[bigshot]

I actually try to stay away from this thread because I end up doing the exact same thing you guys are: obfuscating the hell out of matters! I'm out.

Don't leave! We need more baby animal avatars around here! I just sit back and skim over all the blather and wait for someone to start talking about playing back commercially recorded music in the home again. It can be a long wait at times. Some people never seem to run out of gas.

 

[castleofargh]

I disagree....

If we know so much about precisely how humans hear, and exactly what they can and cannot hear or differentiate, then how come we have so many surround sound systems, and so many competing systems that attempt to record and play back binaural recordings, yet there is not a single one of them that is universally agreed to "work perfectly"? If we know all these details so well, and the technology is so well established, then I should have dozens of systems to choose from, whose main difference is the color of their knobs and the shape of their face plates. Yet, instead, we have a whole bunch of competing systems, none of which seems to be especially close to that goal.

To me, the fact that no such perfect system exists is adequate proof that "we don't know quite all of it yet".

When you can sit me down in a room with a band and a sound reproduction system.
And I literally cannot tell which is playing... ever...
And, when we try it with other bands, and other listeners, nobody can ever tell the difference....
Then I'll be the first to agree that "we've accomplished perfect sound reproduction so there's no need to keep trying".
Until then... we simply aren't there yet.

bad examples. there are quite a few obvious technical issues and deliberate shortcuts in recording, mastering, calibration, and playback of such systems. so of course the result is unlikely to "work perfectly". also what's the reference? as a consumer most of the time I won't know. and even more often it won't care about me. so if all was perfect, could we consumers even tell?
anyway the issues mostly fall under technical and acoustic concerns, not mysterious human interpretation of sound. if you get the right sound at the eardrum, sound involving at least a specific room and speakers, or your specific HRTF cues for binaural, then you're done with sound. even in some utopia of ideal conditions, it wouldn't be surprising for some people not to feel "right". that much is true, but the cause would be what they see or their knowledge of the room. and for binaural, it would have to do with stuff like the lack of head tracking and tactile bass for example.

so bad example to show that we lack understanding about hearing and interpretation. I'm not saying that we know everything about hearing and interpretation, only that this example is bad.

 

[bigshot]

Any port in a storm.

 

[Kammerat Rebekka]

Let me take a stab in the dark here: the ONLY test you will ever see audiophiles greet welcome is the one that proves them right.
The real problem though seems to be that the upgradeitits bug is scouring about in venues that make absolutely no sense whatsoever.
How about upgrading your ears before venturing any further?

 

[KeithEmo]

I agree - that wasn't a great example.
A lot of the gaps involved entail not lack of understanding but simply lack of ability to deliver things we know are missing.
I would also suggest that we really don't necessarily want a perfect reproduction....
For example, nobody wants to actually suffer hearing damage when the rocket takes off at 150 dB in that scene from Interstellar....
We're quite content to have that experience inaccurately simulated by "a really loud noise"....

However, excluding extremes like that, if we start by assuming that some original event actually existed, which is not always the case...

We could replicate what both you and I heard at that original performance (or would have heard there) by exactly replicating every sound wave in the room....
Then, when both you and I listen to it, the sound reaching our eardrums would be exactly the same as it was at the original performance.
This would involve all sorts of things that are annoying in practice - like exactly replicating the acoustics of the original room and each instrument or other sound source.
However, since the sound would be passing through our ears, we wouldn't have to worry about things like HRTF (that part would "take care of itself").

Alternately, we can try to bypass some or all of the variables.
For example, if we had some way to deliver the signal directly to our eardrums, we could bypass both the room and the mechanism of our ears.
This is what we try to do with binaural and IEMs... deliver the signal as close as possible to our eardrum... and electrically simulate or compensate for the rest of the path.
By using headphones and individual HRTFs, we've "taken a step back"...
We've delivered the signal closer to our eardrums, and artificially compensated as many of the differences between the speaker and our eardrum as we can.
(However, IEMs till don't deliver the signal directly to the eardrum and, as you've noted, individual HRTFs are a nuisance to generate and use.)

I should also mention that various attempts have been made to "cover" some of the extra information you mentioned.

For example, both big subwoofers, and "buttshakers", do their best to replicate the tactile feedback... and can do pretty well.
And the Smyth Realizer seems to do a passable job of replicating the head tracking requirement...
The Realizer uses headphones, mounted with an IR sensor, to track head position, and simulates the differences you hear due to changes in head position.
It essentially "simulates listening to the surround sound system of your choice, in the room of your choice, while wearing a pair of stereo headphones".
I guess you could also claim that VR headsets attempt to replicate the full visual portion of the "total experience".

Arguably, in some future system, we could somehow play the music signal directly into the auditory nerve...
If we could do that, in theory we could actually "hear the music through the EARS of the performer" (or, at least, a "designated standin for the reference listener").)
(In fact, if we would somehow "patch into all five senses" then we could replicate the entire experience.
People have been theorizing that for decades... but the technology isn't there yet.

bad examples. there are quite a few obvious technical issues and deliberate shortcuts in recording, mastering, calibration, and playback of such systems. so of course the result is unlikely to "work perfectly". also what's the reference? as a consumer most of the time I won't know. and even more often it won't care about me. so if all was perfect, could we consumers even tell?
anyway the issues mostly fall under technical and acoustic concerns, not mysterious human interpretation of sound. if you get the right sound at the eardrum, sound involving at least a specific room and speakers, or your specific HRTF cues for binaural, then you're done with sound. even in some utopia of ideal conditions, it wouldn't be surprising for some people not to feel "right". that much is true, but the cause would be what they see or their knowledge of the room. and for binaural, it would have to do with stuff like the lack of head tracking and tactile bass for example.

so bad example to show that we lack understanding about hearing and interpretation. I'm not saying that we know everything about hearing and interpretation, only that this example is bad.

 

[sonitus mirus]

I've heard a few recordings of "the sound of the ocean"... and, to me, they just sound like a lot of noise.
But, by applying some signal processing, the guys on a Navy ship can apparently tell me what ships are in the area, what directions they're going in, and how fast they're going.

I'm told they can even identify individual ships by their "sound signatures"

I was in the US Navy as a sonar technician/supervisor for eight years, albeit over 20 years ago. It was mostly trigonometry and an extensive library of recordings made from undersea hydrophones, submarines, and to a lesser extent frigates/destroyers/cruisers equipped with towed arrays that allowed for knowledge of situational information and identification. Generally, the signal processors created a graphical representation of the sound displayed as frequencies with their intensity.

Typically a sound source other than a biologic would first be identified, and further analysis would be needed to determine the range, direction of travel, speed, and type of vessel. It was not instantaneous. Bathythermographs and navigation charts would be used to gather data about the ocean environment with regards to thermal layers and any physical characteristics. Course changes could be used to narrow the target's location. The signal strength could be used to determine if the target distance was increasing or decreasing.

Any identifying auxiliary equipment could be referenced, but in my day we had to study tons of information to be able to do our jobs successfully. The system didn't do this for us. Propellers, turbines, generators, diesel engines, gears, and other noise-makers could help to isolate a class of vessel or even a specific ship/sub. This process could all be automated today with enough CPU resources, but it is mostly going to be a database of information that meets specific requirements and not from signal processing.

 

[KeithEmo]

I would guess you do so by politely asserting that a fallacy exists - and providing actual information to support your assertion.

For example, Gregorio asserts:
Yes, there is *sometimes* something there (>20kHz) that "obviously" modulates in time with the energy peaks of the music.
How do you jump from this fact to the conclusion that's "it's clearly NOT random noise"?

Well, here's a scientific response to that.....

The simplest answer is self-referential (it's a mater of definition)...
If the signal is modulated, then that modulation constitutes information...
Therefore, since the signal contains information, it is not "random".
(By definition, something that contains information is not "purely random".)

Perhaps he misspoke.... and intended to say that: "He is personally certain that it contains no useful information."
If so, then he is entitled to his opinion on that subject, but should avoid asserting it as established fact.

In some contexts noise is defined as the opposite of information....
So, also in general, pure noise would contain no information.
(In reality, with proper analysis, the noise itself will give you information about the processes that generated it.)
But, if that noise is modulated, then that modulation itself contains more information.
If the modulation was deliberate, it has added both the information we chose to modulate onto the signal, and information about our modulation method.
If the modulation was unintentional, then, at a minimum, it contains information about the mechanism that caused it.

So, if we have what appears to be "noise that is modulated in time with the musical peaks"...
Then it contains significant information about how and why those musical peaks were modulated.
(We also have an excellent starting point for our analysis since we have the audible portion of the overall sound spectrum to use for reference.)

To use the example of the Plangent process...
Record bias serves a purpose in terms of tape magnetization - but, as far as the audio signal is concerned, it is noise.
However, some smart fellow noticed that, not only was that noise present, but it was "modulated" by variations in tape speed.
Then, by analyzing that "noise", and precisely how it was modulated, he figured out how to extract useful information about errors in tape speed.
This information is then used by the rest of the process to make audible corrections to the audible portion of the audio spectrum.
(It would be fair to state: "The Plangent process makes audible corrections - based on the analysis of the unintended modulation of noise present on the tape.")

To choose another potential example.
I may notice that the background noise level in the room rises and falls slightly at a certain frequency.
(I would detect this by looking for patterns modulated into the background noise.)
From that, I may determine that a ceiling fan, or something similar, is modulating the noise in the room (a fan is a series of moving reflective surfaces).
And, if I were to detect this, I would look for similar modulations in the levels of audio tracks recorded in that room.... and potentially correct them if I find them.
As a result, I may be able to improve the quality of those audio tracks, by removing an error introduced when they were initially recorded.
This is simply an advanced version of "establishing a noise profile and using it as a basis for corrections".

And, again, the fact that this takes a good programmer a week to do today...
In no way suggests that we won't see it in a $125 AVR in five years.

How do you refute the use of a fallacy without referring to the fact that a fallacy is being employed? How do you refute a ridiculous assertion without explaining/demonstrating why it's a ridiculous assertion?

G

 

[Phronesis]

1. That's entirely possible but as the sort of people I've been hanging around for several decades is world class musicians and engineers, what are we to do? Ignore the actual facts of what musicians and engineers do and just rely on make-up nonsense instead?

2. But he's not making any technical arguments, he's using ridiculous fallacies, non-sequiteur analogies and false assertions. I've asked you whether we should just let all that pass here in the science forum and if not, how to refute it. You have not responded, which implies you think we should let it pass. No problem but then we have to rename this forum!

G

I think you should debate the technical points as forcefully as you want. Just don't make it personal. When you finish writing your posts, before you hit 'Post Reply', just filter out all the personal stuff and everything will be fine. Making it personal doesn't help your arguments, it just diminishes your credibility with respect to being able to rely on the merits of your arguments, inhibits constructive discussion, and makes the forum a less pleasant place to spend time. The same dynamics tend to occur when people talk about politics and religion, and we should be able to do better here in Sound Science where the purported mission is to get closer to the truth by applying a scientific mindset.

 

[castleofargh]

quasi-modo:
personal attack removed and bad bad Gregorio contacted. I wish I could edit a post because most of it was IMO interesting. dunno if Greg will repost an edited version.
@Phronesis maybe don't keep quoting and discussing the very post you think shouldn't be on the forum? that's a little counterproductive ^_^. plus when I come to remove it, I have to chase all the related posts and quotes of it. report, or try your best to comment without actually quoting insults, do it for your modo brother who's really lazy.

 

[KeithEmo]

Yes, today, virtually all of the "number crunching" is done by computer... and there's a major stress on getting it done in real time if at all possible.
Modern analysis techniques are also able to sort out a lot more information a lot faster.
Most systems these days would automate both the lookup process and consolidating that information with other forms of data.
These days, you'll probably end up looking at a screen, with a little icon for each ship or plane, color coded by nationality, class, threat level, or whatever other parameters you set.
The computer will have combined the SONAR and RADAR data, with satellite data, data from each ship's transponder, and reference data about things like "last known location".
(Just click your mouse if you want to see all military vessels, or all freighters of Russian registry, or all ships making over ten knots in the general direction of Hawaii.)

It also might detect and note or flag anomalous information....
For example, it might alert you to an anomaly if it detected an object with "the hull sounds of a destroyer" - but "the engine sounds of a freighter".
Or if the SONAR and RADAR indicate an armored warship, while official records say the only thing near there should be a fishing boat, and the transponder is oddly silent.
This could be really useful for identifying ships trying to camouflage their sound signatures - and so hide or falsify their locations.
(I worked on a project back in the 1970's that involved developing quieter propellers to reduce the acoustic propulsion signatures of large warships.)

Now, when it comes to consumer music, we're usually dealing with "an artificial reality" when multi-track recordings are involved.
However, we may also encounter recordings that claim to have been "recorded live", at a single sitting, and without a lot of processing.
In those situations, it might be useful to detect whether the acoustic signatures of the backgrounds on all the tracks matched.
Was that "audiophile recording" really recorded direct-to-file, with minimal processing, or did they drop in a few tracks later, and add artificial reverb to match them to the others?
In fact, I'll bet we could even recognize certain reverb plugins by their acoustic signatures.
Now, while this is hardly life or death, many people might find it interesting... or even potentially useful.

I was in the US Navy as a sonar technician/supervisor for eight years, albeit over 20 years ago. It was mostly trigonometry and an extensive library of recordings made from undersea hydrophones, submarines, and to a lesser extent frigates/destroyers/cruisers equipped with towed arrays that allowed for knowledge of situational information and identification. Generally, the signal processors created a graphical representation of the sound displayed as frequencies with their intensity.

Typically a sound source other than a biologic would first be identified, and further analysis would be needed to determine the range, direction of travel, speed, and type of vessel. It was not instantaneous. Bathythermographs and navigation charts would be used to gather data about the ocean environment with regards to thermal layers and any physical characteristics. Course changes could be used to narrow the target's location. The signal strength could be used to determine if the target distance was increasing or decreasing.

Any identifying auxiliary equipment could be referenced, but in my day we had to study tons of information to be able to do our jobs successfully. The system didn't do this for us. Propellers, turbines, generators, diesel engines, gears, and other noise-makers could help to isolate a class of vessel or even a specific ship/sub. This process could all be automated today with enough CPU resources, but it is mostly going to be a database of information that meets specific requirements and not from signal processing.

 

[KeithEmo]

Yes, today, virtually all of the "number crunching" is done by computer... and there's a major stress on getting it done in real time if at all possible.
Modern analysis techniques are also able to sort out a lot more information a lot faster.
Most systems these days would automate both the lookup process and consolidating that information with other forms of data.
These days, you'll probably end up looking at a screen, with a little icon for each ship or plane, color coded by nationality, class, threat level, or whatever other parameters you set.
The computer will have combined the SONAR and RADAR data, with satellite data, data from each ship's transponder, and reference data about things like "last known location".
(Just click your mouse if you want to see all military vessels, or all freighters of Russian registry, or all ships making over ten knots in the general direction of Hawaii.)

It also might detect and note or flag anomalous information....
For example, it might alert you to an anomaly if it detected an object with "the hull sounds of a destroyer" - but "the engine sounds of a freighter".
Or if the SONAR and RADAR indicate an armored warship, while official records say the only thing near there should be a fishing boat, and the transponder is oddly silent.
This could be really useful for identifying ships trying to camouflage their sound signatures - and so hide or falsify their locations.
(I worked on a project back in the 1970's that involved developing quieter propellers to reduce the acoustic propulsion signatures of large warships.)

Now, when it comes to consumer music, we're usually dealing with "an artificial reality" when multi-track recordings are involved.
However, we may also encounter recordings that claim to have been "recorded live", at a single sitting, and without a lot of processing.
In those situations, it might be useful to detect whether the acoustic signatures of the backgrounds on all the tracks matched.
Was that "audiophile recording" really recorded direct-to-file, with minimal processing, or did they drop in a few tracks later, and add artificial reverb to match them to the others?
In fact, I'll bet we could even recognize certain reverb plugins by their acoustic signatures.
Now, while this is hardly life or death, many people might find it interesting... or even potentially useful.

 

[KeithEmo]

I'm going to stop responding entirely to "arguments" on this subject...
I will even concede that "at this point in time there is probably little if any information contained in the ultrasonic portion of the spectrum of high sample rate files that is directly useful by any existing home audio gear".

It seems clear that many people here simply fail to find this interesting...
While several others have, quite reasonably, pointed out that it is somewhat speculative.
And some just want to argue.
(And this forum is more intended to address current products and technology than future and potential products.)

If anybody wants to actually discuss it... feel free to PM me, or post an actual question here, and I'll be glad to reply.

quasi-modo:
personal attack removed and bad bad Gregorio contacted. I wish I could edit a post because most of it was IMO interesting. dunno if Greg will repost an edited version.
@Phronesis maybe don't keep quoting and discussing the very post you think shouldn't be on the forum? that's a little counterproductive ^_^. plus when I come to remove it, I have to chase all the related posts and quotes of it. report, or try your best to comment without actually quoting insults, do it for your modo brother who's really lazy.

 

[castleofargh]

I'm going to stop responding entirely to "arguments" on this subject...
I will even concede that "at this point in time there is probably little if any information contained in the ultrasonic portion of the spectrum of high sample rate files that is directly useful by any existing home audio gear".

It seems clear that many people here simply fail to find this interesting...
While several others have, quite reasonably, pointed out that it is somewhat speculative.
And some just want to argue.
(And this forum is more intended to address current products and technology than future and potential products.)

If anybody wants to actually discuss it... feel free to PM me, or post an actual question here, and I'll be glad to reply.

different people with different views and interests. nothing we can do about that on a public forum. we all just try to find what interests us and the people who also are interested in it. and that's rarely going to be everybody. we should be able to direct and answer things in a more common way thanks to controlled experiments and the knowledge base that research provided us with, but that would first require that everybody subscribes to the same knowledge. in practice we can't even get everybody to agree that a listening test shouldn't involve our eye and preconceptions. so, it's a work in progress to say the least.
and then we obviously get the same issues that all scientists fight continuously about, interpretation of data. TBH when I try to follow some modern research(audio or not), the real scientists with real serious experiments, well they really argue and fight just like we do here all the time. it's at an elevated level of control and understanding compared to here, but the issues remain. was the test really properly done? is the data trustworthy? can you really conclude this based on just that data? when there is a consensus like man made climate change, it's only because the evidence is overwhelming. but first we need the evidence. and when it comes to hearing stuff beyond what we nowadays assume to be human hearing thresholds, the evidence isn't overwhelming at all. and the people pushing the idea that there is more, do tend to have a mediocre understanding of human hearing/digital audio, or they are in the business of selling high res gear, and we all have a hard time forgetting that aspect.
I have to be honest, if the paper from Meridian about audibility of ringing had come from actually independent research, I would still have had a few issues with their experiment, but I wouldn't has been so keen on dismissing it. if the paper on checking the brain for impact when music has ultrasonic content didn't have Oohashi, I would also have ended up finding a few issues, but I wouldn't have started with "oh boy, here he goes again". it's unfair and really not scientific, but who can claim that he won't be affected that way? biases will be there. it's really the all story. so as a consequence, a guy like yourself has to prove himself even more than others. again it's unfair, but there is really nothing we can do about it. I can't unlearn what you do for a living, and I can't have the names on the post be randomized so I can read them without my bias.
I mean I ended up in a pure rage quit 2 days ago in a thread because having the "sound science forum modo" tag has some people pass anything I post through the "he's a Sound Science guy" interpreter. which isn't even correct as what it says is that I moderate the section. but the bias is there, rational or not. I know I can't remove that bias, and this time my solution was indeed to rage quit instead of trying to be so perfect that maybe my point would still get through the bias googles. so I believe I kind of get where you're standing right now. but you're a better man than I am, you don't even post anything suggesting that you'd really like to headbutt some people. all hope isn't lost for you .
with that said, I'll keep being a dick and disagree with bad analogies making everything more complicated. and I'll still have the skeptical position on anything without clear evidence, making me the other side of any optimistic view on what might happen. bias or not, a bunch of people here will always be that way. because indeed we're not here to deal with theoretical science and speculations. we already have our hands full trying to settle on known facts.
doesn't mean you can't discuss speculations or poorly supported ideas, it's a public forum you discuss whatever you like(within TOS). you just might not find many people to ride the speculative train in this specific section.

 

[gregorio]

... then we obviously get the same issues that all scientists fight continuously about, interpretation of data. TBH when I try to follow some modern research(audio or not), the real scientists with real serious experiments, well they really argue and fight just like we do here all the time.

To be honest, it's not really just like what we do here all the time. Their fighting is, as you say, about the interpretation of data but they don't fight about falsified/just made-up data, such a person is simply labelled a fraud, has their medical licence revoked or looses their job and becomes an outcast from the world of science. That's effectively exactly the opposite to this forum (and head-fi in general). According to the rules here, there's nothing wrong with someone making up lies, relying exclusively on fallacies or indeed repeating those lies and fallacies ad infinitum. It is however against the rules to mention that someone is lying or label them a fraud. I understand why those rules exist, the financial realities of head-fi but nevertheless, those rules are inconsistent with the name of this sub-forum. Which is why so many of those with real information/facts/science sooner or later feel obliged to leave or are kicked-out. Eventually, the frauds will get what they want, a forum called Sound Science which has nothing to do with science and everything to do with marketing, plus a willing audience who think they're actually getting the facts/science.

G

 

[bigshot]

We just need to come up with euphemism words... W C Fields said that he "prevaricated". Would that be OK to use instead of "lying"?