's should be slightly louder than the base 18's), but the ContiSilent acoustic foam in the tire should offset that somewhat.

jdrsteiny | April 18, 2018

Is there anything to gain by gathering data at highway speeds, 65+?

Frank99 | April 18, 2018

NVHarmony -
Thanks for the authoritative work. Really increases my anticipation waiting for mine (got a VIN, but don't have a DS).

NVHarmony | April 18, 2018

One big advantage the Model 3 has over the Model S is that it does not have a hatchback like the S has, and so the passenger compartment is somewhat isolated form the noise generated in the trunk. This was a huge issue for us on the Model S and made getting the car really quiet (e.g. like MB S550) very, very difficult. The Model 3 closes off the trunk (but does have the "vent" in the rear shelf panel), so it can use less acoustical material and still achieve good noise levels.

NVHarmony | April 18, 2018

jdrsteiny - stay tuned. I may get to that soon :)

chris.pribe | April 18, 2018

Nice. Thanks, NVHarmony.

gundarx | April 18, 2018

@NVHarmony, thank you this is very informative, and for your past work with Tesla, which I'm sure trickled down into the Model 3.

ReD eXiLe ms us | April 18, 2018

NVHarmony: Interesting. I remember reading in an automobile enthusiast magazine that the Lexus LS held something like 200 lbs of sound battening material inside. Of course, that was a couple of decades ago... I suppose they may have used something lighter and hitech since, to help improve fuel economy while still maintaining the serene interior.

rnbows | April 18, 2018

Fantastic report! Thank you so much. I'm just waiting for my invitation, patiently I must add.

Ian.platt | April 18, 2018

Has anyone experimented with closing off the vent in the rear shelf? I don't have my Model 3 yet - delivery next week :), but I can't help think that this vent is a major contributor to cabin noise. The theory is that it is for allowing bass from the sub woofer to pas through into the cabin, but in my experience that is not necessary. Low frequencies from the sub have no problem passing through the rear shelf / back seats in my current car with no vent.

NVHarmony | April 18, 2018

Ian.platt - That is something I will probably experiment with on my car as it is easy to do without harming the interior in any way (it is my personal car, after all!). The vent is there mostly for fresh airflow to reach the "air extractors" in the lower rear corners of the trunk. All cars have air extractors (they look like clothes dryer vents with little flaps that close when there is no airflow) to normalize the pressure between the interior and exterior of the car. This is important when you close the doors and when you are operating the HVAC blowers in fresh air mode. The air has to go somewhere.

Perhaps there is a secondary function to allow the low frequency subwoofer sound waves to come through, but I kind of doubt that since the wavelengths we are talking about are measured in feet, and this opening is too small to have much impact on these frequencies.

patrick.oconnell | April 18, 2018

Great post! I've had my M3 for a week and have been dealing with a whistling noise when i exceed 40mph. I was at the service center today, took them for a ride and they agreed the whistling was an anomaly. He inspected the hood and noticed a small gap where the panel meets the glass (right under the wipers). Quick and easy something to keep an eye out for, well an ear out for!

dd.micsol | April 18, 2018

cool fine by me.

NVHarmony | April 18, 2018

patrick.oconnell - can you give more detail about what the fix was?

Beagle | April 18, 2018

Mr. Harmony,
Thank you for your comments & sharing you knowledge. My car won't be in for a few weeks, but in my trial run in a friend's car, I found it noisy. I wondered if some Dynamat or some similar type material might help. With your professional scientific background and intimate knowledge of the Model 3, it doesn't sound like it will be that easy to quiet some of the various sources. That said, I will certainly follow your future posts with great interest.

patrick.oconnell | April 19, 2018

NVHarmony - He literally just provided pressure with both hand on the panel (about a 1cm gap) and it kinda snapped back in place. I've now driven about 100 miles and no more whistle sound.

NVHarmony | April 19, 2018

Thanks Patrick.oconnell. I'll look more closely at mine and make sure there is no issue there.

Beagle - There are many sources of noise in a running car: road induced noise, wind noise, motor/gearbox noise, fan noise, squeaks/rattles, electric pump noises, etc. They are all more or less independent of each other, and need to be dealt with individually. All of these can be a combination of structure borne (vibrations passed through to the body mechanically) and airborne (noise penetrating the body through the air). Knowing which of these noise sources is most offending and then deciding if it's structure borne or airborne (or both) is the just the starting point of making a car "less noisy".

Also keep in mind that "noisy" is a relative term and depends entirely on context and expectations. Our brains adapt to our environment very quickly and we learn what is "normal" over time. We expect a certain level of noise on, say, an airplane or in an elevator, or in different classes of automobiles. As I mentioned, I find my Model 3 quite "comfortable" from a noise point of view, especially for its class. For that reason, I wouldn't do anything to improve it. Now, if I were in a Model S/X or an Audio A8 or MB S550, I would probably be less than satisfied with the noise level that I now perceive in my Model 3. Like I said, it is all about context and expectations.

With that in mind, can you be more specific about what you found to be noisy in your friend's car? If it was an early build, perhaps there were some build error states that resulted in higher levels of noise than they are building more recently.

geoffalexander737 | April 19, 2018

Yeah, after spending four days driving a P85 Model S and then going back to my M3 I noticed a very, very slight increase in road noise. Minimal.

vp09 | April 20, 2018

NVHarmony, thank you for some of the best info ever posted here.
Could you provide some context for your dB figures by posting examples of noise at 50 or 60 or 70 or 80 dBA? Just to let us here have some context ....
I took the business next to my home to Small Claims Court, and won, over their noise-- 90 dBC at my kitchen door, at midnight-- cogeneration unit running all night--
You are reporting decibels on the A scale-- how about the lower frequency noise?

vp09 | April 20, 2018

NV, could you, to help us understand this noise issue, articulate or explain what the difference between the Model 3 at 67 dBA and the Daimler car means? Like, 6 dB, what is that? Is that a reason to look at MB rather than a Model 3?

Also, if you happen to have this in your back pocket, what are the comparable dBA figures for, say, a Ford F150, or an older whatever type of car, or a VW bug from the 1960s-- just some figures for comparison.

I ask because I absolutely hate stupid, unnecessary, noise. Like my a-hole neighbor running a gasoline-powered motor hedge clipper at 0700. Or the other neighbor's dog that barks at 0640 and randomly throughout the day-- a low frequency noise like someone poking you in the ribs every so often.

At work it seems that the younger men try to make as much noise as they possibly can in the men's room-- slamming the doors, banging down on the paper towel dispenser, on and on.

I wasn't in Viet Nam, but almost, and the slamming rear door of my classroom when students wander in late and let the door slam shut really does remind me of 105 mm incoming artillery ....

cascadiadesign | April 20, 2018

NVHarmony - I've read many reviews and owner comments. One issue that comes up about the Model 3 is wind noise. Do you have any thoughts on the source? Is it just noticed more because EVs are so quiet? Do you think the gaps between the roof glass panels are a possible cause?

phil | April 20, 2018

Thanks, NVHarmony, good to hear.

I agree with jdrsteiny, it would be very helpful to have similar data at highway speeds. From what I've heard, wind noise is a bigger problem with these cars than road noise.

NVHarmony | April 23, 2018

Sorry for the delayed response, I just came back from a weekend trip in my Model 3 to Lake Tahoe (the car was great).

The science behind sound is very complex as it involves the physics and mathematics of wave propagation, frequency analysis, etc, but also the physiological and psychological aspects of human hearing. Rather than taking up space here, I often refer back to an old, but still excellent "primer" on this subject -

dB(A) (A-weighted decibels) is an imperfect but oft used metric for quantifying how loud a sound is. There are actually better metrics (e.g. Loudness in Sones), but dBA is so pervasive, we keep using it. A-weighting is one of four "weighting functions" that can be applied to sound, and it matches the human ear's own frequency sensitivity. There is also the B, C and D weighting functions which are used for various (but hard to explain) reasons. For example, dB(C) is used when low frequencies are dominant.

To try and put dB(A) in perspective, we often report it alongside a perceptual impression scale. In general, the industry uses a perception ratings scale from 1 - 10 (10 is the best) for assessing noise and vibration. In very rough terms, a change of 6 dB would represent about a 1 perception rating point change. This is very tricky territory here since dB(A) is a very simple, single-number metric which attempts to capture a complex, frequency dependent behavior, but cannot adequately describe the overall perception of a sound. Some sounds don't "seem" as loud because of the shape of their frequency spectra. Other sounds "seem" loud (even though they are not as measured by dB(A)), because of the nature of the sound spectrum. That said, the 6 dB/1 rating point change is a good rule of thumb.

In the case of the three vehicles above, ON THE PARTICULAR ROAD SURFACE I MEASURED, I would have given the Model 3 a 6 out of 10, the Model S a 6.5 out of 10 and the S-550 an 8 out of 10. Both the Tesla's suffer a little bit from a frequency spectra that has some strong "tonal" components which blend in with the broadband "white noise" (which is the dominant feature of tire/road noise). The S550 has a much smoother, flatter frequency spectra and the sound is both quieter and more "pleasant".

By comparison, I also happen to have the data on a VW GTI (the car my Model 3 replaced). It had over 100K miles on the odometer, and aftermarket 18" wheels/tires. On this same road surface, I measured it at 74 dBA in the front, and 77 dBA in the rear seat. I would give this car a perception rating of 5 (which is borderline "bad").

I'll leave it at that for now. I could write (and speak) for hours on this subject, but I'll spare this audience the details for the time being.

Frank99 | April 23, 2018

Thank you, NVHarmony. I love both the real live measured numbers, as well as the expert, reasoned analysis behind them.

gundarx | April 23, 2018

@NVHarmony, great post. Thank you.

H271 | April 23, 2018

NVharmony, do you have opinions on other things that can be done to reduce noise? Some people say putting a thicker carpet mat in the trunk can make a big difference. Would replacing all of the mats help noticeably?

NVHarmony | April 23, 2018

@H271 - I mentioned in one of my earlier posts that I wouldn't do anything to improve the Model 3 road noise. However, if I did decide to do that I would focus on suspension, tires and body structure. The reasons is that if you look at the noise spectrum (I have), you'll see that most of the sound energy is structure borne noise below 500 Hz. This part of the frequency range is not affected much by carpets, mats, and other "soft" acoustical materials. In this frequency range, it is all about vibration isolation from the road surface, and low frequency wave radiation from the body sheetmetal panels. One would need softer tires and software suspension bushings and additional body panel reinforcements. That could change the handling behavior and the weight of the car. Tesla has had to balance these oft-competing attributes and I'd say they did it pretty well.

If you had to, I suppose you could "slather" a bunch of Dynomat-like material on much of the body sheet parts: the firewall, door inner sheet metal, rear wheelhouse inner panels, trunk floor, etc. You'd need a lot of it, and probably multiple layers in order for it to make a noticeable difference. I wouldn't do it. It's good the way it is.

H271 | April 23, 2018

Interesting. Thanks for the detail. Someone else had posted he moved his Model S carpet mat to the Model 3 trunk and it helped with road noise. Maybe placebo. | April 23, 2018

@NVH: Excellent discussion of the complexities of noise measurement and the human perception of same. Rings a "Bell" for an old telephone engineer.:-)). For those interested hearing perception of noise is logarithmic rather than linear, hence the decibel measure which is logarithmic as so well explained above. This means that a noise level of 69 dB is twice as loud as one of 66 dB with the same spectral energy distribution for example, just to put the numbers above in perspective. The MB numbers are really impressive.
We recently took our Model 3 on a 900 mile road trip and we found the road and wind noise to be low enough to make conversation very comfortable. Disclaimer: My hearing aids were in place and turned on.

cascadiadesign | April 23, 2018

" ... I suppose you could "slather" a bunch of Dynomat-like material on much of the body sheet parts"

Or you could just crank up the music :)

mos6507 | April 23, 2018

"Or you could just crank up the music :)"

Speaking as someone with tinnitus, noise beyond a certain point becomes damaging to the ear. It's based on the combination of loudness and duration.

"if I did decide to do that I would focus on suspension"

Now I'm wondering what the noise profile might be on a 3 with air-suspension...

Rutrow | April 23, 2018

NVH can correct me if I'm wrong, I'll defer to her/his expertise, but noise is energy. Unwanted noise is wasted energy. If sealing the glass roof gaps reduced noise, Tesla would've done it to reduce the drag coefficient. Road noise may be more difficult to affect at it's source, but any improvements would make the car more efficient and improve range. Tire noise may very well be a give and take between: noise, handling and tire life.

RedShift | April 23, 2018


My 2013 Model S had couple of holes below the C pillars that weren’t blind-plugged. I plugged them when I was doing the dynamite type sound insulation by myself. Ended up doing most of the trunk, rear wheel wells (both inside and outside), all doors and under the rear seat. The improvement in the rear was noticeable.

The front doors and wheel wells didn’t make as much difference.

My S now feels nice and peaceful!

I drove the neighbor’s Model 3. Comparatively it’s as almost as quiet, corroborated by your tests. Driving was it was no match for my S.

3 was way better. :-D

w8ng4m3 | April 24, 2018


walnotr | April 24, 2018

@NVHarmony, thanks for the great info. It is my experience that the greatest contributor to cabin noise is the road surface condition. On a recent drive this was dramatically demonstrated when half of the surface we were traveling had been recently resurfaced and the other lane was not. The fresh surface was almost silent (by comparison) while the old surface required a much higher volume setting to still be able to hear radio conversations and even personal conversations with my passenger. This was in our Prius with low profile tires at 75MPH. If we could only have smooth roads everywhere life would be so much better! :-)

vmulla | April 24, 2018

Would you recommend any steps to seal the gaps around the pano roof to reduce wind noise?

Thanks for sharing your knowledge.

NVHarmony | April 24, 2018

@Rutrow - It's true that unwanted noise is also wasted energy. That said, the amount of power is very small indeed. For example, if you have a spherical source radiating noise uniformly with a source power of 1 Watt, the sound pressure level at 1 meter distance is 120 dB. The noise levels inside a car are much lower. If I assume 70 dBA (the A-weighting is important here), at 100 Hz (that's where the sound level is the highest), then a simple monopole source at 1 meter would need about 10 milliWatts (0.01W). That's just not enough power to spend time and money going after.

Speaking of road noise only, there is however, a lot more energy dissipated in the tire as it encounters increasingly rougher/coarser road surfaces, but improving the efficiency there requires serine effort (and potentially compromise) by the tire OEM. Even more energy is wasted in the shock absorbers, and in this case a clever system can recoup some of that energy by converting suspension motion into electricity. If you were clever, you could figure out how to put that electricity back in the battery.

Reducing wind noise does help improve drag, but again, the actual levels of power we are talking about are still in the milliwatts.

@walnotr - you are 100% correct! The road surface dictates everything. If roads were glass smooth, then engineers could design suspensions with race car setups, and no one would ever feel the difference. Sadly, us automotive engineers have to engineer cars to meet a wide range of use cases, road surfaces, weather conditions, etc, which forces a lot of complexity, cost and weight into the car. The lengths that Mercedes has gone to on the S-550 to isolate road induced vibration from reaching the cabin is remarkable. It is an engineering wonder, but it did not come cheaply and does add weight. But it is also the price of entry into that segment, and more importantly (I presume) it is what is expected from a car with the three-pointed star on the hood.

NVHarmony | April 24, 2018

@vmulla - I have read that some early production models had some wind noise issues. My car (VIN 17xxx) is good for wind noise. If you suspect that your car has issues, grab a role of painters tape and tape over all of the glass seams. If you hear an improvement, then there could be an issue with a leaky seal.

It is true that these small gaps "trip" the flow and create small turbulence eddys, but my experience is that they are not strong enough to generate a lot of noise unless they are combined with a compromised seal in the same area.

vmulla | April 24, 2018

Thank you.
My car VIN 0031xx did indeed have wind noise issues. Tesla service isolated the problem and replaced the seals around the doors. The difference was remarkable.

I was referring to potential reduction in wind noise by using a rubber seal kit sold by rpmtesla ( )
There is a separate thread that's going into the merits of this kit. What is your opinion on the worthiness of this kit?
Thank you.

Bryan.whitton | April 24, 2018

@vmulla I believe he has already addressed the question. His was a good suggestion. Get some painting tape, I would recommend 3M as it removes the most completely and easily in my experience, the cover the gap. If the sound level is noticeably lower get the kit. If not don't worry about it.
I would recommend 1.5" or 2" tape as it is easier to control and remove when the time comes.

M3forMe | April 24, 2018

great discussion. Thx NVHarmony!