Seem is the operative word, and is the one that leads us on the typical goose chases. As you know, those here want empirical evidence.
Yes but I’d consider that a result of degeneration and recruitment.
Just as we don’t consider autism auditory responses increased acuity.
Increased sensitivity does NOT equal increased capability, precision or quality.
If a broken arm is sensitive to the touch, does that mean you have superior perceptual abilities? No, it means you have an injury.
Correct. Like glasses. And this is exactly my point: Hi-fi amplifiers should be considered glorified ear crutches.
Last edited:
You should probably do some research then into perceptual masking?
Here's a starting point for you Auditory masking - Wikiwand
And maybe something more to your liking - you cited ASR before so allow me to do the same & quote one of many statements by Amir about masking
Review and Measurements of Schiit Yggdrasil V2 DAC | Audio Science Review (ASR) Forum
Edit: Or this more comprehensive article by Amir Audibility of Small Distortions | Audio Science Review (ASR) Forum
You casually lumped all the high level hash around 1k but below 2k into the critical bandwidth like it didn't matter. It's nonharmonic, so perception is very different to harmonics. That doesn't follow the very literature you just cited. And you're worried about people doing experiments wrong?!
Well, what you say seems contradictory - I gave the alternative view of how a particular degeneration in a perceptual faculty may actually make us more sensitive to particular qualities of our playback systems - this is increased sensitivity & increased acuity!!
"Yes but I’d consider that a result of degeneration and recruitment." - your analysis of my example doesn't equate with the psychoacoustic research which tests for normal hearing (NH) & hearing impaired (HI) listeners in it's experimental procedures.
"If a broken arm is sensitive to the touch, does that mean you have superior perceptual abilities? No, it means you have an injury." Interesting defense. A broken arm changes the perceptual signals & yes we are more sensitive to touch, gravity, etc.
You seem to have an view of some 'ideal' template of perception, rather than it being a developmental process that forms as we grow/age & provides internal signals which we interpret as a useful representation of the physical world & allows us to best interact with it?
"Yes but I’d consider that a result of degeneration and recruitment." - your analysis of my example doesn't equate with the psychoacoustic research which tests for normal hearing (NH) & hearing impaired (HI) listeners in it's experimental procedures.
"If a broken arm is sensitive to the touch, does that mean you have superior perceptual abilities? No, it means you have an injury." Interesting defense. A broken arm changes the perceptual signals & yes we are more sensitive to touch, gravity, etc.
You seem to have an view of some 'ideal' template of perception, rather than it being a developmental process that forms as we grow/age & provides internal signals which we interpret as a useful representation of the physical world & allows us to best interact with it?
Please define "high level hash"?
From Amir's article, I already linked to
I don't see anything about "high level hash" or nonharmoic distortion, do you?
But if you have citations to show that the above is wrong I would be interested?
In other words what I originally asked - why do the close-in distortions (seen in Miklos plots) not get masked by the strong fundamental we see? You have made statements but no material to back them up
Yes, I can only judge my perceptual abilities in the now & that's my conclusion. Some here seem to maintain that nobody's auditory perception is of value unless empirically tested - considering that they usually consider flawed ABX test results to be 'empirical evidence' - I couldn't disagree with them more strongly
You stated that you were wary of listening impressions without auditory tests - I showed you a logical example of how some age-related changes in auditory perception may lead to increased sensitivity to certain aspects of audio system playback - exactly the sort of reported listening impressions often reported on audio forums.
Is it any surprise, given the typical age-profile of audio forum members, that the listening characteristics I briefly outlined are the ones most often cited i.e. systems which psychoacoustically aid an increase in clarity, soundstage depth, 3D illusion, realism, etc
So does which side of the road you drive on, at least if you drove a lot before air conditioning.
May I ask, does it matter why people hear what they hear if there is no medical fix for any issues? If they hear something and find it disagreeable and they find some gear that sounds satisfactory, then so what? They found something that works for them.
Please define "high level hash"?
From Amir's article, I already linked to
Is the masking threshold curve an arbitrary one, measured one, and what does it look like vs frequency? For example, what does the masking threshold curve look like if the red sound is the primary? As, it is in the more sensitive area, will the masking curve be steeper or less steep off the peak?
Also, is the masking threshold only applicable to a single frequency at any location, or is it a consequence of the energy of a single frequency? IOW, if the area under the curve has lots of frequencies not quite making it to threshold, would their sum of energy make it audible anyway?
jn
Last edited:
Great questions - I will leave it to those who normally cite masking as the excuse for why certain sounds are not audible to answer your well thought through questions.
As you rightly query, auditory perceptual masking is not as simplistic as Amir states on ASR & that's why I was probing DPH - hoping to find out his understanding of his claim that this close-in distortion was audible & why it was not the simplistic claim often made.
Just as a note - auditory masking has been measured for different frequencies - the seminal paper probably being from 1959 "Masking Patterns of Tones" by Ehmer but modern psychoacoustics, critical bandwidths, ERBs are all relevant in answering your questions
You will see this diagram (adapted from Ehmer) in the Wikipedia link I gave - showing the masking shadow for a number of frequencies & amplitudes
Last edited:
You are referring to hearing degeneration as if it’s a fine wine, while in my view it’s more like a juice box left out in the sun too long.
We agree I believe on the basic issues at hand, just not in how we interpret the conclusions.
Last edited:
To me, while a broken arm is hypersensitive, I don’t find that to be an asset or superpower but a reaction to trauma. It’s not in balance with the feeling in the rest of the body.
It’s a response to wear.
So, no, just because deflating, bald tires make me feel the road more doesn’t mean it’s an increased sense of realism and the tire is therefor of superior quality.
Because a problem can’t be presently addressed doesn’t mean it’s not a problem or not worth evaluating. That’s how progress is made- identify problem, then identify solution.
My ideal hearing model would be younger men and women and children.... ears with that “new car smell”.
In other words, everyone’s organic test equipment here has leaky caps and X7Rs littering the signal path to varying degrees.
However if you disagree and would like to take a hearing test and post the results I’d be more than happy to eat my words.
I don’t think any complex test is necessary to make my point clear.
Edit: here are some options:
IPhone / iPad (freeware):
Mimi Hearing Test by Mimi Hearing technologies
Http://itunes.apple.com/us/app/mimi-hearing-test/id932496645?mt=8
Online:
Online Audiogram Hearing Test | Reliable & Calibrated
Last edited:
Ah, thank you.Great questions - I will leave it to those who normally cite masking as the excuse for why certain sounds are not audible to answer your well thought through questions.
As you rightly query, auditory perceptual masking is not as simplistic as Amir states on ASR & that's why I was probing DPH - hoping to find out his understanding of his claim that this close-in distortion was audible & why it was not the simplistic claim often made.
Just as a note - auditory masking has been measured for different frequencies - the seminal paper probably being from 1959 "Masking Patterns of Tones" by Ehmer but modern psychoacoustics, critical bandwidths, ERBs are all relevant in answering your questions
You will see this diagram (adapted from Ehmer) in the Wikipedia link I gave - showing the masking shadow for a number of frequencies & amplitudes
It is funny you cite a paper from 1959. For the sonic attack stuff going on presently, I've been providing content to the higher-ups with many cites from 1947 through 1955. It seems like many of the "experts" disregard any knowledge prior to their birth. Shamefully so.
And much of the best e/m theory and application stem from papers and books like Terman '47 and some from the 50's.
Jn
As to the current discussion, I think too much weight is being given to threshold levels for frequency, and too little on localization ability. I do not know if localization degrades at the same rate as threshold, and I suspect it does not, simply because localization is not directly related to how many hairs we engage in the reception, but rather the timing. It's in the brain...
In personal tests, I find that ITD dominates IID for sources directly in front. This is easily demonstrated on the beach sitting in a beach chair with a source in front of you.** Close your eyes, notice the source direction, pull a towel up in front of you blocking from the nose out on one side, altering the IID. If you concentrate, notice the source position does not change, you keyed to ITD.
I do not know if this dominance is retained throughout the localization field.
** there must be a margarita involved of course...
Edit: I would not fault Amir for brevity though, you always have to consider the audience. As opposed to this site where I have always found someone who knows more than me regardless of the subject.
Last edited:
What is also of interest is how the masking may change as a result of much higher amplitudes. 100 dB? When I'm listening to Zep, or Cheryl Lynn, or even Bruno Mars, the only time I reach 100 dB spl is just before I put the needle in the groove...
I suspect that as the inner ear muscles contract due to high amplitudes, the masking curves may also change. I would expect the transfer function to the cochlea is altered as a result of changing spring constants.
Course at my age, the muscles have gone south for the winter..😱
Jn
I suspect that as the inner ear muscles contract due to high amplitudes, the masking curves may also change. I would expect the transfer function to the cochlea is altered as a result of changing spring constants.
Course at my age, the muscles have gone south for the winter..😱
Jn
Last edited:
I found that to be true as well. Some of the best accelerator texts were from that approximate period. It seems no longer in fashion to explain things the way they did back then.
Yes, far too much emphasis on frequency - maybe because of the over-reliance on FFTs? Temporal acuity seems not to degrade at the same rate with age, as you say.
At least or a mojito?
I've seen you present complex info without over-simplification - simplification is often the enemy of scientific inquiry & progress
mmerrill--
High level hash and the wide jitter-derived sidebands with a huge area under the curve are one in the same (~200 Hz bandwidth around the 1k is when it drops to -70 dB). Remember also, that these are time averaged so you're getting a convolution of the jitter's distribution and the pure tone. Similarly, it should be mentioned that's over 2^16 points, so what exactly of this is within the temporal window is also undefined. Admittedly that's just about a second and a half. Reading Gelfand's sections summarizing critical bandwidth tell you that Miklos's sensitivity is pretty dang close to what Figure 10.7-10.9, starting on pg. 171, would predict.
A Jneutron pointed out, I don't know what *exactly* (and that probably fairly variable person to person) the masking effect of a spot sinusoid at 1 kHz is. We can winnow in that critical bandwidth is around 50-200 Hz. We're less sensitive to broadband noise than we are to pure tones, as well.
To reiterate, one looks at the tones Miklos provides, one sees that there is an enormous amount of energy in the sidebands of the 1kHz. Equivalent to *enormous* amounts of jitter. As I clarified, Richard made a swipe at "armchair experts", and my point was, that this is the closest chance you're going to have *to* audibility (that amount of energy in the sidebands due to jitter) and *still* Miklos struggled to hear it. Now is that generalizable as "no one can hear it"?? No, not at all! But this little snippet makes the opposite point to what he desired. I clarified my off-the-cuff comment when you asked me for "why I thought it would be audible", to "this is the best chance of audibility", as the jitter is pathologically bad at this point that it suggests something is broken with the equipment. Fortunately, it's simulated tones and not Miklos's hardware.
You made the comment that read like the critical bandwidth of a 1kHz tone was essentially 1 kHz itself (out to 2HD), which is something I'd never read before in my life. Critical bandwidth is much smaller, but, yes, wide enough that people struggle with 192 khz .mp3s, which are notably smaller than the equivalent .wav file.
High level hash and the wide jitter-derived sidebands with a huge area under the curve are one in the same (~200 Hz bandwidth around the 1k is when it drops to -70 dB). Remember also, that these are time averaged so you're getting a convolution of the jitter's distribution and the pure tone. Similarly, it should be mentioned that's over 2^16 points, so what exactly of this is within the temporal window is also undefined. Admittedly that's just about a second and a half. Reading Gelfand's sections summarizing critical bandwidth tell you that Miklos's sensitivity is pretty dang close to what Figure 10.7-10.9, starting on pg. 171, would predict.
A Jneutron pointed out, I don't know what *exactly* (and that probably fairly variable person to person) the masking effect of a spot sinusoid at 1 kHz is. We can winnow in that critical bandwidth is around 50-200 Hz. We're less sensitive to broadband noise than we are to pure tones, as well.
To reiterate, one looks at the tones Miklos provides, one sees that there is an enormous amount of energy in the sidebands of the 1kHz. Equivalent to *enormous* amounts of jitter. As I clarified, Richard made a swipe at "armchair experts", and my point was, that this is the closest chance you're going to have *to* audibility (that amount of energy in the sidebands due to jitter) and *still* Miklos struggled to hear it. Now is that generalizable as "no one can hear it"?? No, not at all! But this little snippet makes the opposite point to what he desired. I clarified my off-the-cuff comment when you asked me for "why I thought it would be audible", to "this is the best chance of audibility", as the jitter is pathologically bad at this point that it suggests something is broken with the equipment. Fortunately, it's simulated tones and not Miklos's hardware.
You made the comment that read like the critical bandwidth of a 1kHz tone was essentially 1 kHz itself (out to 2HD), which is something I'd never read before in my life. Critical bandwidth is much smaller, but, yes, wide enough that people struggle with 192 khz .mp3s, which are notably smaller than the equivalent .wav file.
I found Jack Tanabe to be a great mix between theoretical and practical.
For the accelerator school stuff, Many of the teachers are my coworkers down the hall...they are crazy, totally devoid of any common sense.
I enjoy the company😀
Jn
At the moment, what I find to be the most interesting aspect, somewhat related to this discussion but probably warrants a separate thread - is how we differentiate foreground from background in what we hear. In the physical world, we have no problem doing this automatically - we walk into a room/hall/church/field, etc & quickly differentiate the acoustic background from the foreground signal streams. Again, this is done in the brain - through the auditory processing facilities we all have. The auditory clues by which we perform this feat are often very subtle & even are thought to involve summary statistics
In playback systems, it's much more onerous, as the subtle cues need to be reproduced as do the statistical patterns of the background sound. We seem to be very sensitive to these statistical patterns & they quickly lose their illusion of realism in reproduced sound - think of how replay systems often don't produced a completely realistic rendition of textural sounds like applause, rain on a tin roof, cymbal brushing - sounding more like white noise - whereas really good systems present inner detail in these sounds which provides the auditory clues to their realism.
I believe what I said before about lower noise in replay systems has a bearing on this illusion of increased realism & maybe us older gits notice this more?
In playback systems, it's much more onerous, as the subtle cues need to be reproduced as do the statistical patterns of the background sound. We seem to be very sensitive to these statistical patterns & they quickly lose their illusion of realism in reproduced sound - think of how replay systems often don't produced a completely realistic rendition of textural sounds like applause, rain on a tin roof, cymbal brushing - sounding more like white noise - whereas really good systems present inner detail in these sounds which provides the auditory clues to their realism.
I believe what I said before about lower noise in replay systems has a bearing on this illusion of increased realism & maybe us older gits notice this more?
I suspect it may be because us older gits do not have a life..😉
I think much has to do with directional cues. It is not possible to use two channel playback to create the exact sound wavefronts a baby grand in the room would present. I also suspect Bose realized that as well.
Darn, just realized...haven't hooked up that lifestyle 1 2 3 system yet...just not a priority..
Jn
There really are no tools available for ITD analysis. With only a hammer, everything is a nail.
Ok, let's compromise... Mojito in left hand, margarita in right...😉
My wife always used to emphasize the importance of being able to explain complex stuff to those who did not train in a specific discipline. And, more importantly, to do so without talking down..
Important lessons I try to follow. Although I will admit, when my children and grandchildren talk down to me as they explain my phone or IPad to me.....money cannot buy that level of pleasure..
Always remember the important things...
Jn
Btw, I hate statistics. And his brother "summary".
Last edited:
Thanks.
Yes, the critical bandwidth of the Bark scale centered at 1KHz is about 160Hz i.e 80Hz either side of 1KHz. Hard to tell, from eyeballing Miklos plot where the 1080Hz x-axis cursor would meet the plot? But the general concept of masking is that energy within the same critical band or Bark is all added together i.e. is not differentiated from the fundamental. Energy in the next critical band (~1080Hz to 1270Hz) is all added together & may be audible depending on whether it is above threshold for that frequency
Unless I'm looking at the wrong Figs in "Hearing: An Introduction to Psychological and Physiological Acoustics, Fifth Edition", I don't see these figures supporting what you say - what book?
Are we not more sensitive to the masking effects of broadband noise then to pure tones?
Right, thanks - I wondered where you were coming from?
I didn't think I made a comment that read like that?
- Status
- Not open for further replies.