Conclusive "Proof" that higher resolution audio sounds different

[arnyk]

Remember we are talking about people that don't seem interested in listening to and enjoying music but want to turn that pleasure into some sort of science experiment.

That would appear to be yet another rendition of the subjectivist deceit that only they listen to music at all, or only they listen to music for enjoyment.

I will advance the perhaps controversial idea that music is more enjoyable when listened to via a good sounding audio system.

The science experiments are just a means to that end.

I'm wondering if some people seriously believe, as it seems from their self-serving statements, that good sounding audio systems are just something that they pull out of their butts? ;-)

 

[thedudeabides]

I will advance the perhaps controversial idea that music is more enjoyable when listened to via a good sounding audio system.

The science experiments are just a means to that end.

I'm wondering if some people seriously believe, as it seems from their self-serving statements, that good sounding audio systems are just something that they pull out of their butts? ;-)

Rest assured sir that after 35 years plus of working on and listening to my system, it does sound quite good to me. Other experienced listeners who have heard my system agree with my assessment. I didn't pull it out of my butt.

It is obvious that the end, from your perspective, is science determined and is the "means" and is not based on listening to and enjoying of music.

Why is that you and others cannot accept the basic fact that sincere people have an honest difference of opinion on this issue and anyone who doesn't buy into and / or disagrees with yours is full of crap.

How immature, disrespectful, and myopic.

And why, pray tell, do you have to be insulting and combative when you have no idea what my system sounds like.

I'm done with your attacks and insinuations.

Have a great day.

 

[arnyk]

It is obvious that the end, from your perspective, is science determined and is the "means" and is not based on listening to and enjoying of music.

Really? What is there about these two sentences that is so confusing and mysterious as to create such a misapprehension about my true meaning? I didn't mean to write a riddle!

"I will advance the perhaps controversial idea that music is more enjoyable when listened to via a good sounding audio system. The science experiments are just a means to that end."

To lay your confusion to rest, let me express the same idea in a far shorter way: I listen to my audio systems for the purpose of enjoying music and drama.

Why is that you and others cannot accept the basic fact that sincere people have an honest difference of opinion on this issue and anyone who doesn't buy into and / or disagrees with yours is full of crap.

These days sincerity and $5 will get you a cup of coffee in a cheap place. There are such things as right and wrong, and there is such a thing as being sincerely wrong.

And why, pray tell, do you have to be insulting and combative when you have no idea what my system sounds like.

I didn't know that this was a discussion about what your system sounds like.

Please let me remind you that what precipitated this exchange was the following insult: "Remember we are talking about people that don't seem interested in listening to and enjoying music but want to turn that pleasure into some sort of science experiment."

Who wrote that insult?

 

[thedudeabides]

These days sincerity and $5 will get you a cup of coffee in a cheap place. There are such things as right and wrong, and there is such a thing as being sincerely wrong.

Anything you say Jesus. Anything you say.

Praise to the Lord.

 

[amirm]

This really is a deceptive view here. I do believe you know quite well why that is.

The test tones are around a -.5db level. As the jangling keys have a peak of -1 db, several at -2 db and numerous at -3 db such a check for IMD isn't unreasonable.

What??? You can't look at the peak levels of the key jingling part that way. The peaks are not just made up of ultrasonics but both audible and ultrasonics. As such you can't just compare their levels to purely ultrasonic tones Arny has inserter.

We can take one step toward a more accurate comparison by getting rid of lower frequencies. I did this hear by filtering everything below around 23 Khz:

I put a red line where the highest peak is. If you go to the right you see that its level is below -6 db, not -1 db.

But even that is not enough to make the comparison proper. That peak is only hit one and only one time on that clip (denoted by the red circle). Unless someone has proof that the only way to tell the difference is to listen to that one single transient, then we can't go by that peak either. Remember, Arny's test tones are not a single transient. They are continuous tones.

I eyeballed a level that is likely to hit more of the peaks in blue. That level is now below -15 db (sorry, I chopped off part of the "5" in 15 on the vertical scale). Even that doesn't assure us that we are capturing the areas that showed a difference. But even if it did, compare the level of the blue line to the magnitude of the test tones on the right.

The average RMS amplitude of the key jingling part (post filtering) by the way is -41 dbfs. The average RMS amplitude of the tones is just -6 dbfs. That is a difference of 35 dbfs.

Your using an FFT to compare a dynamic signal level to a steady test tone is deceptive. Though the average level of those keys is fairly low the peak levels are plentiful high enough in level to cause IMD if the system is pushed too near its limit and susceptible to that. Really bush league move Amir.

I don't know what a bush league move is . I do know that you can't analyze the high frequency content of a track by just looking at its total amplitude per above. The discussion is about specific part of the spectrum. You cannot take full spectrum numbers and compare them to ultrasonic tones in the context of ultrasonic induced IM distortion. Your analysis must be frequency aware by definition which it was not.

Arny's files might sound not so loud yet cause some IMD which would be uncovered by his test tones. A very minor volume adjustment downward would eliminate them and then make the test of the keys valid so one wouldn't confuse IMD from ultrasonics with hearing a difference from hearing ultrasonics.

I have run his test and my laptop did not generate any IM distortions even at those elevated levels. Arny has a theory but no data whatsoever that says the reason we can tell the files apart is due to high amplitude ultrasonic tones.

Did you hear a difference right at that peak?

So no, there is no way you can represent those ultrasonic tones as being representative of music or the key jingling part. Not even close. If a system does generate IM distortion with those tones as a result, you have no idea if that is the reason you are hearing a difference. Arny has a theory here. He has no proof that any of this IM talk has any bearing on the results. No cause and effect has been shown. We can throw stuff at the board and see what sticks. Doing that takes better attempts and explanation than what is being put forth.

BTW, I appreciate that you abide by the rules of WBF to keep your posting professional with less personal accusations. We are discussing a hobby and not life and death.

 

[amirm]

Anything you say Jesus. Anything you say.

Praise to the Lord.

 

[amirm]

I am going to extend that caution to other personal remarks. Please take a deep breath before hitting that post button that has such tone.

 

[aronjt]

This really is a deceptive view here. I do believe you know quite well why that is.

The test tones are around a -.5db level. As the jangling keys have a peak of -1 db, several at -2 db and numerous at -3 db such a check for IMD isn't unreasonable. Your using an FFT to compare a dynamic signal level to a steady test tone is deceptive. Though the average level of those keys is fairly low the peak levels are plentiful high enough in level to cause IMD if the system is pushed too near its limit and susceptible to that. Really bush league move Amir.

Arny's files might sound not so loud yet cause some IMD which would be uncovered by his test tones. A very minor volume adjustment downward would eliminate them and then make the test of the keys valid so one wouldn't confuse IMD from ultrasonics with hearing a difference from hearing ultrasonics.

Hi,

I'm just curious, do you have any formal education or special training in any electronics or audio related fields, or are your conclusions derived from an armchair perspective? I hope you don't mind my asking.

Thanks.

 

[Phelonious Ponk]

Thank you for again stating the obvious.

Mike, why bother?

Remember we are talking about people that don't seem interested in listening to and enjoying music but want to turn that pleasure into some sort of science experiment.

Oh well.

The best improvement in my listening experience in months arrived today, for only $8 -- 3 flange tips for my new IEMs. They'll give me a decent ear canal seal, so I can block out the horrid pop music at the gym.

Tim

 

[amirm]

Rest assured sir that after 35 years plus of working on and listening to my system, it does sound quite good to me. Other experienced listeners who have heard my system agree with my assessment. I didn't pull it out of my butt.

Maybe it is not said often enough but I don't think anyone thinks you all have built anything but exceptional systems. I have been to Mike and Steve's houses and of course countess demos of high-end audio. And I have heard incredible reproduction of music in them. Beyond fidelity most of your systems are works of art in industrial design and decor.

The "challenge" is that folks on the other side say they can produce the same performance at lower cost. It would be an interesting exercise to downgrade one of your systems and see if the audible performance has been reduced. Alas we won't be able to agree on how to evaluate that.

Why is that you and others cannot accept the basic fact that sincere people have an honest difference of opinion on this issue and anyone who doesn't buy into and / or disagrees with yours is full of crap.

Isn't this a two way street? Why be up in arms over others wanting to use technical analysis as the basis of their audio purchases? Personally I am at ease having a conversation with either camp. Some of us are not however as the emotional outbreaks show.

How about this hypothetical. What if I take your turntable and run it through an AD/DA and DSP and manage to improve the fidelity by a noticeable margin. Do you think this is impossible? If it is not, then doesn't it undo one of the foundations of why turntables sound better than digital?

 

[thedudeabides]

Isn't this a two way street? Why be up in arms over others wanting to use technical analysis as the basis of their audio purchases?

Amir,

I am, of course, not saying that at all. If folks want to rely on tech to determine musicality and long term enjoyment with their system, that's fine with me.

My point is that the folks in the "O" crowd seem to not accept / understand the "S" perspective as equally valid and discredits / demonizes anyone who believes otherwise.

 

[Phelonious Ponk]

Amir,

I am, of course, not saying that at all. If folks want to rely on tech to determine musicality and long term enjoyment with their system, that's fine with me.

My point is that the folks in the "O" crowd seem to not accept / understand the "S" perspective as equally valid and discredits / demonizes anyone who believes otherwise.

Ever occur to you that the opposite applies as well? Maybe even more so, because you assume that because the Os believe in the science and engineering of audio that they "rely on tech to deterring the musicality and long term enjoyment?" The truth, of course, is they know that statement is complete, unadulterated nonsense aimed at marginalizing their point of view, that technology is a tool of system development, not a determining factor in musical enjoyment.

Tim

 

[esldude]

Hi,

I'm just curious, do you have any formal education or special training in any electronics or audio related fields, or are your conclusions derived from an armchair perspective? I hope you don't mind my asking.

Thanks.

Welcome to WBF as this appears to be your first post.

Yes I have some formal education in these matters though it isn't how I make my living.

Educated armchair perhaps?

 

[esldude]

What??? You can't look at the peak levels of the key jingling part that way. The peaks are not just made up of ultrasonics but both audible and ultrasonics. As such you can't just compare their levels to purely ultrasonic tones Arny has inserter.

We can take one step toward a more accurate comparison by getting rid of lower frequencies. I did this hear by filtering everything below around 23 Khz:

I put a red line where the highest peak is. If you go to the right you see that its level is below -6 db, not -1 db.

But even that is not enough to make the comparison proper. That peak is only hit one and only one time on that clip (denoted by the red circle). Unless someone has proof that the only way to tell the difference is to listen to that one single transient, then we can't go by that peak either. Remember, Arny's test tones are not a single transient. They are continuous tones.

I eyeballed a level that is likely to hit more of the peaks in blue. That level is now below -15 db (sorry, I chopped off part of the "5" in 15 on the vertical scale). Even that doesn't assure us that we are capturing the areas that showed a difference. But even if it did, compare the level of the blue line to the magnitude of the test tones on the right.

The average RMS amplitude of the key jingling part (post filtering) by the way is -41 dbfs. The average RMS amplitude of the tones is just -6 dbfs. That is a difference of 35 dbfs.


I don't know what a bush league move is . I do know that you can't analyze the high frequency content of a track by just looking at its total amplitude per above. The discussion is about specific part of the spectrum. You cannot take full spectrum numbers and compare them to ultrasonic tones in the context of ultrasonic induced IM distortion. Your analysis must be frequency aware by definition which it was not.


I have run his test and my laptop did not generate any IM distortions even at those elevated levels. Arny has a theory but no data whatsoever that says the reason we can tell the files apart is due to high amplitude ultrasonic tones.

Did you hear a difference right at that peak?

So no, there is no way you can represent those ultrasonic tones as being representative of music or the key jingling part. Not even close. If a system does generate IM distortion with those tones as a result, you have no idea if that is the reason you are hearing a difference. Arny has a theory here. He has no proof that any of this IM talk has any bearing on the results. No cause and effect has been shown. We can throw stuff at the board and see what sticks. Doing that takes better attempts and explanation than what is being put forth.

BTW, I appreciate that you abide by the rules of WBF to keep your posting professional with less personal accusations. We are discussing a hobby and not life and death.

Well this is more honest as a comparison. I don't think it would be beyond some people trying to hear the difference turning things up those few decibels for a few peaks to cause IMD.

For what it is worth I did test my equipment and also had no IMD worth worrying about. Also, yes, the section I listened to and ABX'd with the original files was centered right about that peak. 2-4.9 seconds. But with better resampling the difference went away which does not support the idea it was IMD. Nor am I pushing that idea. But the previous FFT showing ultrasonics vs a steady test tone were deceptive as a way to compare levels. I should have filtered out the sub-20khz as you have. But the difference in peak levels of jangling keys and test tones is still not as great a difference as the FFT would imply .

 

[aronjt]

Welcome to WBF as this appears to be your first post.

Yes I have some formal education in these matters though it isn't how I make my living.

Educated armchair perhaps?

Hi, thanks.

Okay, you must be a different esldude. There is a esldude on another audio forum who has admitted that he has no EE degree or formal education in any field relating to audio, and this other esldude goes around spreading misinformation and harassing other members, many of whom are real Electronic Engineers, with demands that his armchair conclusions be proven false and, if not, then he'll say they're full of BS and that he is correct. Unfortunately, these other members are working professionals and don't have the time to refute the misinformation and never ending demands, so they give up. He can get quite abusive. This other esldude has, for the most part, been identified as a troll and is mostly ignored now on that other site. I'm glad you're not that esldude. Sorry for the confusion.

 

[arnyk]

I'm just curious, do you have any formal education or special training in any electronics or audio related fields, or are your conclusions derived from an armchair perspective? I hope you don't mind my asking.

Speaking as an EE, Esldude seems to "get it" regardless of his formal credentials. The analysis of the "Keys Jangling" wave files turns out to involve some fairly subtle aspects of signals and systems analysis, which are clearly eluding at least one other poster who also has an EE on his CV.

Esldude made one of two very salient points, which seem to be why some people here who may have credentials but lack some critical understandings are encountering severe conceptual and practical difficulties trying to understand a sound that has a very high crest factor.

Esldude's insight is that because the keys jangling sound has a high crest factor, its peak-to-average ratio is relatively high. Here are some relevant summary statistics:



The confusion factor seems to come from the fact that the wave's "Average RMS Power" is almost 30 dB less than its peak power. This is not exceptional, but it is not commonplace either. The usual range for peak to average ratio or Crest Factor is in the 10 to 20 dB range. Percussion instruments make up most of the situations where this number is larger. Notice the spiky envelope shown below:



If someone says that this wave has an amplitude in the -20 or -30 dB range their problem is that they are looking at the average, not the peaks. In high fidelity audio, we have no choice but to reproduce the entire wave, peaks included, as accurately as we can. We don't have 100 wpc amplifiers so we can illuminate 100 watt light bulbs steadily. We have 100 wpc amps to handle the infrequent peaks.

The other confusion factor relates to how FFTs portray signals.

Let's look at this FFT spectral analysis:



This is a mixture of pink noise and a 1 KHz tone. So, which has greater amplitude the pink noise or the 1 KHz sine wave? Seems like a dumb question since that spike at 1 KHz pokes way up and is standing tall over the noise.

Only, that's the wrong answer. The RMS power of the pink noise is about -15 dB, and the RMS power in the 1 KHz sine wave is more like -20 dB. There is about twice as much power in the pink noise as the sine wave. The difference lies in the fact that the noise is incoherent and non-repetitive and spreads its energy over a wide frequency range, while the sine wave is coherent, repetitive, and focuses its energy on just one frequency.

One other thing. When displaying noise, the apparent amplitude of the noise varies with the number of points in the FFT analysis.



Same waveform, different number of points in the FFT (8k versus 64k). Look carefully, virtually all of the apparent statistics for the signal have changed. The sine wave isn't the same amplitude and the noise floor seems to have jumped up.

Oh and BTW, the two highest peaks in the keys jangling file are in the range from 2 to 6 seconds.

 

[Orb]

Well this is more honest as a comparison. I don't think it would be beyond some people trying to hear the difference turning things up those few decibels for a few peaks to cause IMD.

For what it is worth I did test my equipment and also had no IMD worth worrying about. Also, yes, the section I listened to and ABX'd with the original files was centered right about that peak. 2-4.9 seconds. But with better resampling the difference went away which does not support the idea it was IMD. Nor am I pushing that idea. But the previous FFT showing ultrasonics vs a steady test tone were deceptive as a way to compare levels. I should have filtered out the sub-20khz as you have. But the difference in peak levels of jangling keys and test tones is still not as great a difference as the FFT would imply .

Yeah I have to agree as I embarrassingly got caught out in a brain fart-distracted moment there Amir on the first chart.

BTW just to add and going back to earlier post of mine; the result of not hearing IMD ties in with what JA says with his own actual measuring of hardware using ultrasonic test tone.
IMD was only notable at I think -50db to the -0dbfs ultrasonic tones for a couple of the products he tested, reducing the ultrasonic tones to a point that was still stronger than the relevant peak of the jangling keys proved IMD to be negligible; keys being closer to -10dbfs than 0dbfs in terms of relevant peak levels.

Esldude, the problem then is if someone is allowed to turn things up to cause IMD, then unfortunately Arny created a flawed ABX test if his source deliberately (dbfs levels should be set with ultrasonic content in mind) exacerbates it, because at no point was there any instruction on loudness requirements and in fact the ultrasonic tones at the end were encouraged to be played loud until IMD heard; this may sound harsh but quite a few including Arny go on about flawed ABX tests with cues all the time, although this is mostly negated by JAs ultrasonic IM tones test and measurement unless test subject deliberately pushes electronics into clipping/near clipping but that breaks ABX framework anyway.

However that said subjective listening for IMD was only heard it seems by those who actually turned it up loud on the specific 0dbfs ultrasonic tones and not the keys by various posters on AVSF, the point being having to turn it up loud and this has no sub20khz content to balance how much they turn it up. - I am going by memory so could be very wrong about those tones being ultrasonic at end of the test file.
TBH following various forums over the years I do not remember ever seeing it mentioned anyone passing the hirez comparison using Arny's test, and further reducing IMD being an issue is the strong conclusion that there was no difference between hirez/cd/SRC transparency from his experience running this test.
If cues were originally picked up from IMD I would expect that to be reported and the test modified (meaning the source content).

Cheers
Orb

 

[arnyk]

Amir,

I am, of course, not saying that at all. If folks want to rely on tech to determine musicality and long term enjoyment with their system, that's fine with me.

Still misses the point. Many people rely on tech as a means to enhance the musicality and long term enjoyment potential of an audio system. This seems to require a level of abstract reasoning that many can't seem to fathom, which is that there is a relationship between good tech and good musicality and long term enjoyment. This relationship used to be so bad that we had many cases where stuff "Measured good but sounded bad". Our tech has improved to the point where it is common for good measurements and good sound to come in the same package.

My point is that the folks in the "O" crowd seem to not accept / understand the "S" perspective as equally valid and discredits / demonizes anyone who believes otherwise.

Again misses the point which is that there are reliable facts in this world and some things are right and some things are wrong. Perspectives and opinions don't change whether or not Arsenic in large doses is harmful to humans and most other living things.

If you want to hear some of the craziest BS you ever heard, hang out with a bunch of subjectivists. The the thing is, they are so ignorant of the relevant technology that they don't know what they are saying. They use technological terms like they are meaningless tokens.

The lie of subjectivism is that technology is irrelevant and all you need is a good set of ears to assemble a great sounding system. We now know that ears (which are attached to and dominated by brains no mater how much that disturbs some) have strong limitations.

 

[arnyk]

Esldude, the problem then is if someone is allowed to turn things up to cause IMD, then unfortunately Arny created a flawed ABX test if his source deliberately (dbfs levels should be set with ultrasonic content in mind) exacerbates it, because at no point was there any instruction on loudness requirements and in fact the ultrasonic tones at the end were encouraged to be played loud until IMD heard; this may sound harsh but quite a few including Arny go on about flawed ABX tests with cues all the time.

In some fantasy world I would be able to reach out from my home near Lake St Clair across the world and keep people from turning up the volume control on their audio systems so that they create audible IM. The above post is holding me responsible for not being able to do that. This is why I keep talking about having good test proctors, they can keep stuff like that from happening.

 

[jkeny]

In some fantasy world I would be able to reach out from my home near Lake St Clair across the world and keep people from turning up the volume control on their audio systems so that they create audible IM. The above post is holding me responsible for not being able to do that. This is why I keep talking about having good test proctors, they can keep stuff like that from happening.

Exactly why positive & negative controls are a good idea in any test - it goes some way to assuring that the test is being run within the parameters appropriate for the test. In this case the IMD tones could have been included initially so that the listener & you could verify if the equipment being used was suffering from IMD with these tones? I'm sure there are probably other controls that could also be included, no?