THD in IEM and when to panic...

[Killcomic]

I was lurking around the Audio Science Review forums and something caught my eye.
People were absolutely losing their minds over the THD measurements of the Etymotic ER4XR.
They were all going "Ermahgerd! The ER4XR have over 1% THD! Worst IEM ever! Cats, and dogs living together, mass hysteria!"

Now, to my uneducated peasant mind, 1% distortion sounds pretty high.
I took out the certificate I got with my ER4XR and saw that distortion for my units was over 1.1%.

Funny, I've never heard any distortion with my ER4s.
Then I saw that the THD was measured at 100dB.
There's no way I'd ever listen to my music that loudly.

My question is, am I right to assume that a 1.1% THD at 100dB means very little when listening to music at normal volumes of about 50-60dB?

 

[bigshot]

In my sig file is a YouTube video where they take a horrible buzzing sound and dial it down under recorded music. At around -40dB it disappears (except for the silence between tracks). That is 1%, which I would say is about the threshold. Beyond that, it really depends on what the distortion sounds like.

I sometimes wonder why they list specs in percentages instead of -dB. It would be more intuitive if they did it the other way, and people would have more of a sense of what dB measurements mean in real sound levels.

 

[Killcomic]

In my sig file is a YouTube video where they take a horrible buzzing sound and dial it down under recorded music. At around -40dB it disappears (except for the silence between tracks). That is 1%, which I would say is about the threshold. Beyond that, it really depends on what the distortion sounds like.

I sometimes wonder why they list specs in percentages instead of -dB. It would be more intuitive if they did it the other way, and people would have more of a sense of what dB measurements mean in real sound levels.

Absolutely!
Sometimes I feel like a lot of measurements are provided without much context.
Anyway, considering I didn't hear any distortion with my ER4's, I'm not particularly bothered.

 

[71 dB]

My question is, am I right to assume that a 1.1% THD at 100dB means very little when listening to music at normal volumes of about 50-60dB?

1 % of harmonic distortion means nothing at low frequencies. At some higher frequencies it may be audible with some test signals. However, the distortion should be signficantly lower at lower listening levels, so you probably can listen to DANGEROUSLY LOUD (for example at level 90 dB) with no more than 0.1 % THD, which is inaudible. Also, very low level of audible distortion tend to make the sound richer/warmer, so even if you heard the distortion the effect would be positive rather than negative to the enjoyment.

No panic warranted.

 

[bigshot]

These are IEMs, right? Distortion in IEMs make it sound warmer? That is the opposite of a lot of speakers.

 

[71 dB]

These are IEMs, right? Distortion in IEMs make it sound warmer? That is the opposite of a lot of speakers.

Generally very low level distortion that is audible makes sound warmer, because it adds harmonics and other frequencies, but not so much to make the sound harsh.

 

[castleofargh]

Upon seeing a bad looking measurement, you must immediately stop enjoying whatever it was that you were enjoying. It's the law!

THD is famous for one thing, and that's not being relevant to predict listener preferences. As you suspect, the distortions on the er4XR aren't that high with a lower signal. It's not great but also not particularly strange for a single BA driver. I wouldn't recommend this to people who will listen loudly because some of the mess increases near 1kHz and that's not a frequency range where we want any mess to be loud, but beyond that, your pair has better channel matching than most IEMs, one of the best isolation, and it's small. I can think of worse choices in life.

1% means 40dB below signal. From that you can understand that it's probably audible, but it's one of those times when it might be good to invent a distinction between audible and clearly noticeable, where audible would be exactly what it means, and noticeable would be that we notice and identify it while playing music. It's really hard to notice THD at or below 1%. First because -40dB is already very quiet compared to the full scale music playing over it so a lot of the THD will hopefully be masked. And second because instruments and voices contain a lot of harmonics already. Subjectively THD will alter the sound, but it doesn't mean we'll know where and how, usually we can't tell that anything is out of place thanks to not having a direct comparison with the exact same FR and no distortions. You can go listen to a super low THD IEM and pretend to understand what part of the perceived difference is THD, but I'd bet good money that you'd be wrong. The frequency response difference between the IEMs and the impossibility to volume match them will surely throw you completely off subjectively.

The SR is very close in measurements and while it seems to do a little better at higher listening level in term of distortions, it's still pretty bad compared to most random vented dynamic drivers. If THD wakes you up at night, I would suggest to never touch a balanced armature ever again.

 

[Killcomic]

Upon seeing a bad looking measurement, you must immediately stop enjoying whatever it was that you were enjoying. It's the law!

THD is famous for one thing, and that's not being relevant to predict listener preferences. As you suspect, the distortions on the er4XR aren't that high with a lower signal. It's not great but also not particularly strange for a single BA driver. I wouldn't recommend this to people who will listen loudly because some of the mess increases near 1kHz and that's not a frequency range where we want any mess to be loud, but beyond that, your pair has better channel matching than most IEMs, one of the best isolation, and it's small. I can think of worse choices in life.

1% means 40dB below signal. From that you can understand that it's probably audible, but it's one of those times when it might be good to invent a distinction between audible and clearly noticeable, where audible would be exactly what it means, and noticeable would be that we notice and identify it while playing music. It's really hard to notice THD at or below 1%. First because -40dB is already very quiet compared to the full scale music playing over it so a lot of the THD will hopefully be masked. And second because instruments and voices contain a lot of harmonics already. Subjectively THD will alter the sound, but it doesn't mean we'll know where and how, usually we can't tell that anything is out of place thanks to not having a direct comparison with the exact same FR and no distortions. You can go listen to a super low THD IEM and pretend to understand what part of the perceived difference is THD, but I'd bet good money that you'd be wrong. The frequency response difference between the IEMs and the impossibility to volume match them will surely throw you completely off subjectively.

The SR is very close in measurements and while it seems to do a little better at higher listening level in term of distortions, it's still pretty bad compared to most random vented dynamic drivers. If THD wakes you up at night, I would suggest to never touch a balanced armature ever again.

I wasn’t worried before but I’m getting worried now!
*Sweats profusely and starts looking up DD IEMs*

 

[redrol]

BAs have high distortion compared to dynamic drivers. Get a 2XR if you want less.

 

[Blackwoof]

My question is, am I right to assume that a 1.1% THD at 100dB means very little when listening to music at normal volumes of about 50-60dB?

Your ER4XR would be lower than 0.5% below 90db It wouldn't mean anything at all. It also sounds like there trolling by riding on the "Ew....Single BA" since the ER4SR they tested was 0.3% at 100db pretty much as good as Planar/DD headphones. It the same people who tried to shut me down when a poster showed a ER2XR at 1.98% at 1KHz(also 100db) as they would scoff at the ER4XR being higher THD.

Both Reddit, ASR & SBAF have became foam in the mouth toxic about using the ER4XR as to why BA headphones need fade away. I still remember having folk getting teary eyed at me showing that the ER4XR & ER4SR can handle 6 ~ 12db bass under-125Hz no problem.

 

[bigshot]

I've never understood why people on Head-Fi become like crazy soccer fans over brands and models to the point where they lose sight of fidelity. There's more than one way to skin a cat. No single brand or model is the be-all-and-end-all. I guess people tend to recommend whatever model they bought themselves, and they take any recommendation of something different as an attack on their choice. Validation bias can make people do dumb things.