• 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!

Audible threshold of distortion.

audio_tony

Addicted to Fun and Learning
Joined
Feb 24, 2019
Messages
664
Likes
802
Location
Leeds, UK
I was watching a Youtube video presentation by Rob Watts of Chord, and was rather surprised to see he claims that "Distortion from -130dB to -150dB is surprisingly easily audible".

That's quite some feat for someone of his age.

Full video here:

Screenshot from video:
audible-distortion.png
 
Fortunately, I am not blessed with the ability to perform such astounding feats of hearing.
 
Any current Chord available for measurement?
 
Fortunately, I am not blessed with the ability to perform such astounding feats of hearing.

I think you need to send him a copy of your shout-o-meter...

I found that early on in my wanderings here, and it might have been the single most 'light bulb went on' moments I've had on the forum.

He might need more help than simple science can offer however...
 
what is the basis for his assertion?

did he cite any published work?

No, I'm not gonna waste time watching his video...
 
Let's say you're listening at 85 dB SPL, which for me would be rather loud.

Then we would be talking distortion products at -45 dB SPL or lower. That's like 40 dB below even the most optimistic human hearing threshold, or maybe 20 dB below a dog's.

You could be playing just the distortion products in complete silence... and no-one would be the wiser that anything was even there.

If you had a 85 dB / W / m speaker, that's 0.9 µV at the amplifier output or less. Even when using a very low-noise power amplifier with barely 20 µV of output noise (that's an SNR of 103 dB ref. 1 W / 8 ohms), that would mean an SNR of -27 dB or worse.
I reckon you could stick your ear to the tweeter and barely hear the noise. Anything 27 dB lower? Good luck.
 
This opens up entire new realms of audio gullibility.
 
[Disclaimer: the following are just some vague musings here from a random nobody on the internet etc. etc. blah blah blah. :D And I haven't even glanced at a calculus equation in around 30 years and I present nothing technical here anyway, so anyone with actual expertise in such matters, please save your corrections and criticisms for Rob Watts and his claims, since he seeks profit and glory while at most I might get a few extra 'likes' on a forum. haha! :facepalm: ]


So, even if Chord had an anechoic chamber that was vastly superior to the recent Microsoft one (-20.6 dB, I think), and pretending human hearing was capable of such feats as hearing and differentiating -130 dB or -150 dB, what transducer is capable of producing these sounds at such incredibly low levels of distortion?

Random motion of air molecules is something around -23 dB? I think I read somewhere (is that vague enough? lol) that sets an absolute limit on anechoic chambers on Earth, so Microsoft's chamber is especially remarkable for getting so close to that level.

Can he diagnose an infection by hearing the difference as more white blood cells are in his bloodstream before his temperature begins to climb? He must surely be able to hear the difference between blood flow with his normal cell ratios as compared to when he has the slightest of infections. (I assume since leukocytes are more than twice the size of erythrocytes, it shouldn't be a problem for him to hear differences. Perhaps he can even distinguish even finer differences, such as the sounds from changes in platelet ratios, or even virions?) That should be trivial considering his claims about his hearing acuity. :rolleyes:

Sadly, the rising noise floor from the increasing sound of his own bones and joints creaking as he continues to age will reduce his abilities, even if his hearing itself is miraculously unaffected by the ravages of time. :facepalm:
 
pretending human hearing was capable of such feats as hearing and differentiating -130 dB or -150 dB,

That's getting into Chuck Norris territory...
 
Hearing limits in mammals and birds are pretty close to the physical limits

not counting 'civilized' humans who are exposed to street noise, crawl into speakers at rock concerts, etc.
 
The Lazenby article gives some calculations based on some previously published work. Titles of the references articles are not given so it cannot be determined exactly what he relies on.

It may be based on the Fletcher Munson work, which deals solely with sine waves. Actual tests of more complex waveforms would be nice.

But, it's something - despite the lack of measurements.

Thx for posting it.
 
It may be based on the Fletcher Munson work, which deals solely with sine waves. Actual tests of more complex waveforms would be nice.

To get a first, rough idea of how well your own hearing can detect distortion, I can really recommend the Klippel-Listening test.
And you can discover for yourself what a difference it makes when artificial signals or real music are heard (depends of course on the quality of the recording).

1585772871062.png
1585772883903.png
 
Last edited:
Absurd. Mental note: Never buy a product from Chord.
To be fair, I rather have manufacturers like Chord that try to create products that measure well and thus perform well than brands like PS Audio or Totaldac that just dump shit on the market. Sure you pay too much for Chord, but at least you get a product that is not broken.
 
To answer the OP's question (which may have been rhetorical), from what I read and my own self experiments, distortion around -40 dB or 1% is in the ballpark for trained listeners having normal hearing on high quality recordings and playback equipment. Plus or minus, depending on the source material to which the distortion is applied. This is consistent with results @ctrl posted above. You might get down to around -50 dB or 0.3% under ideal conditions, like a 2-tone test in an ideally quiet environment. Personally, I would consider -60 dB to be safely transparent, and some speakers and headphones have distortion this low at least in the mids & treble where our perception is most sensitive to it.

That said, like @ctrl , I can hear distortion at lower levels in a 2-tone test than with most music. I wonder if there is something even more revealing than a 2-tone test, that enables us to hear distortion at lower levels? Put differently, what is the most revealing type of signal to use for testing perception thresholds for low level distortion?

PS: a few years ago I heard something "off" in recordings I made with a Zoom H4 portable recorder. Turned out to be spurious tones at -46 dB from full scale. I think the reason I could hear them at such a low level is because they were steady tones not correlated to the music, making them "stand out" perceptually; distortion like HD and IMD is correlated to the signal, which can mask them, making them harder to hear.
 
Last edited:
Actual tests of more complex waveforms would be nice.
At later part of the article he did say the results reasonably matched with music study done by olson. But yes more modern test would be nice. And with my rudimentary undestanding of these things I think since speakers have more distortions it is good to have low distortion amplifier to feed clean signals to speakers.
Regards.
 
Back
Top Bottom