• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

THD in Research Headphones

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
Some of the typical headphones/IEMs used for audiometric purposes are Telephonics TDH49, Etymotic ER3 or Beyerdynamic DT48A. I've been given Telephonics and Etymotics (though not those exact models) during hearing tests, for example.

I bring it up because these are considered good enough to do clinical research of thresholds, masking and so on. So they must have exceptional characteristics, right? Not really.

The manufacturers don't post much data. What seems typical is flat response over the mids and falling response at either end, although Etymotics are flat in the lows as well.

Innerfidelity has measurements of the 5 ohm DT48S, similar to the audiometric version, and for 90dB SPL the THD is around 0.1% or -60dB. Not at all different from other kinds of headphone transducers.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
The reason the DT48 were used is because they were fully documented and there was a standard 'compensation' for it available which could be used in hearing test equipment because it had the opposite curve in that equipment so that results are always the same.
The DT48 had a consistent FR over the years of production. It wasn't about 'flat' frequency response or low distortion.
When you bought the hearing test equipment you had to use the headphones that equipment was designed for if you wanted to get calibrated results.
Today other headphones are used with their own curves as well as boneconduction devices, which again, have their own calibration files which have to be selected.

Usually the hearing tests are to determine the lower limits of hearing for speech. Some could only measure up to 4khz, others up to 8kHz.
When using the headphones at the lower hearing limit distortion does not matter as that will be below the hearing limits you test for.
 
Last edited:
OP
pozz

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827

Attachments

  • 744823.pdf
    538.9 KB · Views: 432

Blujackaal

Active Member
Joined
Feb 7, 2020
Messages
221
Likes
102
Etymotic okay with their BA headphones capping at 1% because they said at Head fi their no hard data that people can tell in music or in pure sine tests. They have a dynamic driver version but they said it only good for tests at 90 to 120db where <0.3% would matter more. But with the THD varation on the ER4SR does that mean a 4SR that rates <0.3% at 97db 1KHz is fine for that job?, Which is why i find the ER2SE/DD a odd one.

FYI op the ER4SR & ER2SE are still research headphones. Wouldn't shock me some doctors using ER4SR for that use since some did with the 90's ER4S. Which could explain why the detail so high on them for a music headphone.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,390
Location
Seattle Area
Innerfidelity has measurements of the 5 ohm DT48S, similar to the audiometric version, and for 90dB SPL the THD is around 0.1% or -60dB.
His measurements include noise so not just distortion. I know he used a chamber but I don't see his noise level documented inside there.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
0.04% was about the lowest he could measure it seems (at 100dB SPL) in the Oppo PM3
Noise + THD.
 

lashto

Major Contributor
Forum Donor
Joined
Mar 8, 2019
Messages
1,045
Likes
535
Some of the typical headphones/IEMs used for audiometric purposes are Telephonics TDH49, Etymotic ER3 or Beyerdynamic DT48A. I've been given Telephonics and Etymotics (though not those exact models) during hearing tests, for example.

I bring it up because these are considered good enough to do clinical research of thresholds, masking and so on. So they must have exceptional characteristics, right? Not really.

The manufacturers don't post much data. What seems typical is flat response over the mids and falling response at either end, although Etymotics are flat in the lows as well.

Innerfidelity has measurements of the 5 ohm DT48S, similar to the audiometric version, and for 90dB SPL the THD is around 0.1% or -60dB. Not at all different from other kinds of headphone transducers.
Looks like audiologists are not as picky as audiophiles :)

And be glad, you got really good ones. When I first seen my audiologist's HPs, I thought he was joking. They looked like a billion people tried them during the last +20 years. And the 'design' was kinda similar to the HPs worn by Morse operators in WW2 movies. The whole business and all other devices looked brand new & tiptop, those dingy HPs felt quite wrong in that scenery. He assured me that they were "perfectly calibrated" but I didn't even think of asking about the model/make.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
The system he uses probably is perfectly calibrated and most likely has to have it calibrated every year or so.
For testing the hearing no audiophile or 'flat' or low distortion headphones are needed.
They need be robust, constant in quality over the years, so when you buy a replacement 15 years later, the new one still has the same response.
All they have to do is play very soft single tones or noise bands.
The test gear it is connected to is what is calibrated to the selected (used) headphone which is documented. It can be wildly varying in response as long as those variances are known.
 

Blujackaal

Active Member
Joined
Feb 7, 2020
Messages
221
Likes
102
Looks like audiologists are not as picky as audiophiles :)

And be glad, you got really good ones. When I first seen my audiologist's HPs, I thought he was joking. They looked like a billion people tried them during the last +20 years. And the 'design' was kinda similar to the HPs worn by Morse operators in WW2 movies. The whole business and all other devices looked brand new & tiptop, those dingy HPs felt quite wrong in that scenery. He assured me that they were "perfectly calibrated" but I didn't even think of asking about the model/make.

Yeah it funny seeing audiophiles get upset with Etymotic ER3/ER4 thd, But will gladly defend other gear with as high THD.
 

lashto

Major Contributor
Forum Donor
Joined
Mar 8, 2019
Messages
1,045
Likes
535
The system he uses probably is perfectly calibrated and most likely has to have it calibrated every year or so.
For testing the hearing no audiophile or 'flat' or low distortion headphones are needed.
They need be robust, constant in quality over the years, so when you buy a replacement 15 years later, the new one still has the same response.
All they have to do is play very soft single tones or noise bands.
The test gear it is connected to is what is calibrated to the selected (used) headphone which is documented. It can be wildly varying in response as long as those variances are known.
I did trust his "perfectly calibrated" and you are right, that should be enough for the usecase. Still, would not want a pair of his HPs for home :)

Yeah it funny seeing audiophiles get upset with Etymotic ER3/ER4 thd, But will gladly defend other gear with as high THD.
Most of us engineers seem to be convinced that HD is just evil. It's not. At least not always. It may also sound good/fun.
OTOH, for research/measurement purposes low-THD HPs should be better.
 
Last edited:

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,747
Likes
16,182
The Correlation between Distortion Audibility and Listener Preference in Headphones

It is well-known that the frequency response of loudspeakers and headphones has a dramatic impact on sound quality and listener preference, but what role does distortion have on perceived sound quality? To answer this question, five popular headphones with varying degrees of distortion were selected and equalized to the same frequency response. Trained listeners compared them subjectively using music as the test signal, and the distortion of each headphone was measured objectively using a well-known commercial audio test system. The correlation between subjective listener preference and objective distortion measurement is discussed.

https://www.aes.org/e-lib/browse.cfm?elib=17441
https://www.researchgate.net/public...ibility_and_listener_preference_in_headphones
 

Blujackaal

Active Member
Joined
Feb 7, 2020
Messages
221
Likes
102
I did trust his "perfectly calibrated" and you are right, that should be enough for the usecase. Still, would not want a pair of his HPs for home :)


Most of us engineers seem to be convinced that HD is just evil. It's not. At least not always. It could also sound good/fun.
OTOH, research/measurement HPs should be much better at low THD.

Etymotic been making hearing aids/research IEMs since the 80s. One rep on head fi said they've yet see any data proving Estats/dynamic drivers at 0.35% gains over the ER4XR(0.7 ~ 1.8%) & ER3/ER4(0.5 ~ 0.9%). For music/research/studio use.

Also this all at 90 ~ 105db, Under 90db the ER4XR at 1% would <0.5%. The only time they said a <0.1% was useful if they were doing 90 ~ 120db tests.
 

lashto

Major Contributor
Forum Donor
Joined
Mar 8, 2019
Messages
1,045
Likes
535
The Correlation between Distortion Audibility and Listener Preference in Headphones

It is well-known that the frequency response of loudspeakers and headphones has a dramatic impact on sound quality and listener preference, but what role does distortion have on perceived sound quality? To answer this question, five popular headphones with varying degrees of distortion were selected and equalized to the same frequency response. Trained listeners compared them subjectively using music as the test signal, and the distortion of each headphone was measured objectively using a well-known commercial audio test system. The correlation between subjective listener preference and objective distortion measurement is discussed.

https://www.aes.org/e-lib/browse.cfm?elib=17441
https://www.researchgate.net/public...ibility_and_listener_preference_in_headphones
For non AES members, youtube seems to have most of the info.

It's an interesting paper but not sure why would anyone still try to correlate THD (as a number). Or they only included it to show (again) that it's almost useless !?
OTOH, their non-coherent-distortion (free access) looks interesting and it was the only one with ok(ish) correlation. And it looks good enough, maybe @amirm can start using it. IMO, we need a 100% new metric for audibility and preferences. Even those two might be so different that we'll need two separate metrics.
 
Last edited:

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,747
Likes
16,182
It's an interesting paper but not sure why would anyone still try to correlate THD (as a number). Or they only included it to show (again) that it's almost useless !?
Fully agree and was asking myself the same.
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,424
Likes
4,030
Location
Pacific Northwest
It's common for hearing tests to only go up to about 8 kHz. Usually they're only testing for obvious deficiencies that would make it difficult to discern speech.

One exception is (long ago) the medical I took for entering the Naval Academy. I don't remember what headphones were used, but I do remember the high tones sounded like the whine of an old TV set (I believe around 15-16 kHz, which I could hear back then). They played tones of random duration & frequency at random intervals, and you lifted your hand when you heard it. When I took that medical I was paired up with another guy who wasn't raising his hand as much as I was. After the hearing test the examiner asked the other guy, "Do you attend rock concerts?". The guy answered "Yes". The examiner nodded and noted that on the clipboard.

As mentioned above, medical equipment doesn't have to be super low distortion or linear response. It just has to have consistent, known response and be robust & durable.
 
OP
pozz

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
OP
pozz

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
Also, I should add that I have another paper (can't recall the name right now) that deals with techniques for cancelling distortion for acoustic and listening tests when using test tones.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Top Bottom