The -120 dB wasn't for the "average" ear. It was for a highly trained expert using special listening techniques.
Rick "way out on the tail of that distribution" Denney
-
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
- Thread starter flipflop
- Start date
Rick "way out on the tail of that distribution" Denney
But you'd be able to hear below 1% distortion in sine tones, but that's not the same thing as the music you mention, but I reckon you would with sine tones. I tested myself the other day and could hear 0.5% distortion in all of the tones I tested and in one tone at 125Hz I could hear 0.1%. But yes, not the same as with music as I know I wouldn't be able to detect those levels in music, except maybe on only specific segments of some tracks, but that's a big maybe and also a big don't know!And added to that is the requirement for engaging special listening techniques to get anywhere near Amir's most demanding threshold, including gainriding the tails of reverberation and listening to floor-level sounds at high volume. That's how Amir can distinguish between recordings with low levels of distortion and noise.
When I have noise at -90 dB, I have to put my ear right up to the tweeter to hear it. That's the equivalent of listening at high volume--if my ear is there when full-scale signal appears, it's gonna hurt.
But, yes, the 120 dB SINAD threshold exceeds any that is found in the literature, and least that has been cited where I've seen it. Published thresholds are much lower, even for listeners able to adjust the volume during the test. For controlled-level tests, it's lower still. A search on "thresholds" in the title of threads yield a page and a half, and many of these threads include citations. But here is the best attempt at a summary:
Audibility thresholds of amp and DAC measurements
Introduction A recurring theme on ASR is whether or not the various measured qualities of the devices are audible. In this post, I'll present some clear and visual thresholds for when those imperfections can be considered a potentially audible concern. I will not explain the basics of amp/DAC...
Rick "who, with music and without gain-riding, can't hear harmonic distortion below 1% in blind testing" Denney
Well, "in practice," very few recordings have intended signal anywhere near 96 dB full scale. I'd be that the range is more like 20-60 dB, with 60 dB being live acoustic performances with large ensembles and energetic program material. Most of what I hear these days is so processed, you can't tell anything about it. They gate off the noise floor before applying the compression algorithm to avoid noise pumping, so what we might be hearing in the tail of that reverberation is more plastic than meat, like Darth Vader's helmet.
One important distinction is where the noise floor of equipment is in the gain structure. A noise floor of, say, -50 dB SPL in a source component will be at a very hearable 60 dB SPL when the peaks are playing at a loud but not painful 110 dB SPL. Source signals get up to ~40 dB of amplification (more than that for microphones and moving-coil cartridges) before arriving at the speakers. But power amplifier noise is already at listening level. Thus, a power amp with a noise floor of -80 dB is not wasting the capabilities of a source component with a noise floor of -120 dB, if we listen at full scale. That is, of course, as long as the amp is not spectrally steered by the speaker loads, or underpowered and clipped on the peaks, no matter how forgivingly it handles clipping.
Rick "power amplifiers with SINADs greater than 90ish are fun to talk about, but that's all" Denney
You are right, but only if the tones are below about 5000 Hz (and maybe WELL below 5000 Hz). It's easier to hear when the signals are in the low hundreds, like the 125 Hz you mention.
As a tuba player, I'm sensitive to harmonic distortion that makes tubas sound like euphoniums or euphoniums sound like flügelhorns or French horns sound like trombones. But that takes harmonic distortion levels higher than -40 dBFS. I can't think of any electronics I've ever owned that was bad enough to cause that problem for me (well, maybe except for some old cheap car stuff), but speakers are a whole other thing. Distortion levels of 10% (-20 dB) are not that uncommon for small speakers played loudly, particularly when equalized to bring the bass up to correct levels. I once went to a Magnolia to listen to speakers, and was surprised at how many of them made it sound like Chuck Daellenbach (of the Canadian Brass) was playing a euphonium when they were cranked up. Say what you will about my old Advents, they never did that.
Rick "who buys DACs based on features, price, and form factor, not SINAD" Denney
Great. Are the details published in the literature somewhere? I would love to read the paper. Thanks!
Nothing new there, at least that I could see. Thanks for the link though. To kind of get back to the point I was trying to make maybe I could try to explain this way: Absolute limits of hearing are exclusively measured with single tones for very good reasons. They are of course estimates of an average level for a population, not the best or the worst. Multiple tones are disallowed because the presence of other tones can affect the perception of tones below the absolute threshold. Why? What is the physics? Please take for example the waveforms below. Let's say the one small amplitude in the middle is below the absolute threshold. Now add the higher amplitude tone above it. The sum of the two tones is shown in the lower waveform. The peak pressure level of the combined tone is higher than either of the single tones alone. It is possible the peak level of the low amplitude signal in the middle is now at a pressure level at least some of the time where it is above the threshold pressure level. That's why only single tones are used for absolute thresholds. Its because adding tones can affect the maximum absolute SPL, and thus shift detection of low level tones.
View attachment 343416
Hope that helps!
All it helps me to understand is that you did not actually read or listen to the actual tutorial. You're showing multiple frequencies far removed from each other, which shows without doubt that you do not even begin to understand the tutorial. As said before, those widely separated sounds are detected on different parts of the cochlea, and as such do not interact.
Your claims that there's "nothing new I can say" simply shows that you didn't watch and understand it. I have some doubts you even watched the tutorial.
You get one more chance. ONE more chance.
DualTriode
Addicted to Fun and Learning
- Joined
- Oct 24, 2019
- Messages
- 903
- Likes
- 594
You are neglecting sidebands caused by intermodulation.
Two Tone Intermodulation Distortion can be clearly audible.
Take a look at the Purifi web page for discussion and sample test plots. Also see the Puirifi driver data sheets for Two-Tone for IMD data.
Different parts of the cochlea, not so true.
DT
Last edited:
How do you know what I'm ignoring? Hint: You have no idea. Please read the hearing tutorial and get back to me how that's a germane issue?
IF the IMD gets to the cochlea, handle it like the signal it is. Then you're done. Now, if you're talking about dangerous levels, yes, all sorts of things happen. You shouldn't be doing that, too.
DualTriode
Addicted to Fun and Learning
- Joined
- Oct 24, 2019
- Messages
- 903
- Likes
- 594
Hello,
Neglecting vs ignoring, we will let someone else score that one. I believe that it is intent. Perhaps it is a question for the next Wechsler Adult Intelligence Scale.
As far as complex tones and Beat Frequencies go.
Score that point for @Wanman.
Beat Frequency Formula: Definition, Formula Derivation, Solved Examples
In sound, we hear the beat frequency as the rate at which the loudness of the sound varies whereas we hear the ordinary frequency of the waves as the pitch of the sound. Here we discuss the concept of beat frequency and beat frequency formula in detail.
www.toppr.com
Thanks DT
Last edited:
Hello,
Neglecting vs ignoring, we will let someone else score that one. I believe that it is intent. Perhaps it is a question for the next Wechsler Adult Intelligence Scale.
As far as complex tones and Beat Frequencies go.
Score that point for @Wanman.
Beat Frequency Formula: Definition, Formula Derivation, Solved Examples
In sound, we hear the beat frequency as the rate at which the loudness of the sound varies whereas we hear the ordinary frequency of the waves as the pitch of the sound. Here we discuss the concept of beat frequency and beat frequency formula in detail.
Thanks DT
As you still haven't addressed the issue at all, why don't you go read through the hearing tutorial. If it gets to the eardrum, it will get analyzed on the cochlea. Are you now arguing for massive IMD in the basilar membrane? Others have done that, and come up with quite overly complex explanations for very simple things.
Yes, we all know what beat frequencies are, or, well all of us who are actual professionals in this field, like me, at least.
But pretending that your opponent does not understand the most basic kind of product of a complex exponential is rather insulting. In fact, your insinuation that I don't know the basics of my own science is very nearly actionable disparagement. I suggest that you start apologizing.
As far as the cochlea is concerned, there is nothing TO neglect. The various frequencies, "beat" or not, simply fall where they should on the cochlea, and will be detected accordingly. Your complaint fails to even be wrong.
You are, however, rather insulting and quite offensive.
Trouble with the 'net, everyone's an expert except the real experts, natch...
Curvature
Major Contributor
- Joined
- May 20, 2022
- Messages
- 1,115
- Likes
- 1,402
Remember that there are very sensitive speakers and headphones out there that don't need much power to play. That also translates to picking up and pushing out noise given it's own contribution to the signal.
It is a question of specifics. In most cases, the assumptions that many engineers of consumer products make about the limited range the end listener will practically use are correct. That kind of thinking is even good engineering, and there are studies out there about optimal use of suboptimal specs. But it is also very common to hear defects produced by that equipment. Turn it up too much or connect an unexpected component, or set gain randomly. Easy. Although a lot of such problems have been dealt with by standards. And then these days it is even easier to get good, performant, reliable, stable gear because of measurements by reviewers.
As soon as you start looking at uses in the studio and in a PA context, this question of necessary specs becomes more complicated.
I don't see any problems with the first post. It represents a good effort to simplify things for people. Anyone who has looked at the science earnestly will understand that there is much more complexity to the whole thing.
I do have my own irritations at audiophile turns of phrase ("this speaker is flat"), and use of numbers ("I listen to music at 79dB"; "you can't hear distortion unless it's at 10% or more"). I'd put a lot of the discussion in this thread into that genre: irritation at simplifications.
Maybe someone is willing to do comprehensive work that will, with as much clarity as @flipflop, present better sense of thresholds, although I'm not holding my breath. It's a lot of work.
DualTriode
Addicted to Fun and Learning
- Joined
- Oct 24, 2019
- Messages
- 903
- Likes
- 594
Fun.
The basics are pretty simple. You can do a frequency analysis on ERB's or 1/3 ERB's and relate that to your overall system gain. But to do that, you have to analyze for a specific signal, obviously.
And, of course, always remember the 6dB or slightly more white noise component of the atmosphere that always exists at room temperature. So that does give you an absolute bottom level to be concerned about.
This kind of signal is also very revealing if you're listening to a "compressed" signal via a low-rate codec. "things" happen with codecs that ignore Binary Masking Level Depression. I won't name any, but trying the experiment will tell you where to look PDQ.
Start here:
Audio Blind Tests and Listener Training (video)
I see so many people claiming to have done blind tests in youtube videos recently without knowing what it means to perform one. Also, the topic of training comes up with many abusing the term. So I decided on a whim to create a video on it. It is rather long at 44 minutes but you can speed it...
www.audiosciencereview.com
There is more than that on the forum, but this is what came up in two seconds of using the Search feature.
Rick "look also for Harman listener training, which Amir also completed" Denney
Here's another one:
What are trained listeners actually listening to/for during an ABX blind test?
We've all tried to see if we could tell the difference between a lossless file and a lossy one and I was wondering (since I personally can't), what are we supposed to listen to/for? What does being a trained listener actually mean? For those who are able to tell the difference, where to do you...
www.audiosciencereview.com
The link points to a post, but read the whole thread--others also talk about specialized techniques.
We can answer these questions for ourselves. Rather than stating one hears a difference without supporting evidence, try this protocol:
1. Switch devices into the same system with levels matched. Sighted is fine for this first test. If no differences, then you're done. If you perceive a difference...
2. Plug two devices into a preamp with a remote. Switch back and forth at least a hundred times while listening to music you love. If you lose track of which device is which and can't tell the difference, you're done. If you perceive a difference...
3. Hand the remote to someone else. Make sure the two devices are level-matched using a precision voltmeter. Have the the other person do the switching from out of view. If you can't tell the difference, you're done. If you think you can, you still haven't tested out all the potential bias. So...
4. Create a test situation where the other person is also blinded to which device is being tested, preventing any possibility of a tell from the person doing the switching. Ideally, have the tester declare that one of the devices is item X in addition to being item A (or B). Then, the question is whether you can identify which device X is, A or B. If you can't with statistical significance, you're done. If you can, then you have some evidence to demonstrate a perceived difference.
You don't need controlled testing to prove you cannot perceive a difference, only to demonstrate that you can.
For hearing distortions, one can simply clip out the tails of reverberation, the recorded quiet bits in the room, and other samples that those with training know will show the effects you are searching for. Then, play those clips as comparisons. This is more difficult, and it's probably impossible without specialized training. As Amir said in the post I first linked, the training comes from listening to exaggerated samples to learn what to listen for. With that training, one might hear certain distortions in the most sensitive range to a level below 100 dB, if those artifacts are amplified enough to be out of the noise. Whether it's significant in real music listening is a whole other question.
Note that spectral tilt (even just a little of it) and clipping or compression somewhere in the chain from an inappropriate gain structure totally overwhelm any of the low-level distortions one might be talking about.
I find that if I take myself out of my usual listening mode, much of what I think I perceive vanishes. Hence, I recently advised someone to play the system to a person to whom one is sufficiently attracted to make impressing them actually important. Suddenly, it's not enough for us to perceive a difference, but for someone who basically doesn't care about our toys to notice a difference. This is easily manipulated, however, and the usual reason that the "wife in the kitchen" is part of the usual story of an "obvious" improvement. But I'm suggesting doing it for real, not making it up.
Rick "self-delusion is the most powerful kind" Denney
Similar threads
