New Research on Audibility of Distortion in Headphones

[Jiraya369]

I don't know... the few times I was around some young folks with closed headphones on that I could clearly hear would be evidence at least a number of them do.
Arguably these kids could not care less about distortion and sound quality. They want loud sound and a lot of bass.

I agree that unless one listens at loud levels non-linear distortion is not a contributing factor that defines 'sound quality' but rather mostly linear distortion does.
The biggest problem is when measurements show a lot of distortion at 104dB and dismisses that headphone while the same folks never listen loud to protect their hearing.

They are probably the same folks that believe 112 SINAD is preferred over 95 SINAD and will not buy the 95 SINAD device even though it is inaudible.
It's just how people are wired.

that's what im saying, people just focusing on things that are measured instead of understanding what those mean for our ears and just wasting their time and money as well as others'.

 

[solderdude]

Nothing wrong with focussing on things that are measured but is only really important when one understands the audibility of the measurements when listening to music.

Then there is a difference between thinking one knows how to interpret measurements and actually knowing how measurements translate to actual music enjoyment.
I don't really mind if people are wasting their time and money on any hobby.
What I kind-of do care about is deception but then again, only what counts as deception to me and not in an absolute sense.

In any case .. I always like to read 'new research' especially when it confirms what I already knew but was never really 'tested'.
This article is just one of those.

It is just a small sample of headphones being used as well as a small amount of testers ... so there's that.
Not conclusive in all cases for all people. Just conclusions based on a small amount of test samples.

 

[illusonic]

Night clubs and rock concerts can be deafening, again due to treble frequencies, not deep bass.

According to below study, exposure to 63Hz low‐frequency octave-band noise at 90 dB SPL for 6 h leads to significant hearing loss over an extensive range from low to high frequencies.

Extensive hearing loss induced by low‐frequency noise exposure - PMC

With little attention given to low‐frequency traffic noise and our understanding that cochlear function may be highly susceptible to low‐frequency noise, there is an urgent need to determine traffic noise‐induced hearing loss (NIHL), not only the ...

pmc.ncbi.nlm.nih.gov

 

[Transistor]

Amir deliberately listens at maximum volume, bordering on hearing damage, to give distortion at least a chance

 

[illusonic]

With respect to distortion in bass, no, you don't need controls. When the woofer falls apart, distortion is massive and easily audible beyond any bias. Listen to the start of this track on your speakers:


Now gradually increase the volume (use a proper streaming service such as Tidal or Quobuz, not youtube above). Unless you have world-class, flat response down to 20 Hz, you are either going to hear the low frequencies faintly or with severe distortion. Distortion that sets in at levels that are easily tolerable.

Just listened to your proposed track again (Fading Sun from Terje Isungset) on my Palmer Orbit 11s. No sub, 5x7m room, 2.5m listening distance. I cleary hear the 30Hz rumble that my smartphone analyzer shows. No obvious, let alone objectionable distortion. Maximum listening volume is limited by what I feel pleasant with regard to midrange. When I turn the volume up past that point, things in my room start shaking and produce audible noise. Not the speakers.

Your review suggested that the speakers are somewhat flawed because they provide extension down below 30Hz with high distortion and limit above 86dB SPL (actually they start around 89dB in the bass ONLY). You also reported audible distortion in bass, which actually only becomes obvious/objectionable when the limiter kicks in while playing sine wave or other band-limited content that includes bass without harmonics (like above mentioned track). It's limiter-distortion not distortion from the speaker that's quite audible (and could be improved).

In a normal livingroom situation including 2 speakers close to walls, there will be a lot of room gain (you also complained about too much bass in some songs). I simply compensated with the near-wall DSP setting of the Orbits reducing bass by 6dB with a shelving filter starting at 150Hz (intstead of 100Hz for the normal bass low-shelves). Compared to your single speaker test, this gives me 12dB higher level before the limiter kicks in. In this way, I simply don't encounter any limitations in everyday listening while enjoying full frequency range. Your review suggests otherwise, which in my view makes you miss the "Thruthear Zero moment" of loudspeakers. Fullrange speakers in good quality and small form factor have finally become affordable.

Bottom line is that harmonic distortion in the bass is simply a weak metric for listener preference. Even if hamonic distortion in bass is audible, especially compared directly to a low distortion signal, it may still not be objectionable. I had the chance to listen to a system that actually reduced speaker distortion by up to 12dB and could be switched on/off for direct comparison. Test device was an 8" bass speaker running with low-pass. Me and my collegues (all engineers) either did not hear an improvement (no preference) or even preferred sound with distortion. That was when I learnt that the uggly low-frequency harmonic distortion hump is mostly a visual problem. In that sense, looking at measurements before listening definitely produces strong bias.

Especially in fullrange or bass+midrange speakers, IMD can be a much bigger issue than (T)HD. Voices modulated by bass are easily detected once IMD gets in the 10% region. If THD is low, IMD will also be. At least THD gives a hint regarding audible speaker distortion. However, multiway systems and speakers with Kms-based HD don't have IMD issues. This makes multitone distortion the most important distortion metric. Anyone wondering how IMD sounds can take the Klippel listening test.

Listening Test

www.klippel.de

 

[Vladimir Filevski]

there isn't any and there is no NEED for it, just do a distortion test to find your own distortion identification threshold and then just base your purchase around that perhaps, that's it.

No need for blind listening tests??? Are you joking, or...?
I am testing myself often on distortion tests and I know my hearing threshold on (harmonic) distortion. But it is very far from "that's it"! What about linear distortion - i.e. deviations from flat frequency response? And what about bad waterfall diagram (bad resonances)? And what about...
As you can see, your criteria for buying something based solely on harmonic distortion is very, very wrong.
So, there are no blind listening tests published on regular basis anywhere... and you don't trust sighted subjective tests because they are biased. Looks to me as a dead-end for you.

There must be a point to things, what is the point of looking for someone's subjective listening impressions if they don't align with your own personal thresholds and stuff? What if they're completely biased?
Why should I care if there's some bloke out there doing blind testing on this if I have products I'm happy with whose distortion I could not find during test sweeps even at loudness levels where I'm uncomfortable? There's no point for me.

Why do you trust yourself when doing only sighted listening tests? What if your tests are completely biased?
That is why you need "some bloke out there doing blind testing".

Btw erin does his listening before looking at graphs.

Yes, I know, but that should be irrelevant to you, because you don't trust sighted listening tests, anyway.

 

[Jiraya369]

No need for blind listening tests??? Are you joking, or...?
I am testing myself often on distortion tests and I know my hearing threshold on (harmonic) distortion. But it is very far from "that's it"! What about linear distortion - i.e. deviations from flat frequency response? And what about bad waterfall diagram (bad resonances)? And what about...
As you can see, your criteria for buying something based solely on harmonic distortion is very, very wrong.
So, there are no blind listening tests published on regular basis anywhere... and you don't trust sighted subjective tests because they are biased. Looks to me as a dead-end for you.


Why do you trust yourself when doing only sighted listening tests? What if your tests are completely biased?
That is why you need "some bloke out there doing blind testing".


Yes, I know, but that should be irrelevant to you, because you don't trust sighted listening tests, anyway.

idk if you're intentionally dodging my points or actually fail to understand what i said, i wont bother, you do you bud.

 

[Vladimir Filevski]

idk if you're intentionally dodging my points or actually fail to understand what i said,

You failed to understand that you are contradicting yourself - I tried to explain that to you, but no avail...

 

[Somafunk]

Now gradually increase the volume (use a proper streaming service such as Tidal or Quobuz, not youtube above). Unless you have world-class, flat response down to 20 Hz, you are either going to hear the low frequencies faintly or with severe distortion. Distortion that sets in at levels that are easily tolerable.

Had a listen to that track on tidal/roon and i eventually gave up on the volume as I could feel the sub bass in my chest and hear the mugs/plates in the kitchen units sing n’ rattle

I definitely do not have a world class system (in sig below) but it’s decent enough and the manufacturers specifications state (rightly/wrongly ) that the subs are -3db at 16hz for 85db, with a max SPL of 110db - whether that’s entirely accurate or not I dunno?…..but I’ve never sat here and thought……I want more bass in my electronic music, James Blake/Aphex Twin/Underworld/Craven Faults etc…etc…

 

[Ype]

idk if you're intentionally dodging my points or actually fail to understand what i said, i wont bother, you do you bud.

It is in fact you that needs to learn how to listen, in every sense of the word.

that's what im saying, people just focusing on things that are measured instead of understanding what those mean for our ears and just wasting their time and money as well as others'.

We do not have a standard/trustworthy method of quantifying the audibility of non-linear distortions so we need trained listeners. You are the one wasting our time.
If you don't want to trust Amir that is your choice but from your posts you give me the impression that you are not interested in learning how any of this works.

See this as my final response to you until you engage with people here while showing a willingness to listen and learn.
There are people out there that do want to learn, like me.

 

[Transistor]

They are probably the same folks that believe 112 SINAD is preferred over 95 SINAD and will not buy the 95 SINAD device even though it is inaudible.
It's just how people are wired.

Even an amplifier with a SINAD of 90 is completely transparent and in no way inferior to an amplifier with a SINAD of 120. But many swear by SINAD. Although no one would ever be able to distinguish between two such amplifiers in a blind test, an amplifier with a higher SINAD can be considered superior only from an engineering standpoint. In this sense, it is a more sophisticated device. But for music lovers, there will be no improvement in sound quality.

 

[solderdude]

For those interested if they can hear anything below -90 dB attached a link to FLAC file.

Secure File Sharing - Filemail

Share files of any size, any format, any type, securely

www.filemail.com

The audio-file can ONLY be downloaded till May 2nd and is 8.5MB in size... after May 2 the link will be 'dead'

It contains music played at normal level. When starting the test use the first 10 seconds to let the music play loud.
After a while the music is attenuated for a few secs by 50dB, comes back on for a few seconds only to drop -60dB, goes up, down to -70dB, up again, down -80dB, up again and to -90dB and back up.
An easy test to see how much lower a signal has to become before you don't hear anything anymore (while music is actually playing).

Very enlightening. When you can't hear anything at a certain attenuation anymore how can one hear any other signal below that point anyway.
Highly recommended for people that keep wondering if -100dB or -120dB is what they really need.

 

[amirm]

A good blog from Purifi in audibility of distortion from woofers: https://purifi-audio.com/blog/tech-...o-whats-the-point-of-low-distortion-drivers-4

There are audio samples there as well. It follows the same type of analysis/conclusions as this study and that if Fielder:

"The science was right: even gross bass distortion is remarkably difficult to hear. The fine print says yes, provided it stays in the bass."

Note that this distinct from the type of distortion where the excursion limits of driver is reached (what I have been discussing). That type of distortion is clearly audible and not subject to debate.

 

[amirm]

Amir deliberately listens at maximum volume, bordering on hearing damage, to give distortion at least a chance

I listen until I find limits of the speaker. No different than a car reviewer pushing a car to its limits in handling, acceleration, and breaking. But maybe you complain to them as well that you are a Sunday driver and only go 35 mph to local store 5 minutes away so their tests are useless.

 

[amirm]

According to below study, exposure to 63Hz low‐frequency octave-band noise at 90 dB SPL for 6 h leads to significant hearing loss over an extensive range from low to high frequencies.

Extensive hearing loss induced by low‐frequency noise exposure - PMC

With little attention given to low‐frequency traffic noise and our understanding that cochlear function may be highly susceptible to low‐frequency noise, there is an urgent need to determine traffic noise‐induced hearing loss (NIHL), not only the ...

pmc.ncbi.nlm.nih.gov

If your listening habbits include this:

"The current study aims to investigate the potential for extensive hearing loss induced by exposure to 0.063 kHz octave band noise (OBN), which is an important component of low‐frequency traffic noise."

you have more serious issues than what distortion means in a speaker.

And let's hope you are not doing this:

"Exposure to 0.063 kHz low‐frequency OBN at 90 dB SPL for 6 h leads to significant hearing loss over an extensive range from low to high frequencies."

6 hours of continuous noise???

 

[amirm]

. But many swear by SINAD.

Nope. We praise engineering excellence as we should. What you say is a talking point by anti-measurement folks. I don't think a review of high sinad device exists without someone saying what you said.

 

[amirm]

Even an amplifier with a SINAD of 90 is completely transparent and in no way inferior to an amplifier with a SINAD of 120.

What if sinad was 30? Or 20? Or even zero? This is what we are talking about with transducers on regular basis. Why not strive for better?

 

[amirm]

Just listened to your proposed track again (Fading Sun from Terje Isungset) on my Palmer Orbit 11s. No sub, 5x7m room, 2.5m listening distance. I cleary hear the 30Hz rumble that my smartphone analyzer shows.

If you can't hear nearly 50% distortion, I don't know what to tell you:

Other than turn it up until you clearly hear it. Then back down the volume and you will perceive the distortion at lower level. Classic way to train your ears.

 

[illusonic]

If your listening habbits include this:

"The current study aims to investigate the potential for extensive hearing loss induced by exposure to 0.063 kHz octave band noise (OBN), which is an important component of low‐frequency traffic noise."

you have more serious issues than what distortion means in a speaker.

And let's hope you are not doing this:

"Exposure to 0.063 kHz low‐frequency OBN at 90 dB SPL for 6 h leads to significant hearing loss over an extensive range from low to high frequencies."

6 hours of continuous noise???

Noise was around 60Hz and hearing loss was wideband. This is the opposit of what you claimed. So I would advise to be carefull with extreme bass levels.

 

[illusonic]

A good blog from Purifi in audibility of distortion from woofers: https://purifi-audio.com/blog/tech-...o-whats-the-point-of-low-distortion-drivers-4

There are audio samples there as well. It follows the same type of analysis/conclusions as this study and that if Fielder:

"The science was right: even gross bass distortion is remarkably difficult to hear. The fine print says yes, provided it stays in the bass."

Note that this distinct from the type of distortion where the excursion limits of driver is reached (what I have been discussing). That type of distortion is clearly audible and not subject to debate.

The Purifi examples are the perfect demonstration of (T)HD vs. IMD. First example is HD due to KMs-nonlinearity. Second track is IMD due to BL-nonlinearity. IMD is much worse than HD.