• 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!

Audibility thresholds of amp and DAC measurements

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Hi Don,

What I actually meant is if your source music being played is recorded at 0 dBFS or close to it (like most pop music). The volume level of the amp would be calibrated to output 96 dB when your source music is recorded at 0 dBFS. Now, 96 dB is very loud, and I usually listen at much lower levels, but I used it for argument's sake.

I am not sure the question, sorry, juggling too many things today. If your DAC adds no distortion, and assuming by "DAC" you mean the actual DAC plus the filters and amplifiers that deliver the analog signal to you rears, then if you set the DAC to full-scale (0 dBFS) and the SPL at your ears is 96 dB then the IMD is at 0 dB -- again assuming no other distortion sources (unlikely). If you turn the DAC down to -30 dBFS the output SPL would be 66 dB and the distortion products will be about 76 dB below that.

The distortion in the recording depends upon the chain up to and including the ADC.

Generally if you turn down the volume the distortion goes down but the noise stays flat thus SNR goes down but THD/IMD improves.
 

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,687
Likes
4,068
Hi,
What does it mean for an amplifier to have an IMD of 0.05 %?
Bad or ok?
I read somewhere that IMD was worse than THD to the ears.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
What does it mean for an amplifier to have an IMD of 0.05 %?

http://www.sengpielaudio.com/calculator-thd.htm

1553085508489.png
 

sergeauckland

Major Contributor
Forum Donor
Joined
Mar 16, 2016
Messages
3,440
Likes
9,100
Location
Suffolk UK
I
Hi,
What does it mean for an amplifier to have an IMD of 0.05 %?
Bad or ok?
I read somewhere that IMD was worse than THD to the ears.
Harmonic distortion is just that, harmonically related to the fundamental, and so can add to overtones of musical instruments without greatly changing their timbre. As such, not only do we tolerate harmonic distortion, but some even find it pleasing. Something like a violin, or better still, a flute can be thought of as a sine wave generator with a lot of harmonic distortion.

Intermodulation distortion is what happens when two (or more) frequencies modulate each other, and create tones which are not harmonically related to any of the tones. These extra tones are generally thought unpleasant.

However, as HD and IMD are created through the same process, that of non-linearity of the amplifier, one can't have one without the other.
An amplifier with 0.05% IMD will probably have much the same order of magnitude of THD. Whether this is audible or not, depends on many factors, but if 0.05% is a worse case, then nothing to worry about. If a best case, then the reality could be a lot worse and become audible.

S.
 

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,687
Likes
4,068
Coming back to this thread's first post:

THD, IMD: -66 dBFS / 0.05%
Noise: -85 dBFS / 0.005%
SINAD: 85 dB

If you have an amplifier which meets:
THD, IMD: -66 dBFS / 0.05%
Noise: -85 dBFS / 0.005%

then it won't meet
SINAD: 85 dB.

So there's a problem here. It seems like lenient SINAD doesn't mean anything.
 
OP
F

flipflop

Addicted to Fun and Learning
Joined
Feb 22, 2018
Messages
927
Likes
1,240
If you have an amplifier which meets:
THD, IMD: -66 dBFS / 0.05%
Noise: -85 dBFS / 0.005%

then it won't meet
SINAD: 85 dB.

So there's a problem here. It seems like lenient SINAD doesn't mean anything.
SINAD doesn't mean much to me, which is why it wasn't included in the original draft. I added it later per request.
It's more useful to look and distortion and noise separately when possible.
 

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO

HuskerDu

Member
Joined
Mar 27, 2019
Messages
62
Likes
44
Location
Houston
This is really, really good info. At the risk of taking the fun out of it, is there a version in Excel somewhere? I want to add a second row for PRICE on everything that meets "strict." Why look at extraneous features on pointlessly expensive gear. Another stack-rank near the "lenient" boundary would also take the "mystique" off the price for perfection, I would think. I'm having a hard time finding a better/dollar DAC than a handful of close-out Chromecast Audio pucks, at $15 each. (But heck, any DAC under $150 is essentially "free" compared to any worthwhile speaker.)

Um, can you do the same "schema" for room noise... :D Puhlllllleeeeeease...? (I'm prepared to help!)

P.S. I've read the room noise I and II posts. (I don't see a third, or I missed external ref...?)
 
OP
F

flipflop

Addicted to Fun and Learning
Joined
Feb 22, 2018
Messages
927
Likes
1,240
At the risk of taking the fun out of it, is there a version in Excel somewhere?
No, I don't have any related content that wasn't included in the opening post.
I want to add a second row for PRICE on everything that meets "strict."
You would have to pick a single measurement then. No device falls below strict thresholds in all of them.
Um, can you do the same "schema" for room noise... :D Puhlllllleeeeeease...? (I'm prepared to help!)
It would have to be for another thread. This one deals with amps and DACs only.
 

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,687
Likes
4,068
In THD+N vs frequency plots, it's not always possible to tell the distortion and noise apart, so the lenient noise threshold of -85 dBFS / 0.005% THD+N will be used.

View attachment 18973
Behringer doesn't manage to stay below the lenient threshold. The strict threshold is off the chart, so it hasn't been marked.

Hi,
In the text you speak of 0.005%, but in the graph the line is at 0.05%.
 

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,687
Likes
4,068
Thanks. The text was right, the graphs were wrong. They have now been updated. The first one has also been replaced with a better example.

The Behringer would have seemt to be a big piece of crap if you had kept the graph but with the corrected green line!!
 

abm0

Active Member
Forum Donor
Joined
May 12, 2019
Messages
126
Likes
58
Frequency range of human hearing
Humans cannot hear sounds of every frequency. The range of hearing for a healthy young person is 20 to 20,000 hertz.
For the eardrum-targeting headphone use case this should be the end of the story, but I wonder how much relevance it has for listening through bone-conduction headphones (skullphones?) or through speakers that there's some capacity for humans to "hear" things via ultrasonic vibrations, as in https://science.sciencemag.org/content/253/5015/82

Frequency response, channel balance
Going back to NwAvGuy's amp guidelines, he recommends a maximum of 0.5 dB deviation (from 0) in the frequency response.
Psychoacoustics: Facts and Models by Hugo Fastl and Eberhard Zwicker is not a very quotable book, but on pages 180-181 it makes it clear that a change in SPL of less than 0.2 dB can be heard by humans.
In 'Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms' Floyd Toole says the following: "The simplest deviation from flat is probably a spectral tilt. There is some evidence that we can detect slopes of about 0.1 dB/octave, which translates into a 1 dB tilt from 20 Hz to 20 kHz — not much."
0.1 dB is therefore the strict limit.
Mhhh, not sure the tilt case is measuring the exact same thing though. When I went looking for info on this I found it reported that 0.41 dB is the loudness difference where subjects' guesses on whether they heard a difference or not become no better than chance (could be untrained listeners though, so I could see 0.2 being the right value for trained listeners): https://asa.scitation.org/doi/abs/10.1121/1.1916528

When I tried this myself at 0.5 dB I could guess the difference with a 77% success rate out of 200 tries of 3 choices each (louder, quieter, same), so well over the blind-chance 33%: https://www.audiocheck.net/blindtests_level.php?lvl=0.5
(This was in the silence of night, using the Superlux HD668B.) Don't remember how I did for 0.2, but at 0.1 dB I'm almost sure I was no better than chance no matter how much I tried, so at least for me this threshold is not something that's going to make me spend more money on reproduction equipment. :)

Then again one thing that maybe needs mentioning in the context of any thresholds guide like this is that the more up-stream a device is (particularly if it's before all the amplifiers), the more it has to account for what might happen after it in the chain and perform somewhat beyond the thresholds of hearing just because of the additional degradation that might be introduced downstream.
 
Last edited:

Kvalsvoll

Addicted to Fun and Learning
Audio Company
Joined
Apr 25, 2019
Messages
878
Likes
1,643
Location
Norway
A simple and robust method for showing the practical consequences on sound quality when measured performance criteria are met, could help spreading the message and bring things into context.

DACs and good amplifiers have been transparent for decades now, yet we still have a situation where there is an established opinion in the audiophile world that very significant differences in sound quality exist, a story that simply ignores measurements performed on actual devices and the scientific knowledge of how those measurements correlate to audible differences.

We know a good dac has no "sound", we know why people still perceive differences in sound. The science behind all this is known, and has been for a long time.

Some people will never accept this, some have financial interests in keeping this misconception alive. But some will want to learn and know the facts.

One way to show it is to provide sound samples recorded from playback through the test object. The recorded sample can then be compared to the original, by everyone, on their own system, speakers or headphones. If there is a difference, that can be confirmed in an abx test, quite easily.

I did this in the amplifier test some years ago. On music program material, it turns out it is very difficult to hear any difference between very different amplifiers, and even comparing to the original there is to this day not one single report of anyone being able to hear and verify a difference.

For people who are curious and want to learn, this is a useful exercise. Looking at some measument graphs is fine, but when you can hear what this really means, that can be a real ear-opener.
 

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
[...] we know why people still perceive differences in sound. The science behind all this is known, and has been for a long time.

Some people will never accept this[...].

No sure everyone really knows why, so is it related to "loony-bin audiophiles" or is it simply psiyhoacoustics?

"Hearing is not a purely mechanical phenomenon of wave propagation, but is also a sensory and perceptual event; in other words, when a person hears something, that something arrives at the ear as a mechanical sound wave traveling through the air, but within the ear it is transformed into neural action potentials. The outer hair cells (OHC) of a mammalian cochlea give rise to an enhanced sensitivity and better[clarification needed]frequency resolution of the mechanical response of the cochlear partition. These nerve pulses then travel to the brain where they are perceived. Hence, in many problems in acoustics, such as for audio processing, it is advantageous to take into account not just the mechanics of the environment, but also the fact that both the ear and the brain are involved in a person’s listening experience."
 

Kvalsvoll

Addicted to Fun and Learning
Audio Company
Joined
Apr 25, 2019
Messages
878
Likes
1,643
Location
Norway
No sure everyone really knows why, so is it related to "loony-bin audiophiles" or is it simply psiyhoacoustics?

"Hearing is not a purely mechanical phenomenon of wave propagation, but is also a sensory and perceptual event; in other words, when a person hears something, that something arrives at the ear as a mechanical sound wave traveling through the air, but within the ear it is transformed into neural action potentials. The outer hair cells (OHC) of a mammalian cochlea give rise to an enhanced sensitivity and better[clarification needed]frequency resolution of the mechanical response of the cochlear partition. These nerve pulses then travel to the brain where they are perceived. Hence, in many problems in acoustics, such as for audio processing, it is advantageous to take into account not just the mechanics of the environment, but also the fact that both the ear and the brain are involved in a person’s listening experience."

What is perceived can be different from the actual physical sound, because hearing is affected by other stimuli and memory and expectation, and everyone are subject to this. To the extent that it does not help that you know how this works, you will still be affected and may hear things that actually does not exists, and what you think you hear is percieved as real.

This is caused by the way hearing works. Acoustic sound enters the ear and physical hearing mechanisms, then there is a psychoacoustic part that interprets this acoustical-mechanical input - frequency/tone, loudness, location/direction. Then there is a psychological part, where the brain tries to make sense of the sound - what the sound is, decoding of spoken words. This last part searches for solutions that are most likely, and uses other information in this process. Such as if a spoken word is partly masked, the missing details will be filled in so that it matches something that is likely. This comes into play when listening for differences that are barely audible, or not present at all. If there is no difference, the brain fills in using all sorts of other non-acoustical cues, and creates an illusion of a sound that simply does not exist.

I am not an expert on hearing and perception and psychology, but someone is, ask them of you want to learn more about how hearing actually works.

To experience how this works, you can do some experiments. Find something you know from a technical analytical perspective is likely to sound the same, but still you suspect there is a difference. Listen to both, find how they sound different, continue to focus on those differences. When you are very sure you have identified how they sound different, now try to identify which one is playing in a abx. The differences suddenly disappear. Now, how can this be - either your hearing had you fooled, or blind testing does not work.

One practical way to test this:

- Install foobar with foobar abx plugin.
- Find a music sample in flac/lossless.
- Encode sample in 320k lossy format, use a newer codec, mp3/aac/ogg.
- Compare flac vs 320k until you are sure you can hear the difference.
- Then try to abx the files, to get a reliable verification.

The exact same method can be used to do controlled listening tests of electronic devices - dacs, amplifiers.
 

watchnerd

Grand Contributor
Joined
Dec 8, 2016
Messages
12,449
Likes
10,408
Location
Seattle Area, USA
I would love to see these types of measurements for phono stages, particularly MC phono stages.

Mine is digital, and at -20 dB gain I can't hear anything unless my ear is at the tweeter, but at 0 dB the noise floor is obvious (if music isn't playing).
 

Miguel

Member
Joined
Jan 22, 2019
Messages
10
Likes
15
I have a question about the thresholds throughout the frequency range, are them really a flat line or should them be some kind of curve. For example does a THD+N of 0.005% have the same importance at 1kHz or at 50Hz. Thanks.
 
Top Bottom