• 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!

SINAD vs Dynamic range vs Human hearing limitations

EB1000

Senior Member
Joined
Jan 8, 2020
Messages
484
Likes
579
Location
Israel
Hi

As you know, SINAD represents THD+N in dB, so a SINAD of -60dB should correspond to the human limit of 0.1% THD. DR on the other hand, is the maximum ration (in dB) between the highest and lowest sound intensity. But these two indices are correlated, as the higher the noise floor, the lower the the DR and SINAD. But human ear limit for DR is around 110dB, so how can we hear 110dB if we can't tell the difference between -80 and -60dB SINAD.

Can an audio device have a SINAD of 70dB and a DR of 100dB for example?

Thanks
 

dfuller

Major Contributor
Joined
Apr 26, 2020
Messages
3,339
Likes
5,063
As you know, SINAD represents THD+N in dB, so a SINAD of -60dB should correspond to the human limit of 0.1% THD.
I dunno where you got that from. You might not hear it as distortion, but I'd bet good money trained listeners can hear below 0.1% distortion.

Can an audio device have a SINAD of 70dB and a DR of 100dB for example?
Sure, it would have much higher distortion than noise.
 

MrPeabody

Addicted to Fun and Learning
Joined
Dec 19, 2020
Messages
657
Likes
942
Location
USA
You've asked a thought-provoking question, but the way I would approach it is by first asking questions about whether and how the concept of dynamic range is applicable to the output of an audio device. The output of an audio device is a signal, and for any signal, the signal-to-noise ratio is well-defined. One scenario where the concept of dynamic range is well-defined (and fully distinct from signal-to-noise ratio) is the scenario of a digital encoding format, where dynamic range is defined as the ratio of the maximum possible signal to the weakest possible non-zero signal, expressed in decibels. In order to extend the concept of dynamic range to analog encoding and transmission formats, it is generally necessary to refer to some typical or de facto value for the noise floor, using this in place of the weakest possible non-zero value. Consequently, in the context of analog encoding and transmission formats, the distinction between dynamic range and signal-to-noise ratio is not a clear distinction.

You have added another layer of obfuscation: the messy business of the level of distortion that may be audible to people. The best answer to your question is "no". In order for the answer to be "yes", there will need to be a non-arbitrary way to assign a theoretical minimum noise value to the analog signal, lower than the actual noise in the signal. Anyone who is sufficiently motivated to think up a way to do this will likely succeed, but it will be a non-productive exercise nevertheless.
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,445
Likes
15,781
Location
Oxfordshire
I dunno where you got that from. You might not hear it as distortion, but I'd bet good money trained listeners can hear below 0.1% distortion.
I would not expect anybody to, based on my experiment, which is admittedly only on me, but I have done it more than twice.
If I listen to a tone, or music, at my normal listening level and turn the level down by 60dB I can either barely hear it, or not at all, depending on frequency.
In order to be able to hear a distortion product at 0.1% I would have to hear something I can't hear on its own in a quiet room whilst music is playing at my normal listening level.
I would take a lot of convincing that I would be able to do that.
I would certainly want any background noise to be lower than that but distortion? Nah.

Also whilst it is true that humans (depending on the individual) can hear from 0dB to 120dB pretty well nothing in real life has that dynamic range, certainly none of the music I know of.
On top of that if you have been listening to something at 120dB it will be seconds, hours or maybe never before you could hear something at 0dB, depending on how long you were exposed to 120dB, so this sort of dynamic range is of academic interest rather than a practical requirement for music.
I suppose it is possible to imagine a film soundtrack containing birdsong in one part of the film and an aircraft taking off elsewhere in the film but pound to a penny they won't be recorded, or expected to be reproduced, at the dynamic range of real life.
 

q3cpma

Major Contributor
Joined
May 22, 2019
Messages
3,060
Likes
4,416
Location
France
what to look for here? If this can be simply summarised
The higher the harmonic the lower the masking. See this pure tone masking chart from Psychoacoustics: Facts and Models (3rd edition) by H. Fastl & E. Zwicker.
index.php


And that's for pure tones, real music has even more masking (including temporal) going on. But some example to explain what I do:
index.php

index.php


The MOTU M2 has a higher SINAD including a slightly higher SNR, but the Babyface has nothing of note after H3 while the M2 keeps going on. Both are more than probably transparent, but if I were to score them, the Babyface is better on the distortion side of things.

Or for a more extreme example:
index.php
 

Attachments

  • pure sine masking.png
    pure sine masking.png
    34.2 KB · Views: 1,204
Last edited:

Head_Unit

Major Contributor
Forum Donor
Joined
Aug 27, 2018
Messages
1,341
Likes
688
I'd bet good money trained listeners can hear below 0.1% distortion.
I won't speak to that, but as an outgrowth of doing rub/buzz tests on woofers in free air I can say that 0.1% is sure the heck audible with low frequency tones-we had to get a cleaner signal generator to avoid the woofers all sounding bad. Note that in this situation there is pretty much no masking going on.
 

Mnyb

Major Contributor
Forum Donor
Joined
Aug 14, 2019
Messages
2,645
Likes
3,624
Location
Sweden, Västerås
You have to look at the graph and take inte account that there is two components at -60dB sinad I would look if the results also a have an excessive noise component.

And what component is it ,your phono stage ok vinyls is a bit noisy and the noise follows your volume control.

Is it your power amp with a noise at only -60dB relative to max power :) this will be very noisy indeed
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,445
Likes
15,781
Location
Oxfordshire
Is it your power amp with a noise at only -60dB relative to max power :) this will be very noisy indeed
Yes, I allowed 20dB of margin on my "test of me for me" which is only valid for me of course, so whilst I am pretty sure I can't hear 0.1%, ie -60dB of distortion 80dB would be required for noise and the real effect of that is different between a power amp, where tho noise is there all the time, compared to a DAC where the noise is reduced by the volume control of the preamp so I would expect a noisy power amp to be more audible than a DAC with the same snr, sadly and amusingly getting a low snr from a power amp is much harder than from a DAC :)

@Pluto demonstrated noise shaping to a stream of enthusiasts at the Scalford show a few years ago.
He took a well recorded pretty high dynamic range 24/96 file (from Eric Whitacre's Water Night album) and we got to listen to it as recorded, as a 8-bit file and as 8-bit with noise shaping.
Without noise shaping the background noise at -42dB was very obvious in the quieter parts but with noise shaping the difference between 24/96 and 8/96 was surprisingly small - barely audible.

This got abuse from a lot of the enthusiasts, some even got really angry and absolutely didn't believe what they were listening to was 8-bits - both the people who were supportive of "high res" digital and those who were sure analogue was superior to CD because 16/44 "isn't enough" were both outraged.
Anyway I think it does me good to get a "feel" for what the numbers mean by listening. I suspect it would help others, it isn't that hard, actually, then one has a much more realistic perspective on what distortion, noise and dynamic range means to oneself in actual sound. It "calibrates" the numerical values of the measurements in the mind.

I accept that this is only valid for me and I would encourage anybody to do the same for themselves, since they can easily do so.
 

bigjacko

Addicted to Fun and Learning
Joined
Sep 18, 2019
Messages
721
Likes
359
Does the human DR around 110dB mean the smallest sound human can hear is 0dB and loudest sound human can hear is 110dB without damaging ear? If that is the case then problem is easy, when you hear loud sound the small sound gets covered up so you can't hear them.
 

Mnyb

Major Contributor
Forum Donor
Joined
Aug 14, 2019
Messages
2,645
Likes
3,624
Location
Sweden, Västerås
Yes, I allowed 20dB of margin on my "test of me for me" which is only valid for me of course, so whilst I am pretty sure I can't hear 0.1%, ie -60dB of distortion 80dB would be required for noise and the real effect of that is different between a power amp, where tho noise is there all the time, compared to a DAC where the noise is reduced by the volume control of the preamp so I would expect a noisy power amp to be more audible than a DAC with the same snr, sadly and amusingly getting a low snr from a power amp is much harder than from a DAC :)

@Pluto demonstrated noise shaping to a stream of enthusiasts at the Scalford show a few years ago.
He took a well recorded pretty high dynamic range 24/96 file (from Eric Whitacre's Water Night album) and we got to listen to it as recorded, as a 8-bit file and as 8-bit with noise shaping.
Without noise shaping the background noise at -42dB was very obvious in the quieter parts but with noise shaping the difference between 24/96 and 8/96 was surprisingly small - barely audible.

This got abuse from a lot of the enthusiasts, some even got really angry and absolutely didn't believe what they were listening to was 8-bits - both the people who were supportive of "high res" digital and those who were sure analogue was superior to CD because 16/44 "isn't enough" were both outraged.
Anyway I think it does me good to get a "feel" for what the numbers mean by listening. I suspect it would help others, it isn't that hard, actually, then one has a much more realistic perspective on what distortion, noise and dynamic range means to oneself in actual sound. It "calibrates" the numerical values of the measurements in the mind.

I accept that this is only valid for me and I would encourage anybody to do the same for themselves, since they can easily do so.

Like you I'm fascinated why the DAC the probably best performing component in anyone's hifi gets such much attention and are exchanged so often to next years model ?
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
The reasons are:

Specs keep increasing.
Support for even higher bitrates are increasing.
People think it matters.
People are used to buying new 'computer and phone' stuff every few years.
Reviewers keep telling us new stuff is 'better'.
Fear of missing out on max. sound enjoyment.
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,445
Likes
15,781
Location
Oxfordshire
Does the human DR around 110dB mean the smallest sound human can hear is 0dB and loudest sound human can hear is 110dB without damaging ear? If that is the case then problem is easy, when you hear loud sound the small sound gets covered up so you can't hear them.
More or less.
It is a bit more complicated because it is frequency dependant and because loud sounds do temporary or permanent hearing damage.
This means a bass sound actually has to be pretty loud before you can hear , 60dB at 30 Hz is the threshold of hearing and sounds about as loud as 0dB at 1kHz.
Also the range is a bit more, though it depends on the individual, and if you have been listening to 110dB not only does it obviously mask most quieter sounds but it will be seconds, minutes, hours or maybe even never before you could hear something at 0dB depending on how long you had been abusing your hearing.
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,445
Likes
15,781
Location
Oxfordshire
The reasons are:

Specs keep increasing.
Support for even higher bitrates are increasing.
People think it matters.
People are used to buying new 'computer and phone' stuff every few years.
Reviewers keep telling us new stuff is 'better'.
Fear of missing out on max. sound enjoyment.
And the problem is people aren't informed on what is actually important.
 
Top Bottom