Sampling Rates

[Sir Terrence the Terrible]

Here ya go, Terrence, a couple of simple tests for you to peform of audibility of ultrasonic interaction, intermodulations producing audible sound:

take two tweeters that can reproduce ultrasonic sound to, say 30kHz, not difficult I hope. Send one a 25kHz signal and the other one gets 24kHz. No, since you cannot individually hear either one, you seem to claim that they will intermodulate and produce a 1kHz audible sound. Is this what you claim? Please put this simple test to work. While you are at it, use a mic and also measure it.

Mtry, this is probably the most stupid thing you have ever wrote about this subject at least in this thread. I can see this has become an absolutely fruitless excercise in futility. You are missing, and continue to miss my point entirely. And I also made no such claims that you mention in ANY of my posts. You are making this up.

Second one. Try a pet shop that has dog whistles. There should be sample to sample frequency differences between enough to modulate audible frequencies when they intermodulate. Now take these two, make sure they don't produce audible sound individually by themselves, blow then simultaneously and see if there is an audible byproduct. This experiment is Free!

Unbelieveable. that is all I can say.

Don't be surprized if there is no audible byproduct. For this to happen, the air molecules would have to produce audible sound when they interact, collide. What a noizy place this would be full of audibel noize from air molecule collisions.

Test three. Check your components IM distortion which is unwanted intermodulation of different signals. Guess what, the designe is trying to eliminate this, not hear it. Now why would this be desired byproduct of audible tones intermodulating to produce audible harmonics? IT isn't.

Oh, by the way, you have totally misunderstood about the harmonics and believe as that is all it can be, a belief on your part, that I would claim instruments don't produce harmonics but only pure tones? The harmonics are produced by the instrument, not the air molecules intermodulating once they are outside of the instrument. But, who knows, maybe you have discovered something new.

As to ultrasonics, I would be interested in something more solid information than what you seem to want me to believe. Any interesting reading you can point me to?

Mtry,

Based on the so called "experiments" you mention, it is very clear to me that you have missed my point entirely. It was VERY clear in my previous post that I said that we as humans cannot hear a direct impulse above 20khz. What I have stated over and over is that there are instrument harmonics above 20khz(that are out of the range of hearing) that effect the timbre of instruments below that frequency. If the ultrasonic components are filtered out, then it changes the timbre and transient response of the instrument itself within the audible range. I NEVER MADE THE CLAIM THAT WE COULD HEAR DIRECT SIGNALS ABOVE 20KHZ, IS THAT CLEAR TO YOU!!!!!

If you do not believe my assertions, check out AES preprint 3207 by Oohashi, which fully supports my claims and is the basis of my disagreements with you.

It is obvious by your diffused and totally off subject comments above that you are grasping for straws to support your lack of knowledge of this subject, and are trying to steer the argument away from the core comments. These comments while design to simulate knowledge, have no core basis to the subject matter we are discussing. They do nothing to support your arguments whatsoever since they have nothing to so with harmonics, but are primary ultrasonic frequencies with no audible component. It is VERY clear to me that you lack of understanding of the recording process, sample rates, and various other subjects involving digital audio make this discussion(like the ones we had regarding Dts and DD) fruitless. You have absolutely no hands on experience in recording audio, and only a very basic fundimental understanding of digital audio(much of what you gleaned off the internet which is very obvious). Without a more indepth understanding of both the recording process, and digital audio in general, this discussion is fruitless and serves only to confuse the issue, and the people reading this thread. That is not my desire. These so called "experiments" are a total waste of my time, and really miss the point of what I am trying to convey to you. You attempts to "spin" this in another direction only muddle the subject matter, and have totally no relationship to sample rates whatsoever.

If this thread is to bring any understanding of the value of sampling rates, please remain on topic, relevant, and alot more precise in your comments. Or this thread is a useless waste of bandwidth.

[rb122]

If the ultrasonic components are filtered out, then it changes the timbre and transient response of the instrument itself within the audible range. I.

It seems to me (as a layperson and forgive me if I'm misinterpreting) that you originally said (and said again above) that ultrasonic sounds can produce something we can hear in the audible range. It also seems to me that Mtrycrafts reads your claim in a similar manner. If both he and I are off base, can you please clarify? How do ultrasonics affect what we do hear?

I'd like to see this thread continue. Thanks for your help.

[Sir Terrence the Terrible]

It seems to me (as a layperson and forgive me if I'm misinterpreting) that you originally said (and said again above) that ultrasonic sounds can produce something we can hear in the audible range. It also seems to me that Mtrycrafts reads your claim in a similar manner. If both he and I are off base, can you please clarify? How do ultrasonics affect what we do hear?

I'd like to see this thread continue. Thanks for your help.

RB,

A cymbal crash contains measureable content out to 40khz(as do muted trumpets, french horns, and some stringed instruments). These are called harmonics or overtones. The fundimental note, and its harmonic components(overtones) make up the timbre. If you record this cymbal crash, and remove the inaudible ultrasonic energy, it will effect the overall timbre(and transient response) of the cymbal crash within the audible portion of the spectrum. According to the AES preprint 3207 this is an accurate statement fully supported by research, and my experience recording audio. While we cannot hear direct stimulus above 20khz, we DO hear a timbral shift when these ultrasonic components are filtered out.(This is fully documented by Oohashi's paper)

Another argument that I outlined for higher sampling rates lies in the fact that more samples are taken within the audible frequencies. More samples, more resolution, better audio, clearer imaging. Basic digital audio 101. If you have 6-12 samples between 6-8khz, it will sound audibly better than a sample rate that only has 2 samples in the same frequencies. This is the crust of my argument, not ultrasonic theory. Mtry is completely side stepping this issue in favor of one that everyone agrees on, you CAN'T hear a direct impulse over 20khz. You CAN however hear when ultrasonic harmonics are lopped off as a timbral shift.

My arguement is just that clear and simple. A high sampling rate is not a marketing ploy, and I think that everyone that records audio(not the people who write about it)generally agree on that as demonstrated by a polls taken at Surround 2003.

[mtrycraft]

A cymbal crash contains measureable content out to 40khz(as do muted trumpets, french horns, and some stringed instruments). These are called harmonics or overtones. The fundimental note, and its harmonic components(overtones) make up the timbre. If you record this cymbal crash, and remove the inaudible ultrasonic energy, it will effect the overall timbre(and transient response) of the cymbal crash within the audible portion of the spectrum. According to the AES preprint 3207 this is an accurate statement fully supported by research, and my experience recording audio. While we cannot hear direct stimulus above 20khz, we DO hear a timbral shift when these ultrasonic components are filtered out.(This is fully documented by Oohashi's paper)

Nonsense. Did you even read that paper? That paper is so flawed that it should be obvious to even you. Its statisics is next to worthless. Yes, I do happen to have that paper and two others that purport to show audibility of ultrasonic music. Too bad they didn't do a peer -paper.

Not to worry. Get a hold of a much newer paper:

"Detection Threshold for tones above 22kHz," Ashihara Kaoru, et al, preprint 5401, 2001 AEs convention paper

It shows the gross mistakes the previous researchers made, I mean gross. No such audibility is demonstrable, zippo.

Oh, what they heard was IM introduced by the speaker itslef. When multiple drivers were used for the ultrasonics, nothing is audible, no speaker IM. That should have been obvious in the early attempt.

Another argument that I outlined for higher sampling rates lies in the fact that more samples are taken within the audible frequencies. More samples, more resolution, better audio, clearer imaging. Basic digital audio 101. If you have 6-12 samples between 6-8khz, it will sound audibly better than a sample rate that only has 2 samples in the same frequencies. This is the crust of my argument, not ultrasonic theory. Mtry is completely side stepping this issue in favor of one that everyone agrees on, you CAN'T hear a direct impulse over 20khz. You CAN however hear when ultrasonic harmonics are lopped off as a timbral shift.


Too bad that the sampling theorem doesn't agree with you on this. But you are allowed to argue.
You need to get the CD The Digital Audio and see how well you can detect band limiting below 22.05 kHz. Tracks 2-13 has some limiting down to 5.5kz, 11 kHz and 16 khz. Good luck with the 16 kHz..

The CD also has some bits per sample test but not enough.

My arguement is just that clear and simple. A high sampling rate is not a marketing ploy, and I think that everyone that records audio(not the people who write about it)generally agree on that as demonstrated by a polls taken at Surround 2003.

They agree by poll? Is that you evidence? How wonderful. Now I am convinced. That is how audio should be advanced, by polls. All of science should be done by poll. Great!! Hell with credible evidence.

[Sir Terrence the Terrible]

It is clear and apparent that this issue(just like at AES) will not be solved on this forum. For every white paper you turn up that states that any energy over 20khz is insignificant, I can turn over one that says while there is no fundimental signals audible, harmonics that contribute to establish timbre could possibly be. I also stated that more samples in the audible frequencies equal more resolution and better sound. You disagree and stay with the basics even though the basics only give passable results. I state the anti aliasing filters used in most CD and DVD players create distortions and phase shift, you argue they don't even though tests have proven they do contribute to audible distortions. You say oversampling eliminates phase shift and distortions caused by steep filters, however Stanly Black and Julius Dunn research disagrees with that. Even Bob Katz's work disputes you assertions. I think we just need to agree to disagree because we are never going to see eye to eye on this issue. You can believe what you desire, and I'll continue to let my experience and continued education guide me. I personally would rather dispute this with someone who actually has some recording experience rather than someone who does not.

[RGA]

Sir Terrence

Just want to check something. High sampling rates mean better sound is not the case necessarily at the player's end. Sterophile and other publications have said the best DAC on the planet happens to be one that is 1X oversampling unit which means no times oversampling and is the Audio Note Dac 5 - which apparently is based off the very first Sony Phillips units only Audio Note got it right and those two couldn't or wouldn't.

It was Martin Colloms who reviewed it I believe and he is certainly competant in electrical engineering.

[mtrycraft]

Aren't there issues with anti-aliasing filters? E.g. phase shifts that affect audible frequencies? Or have the problems been beaten? Wouldn't it be better to virtually eliminate anti-aliasing by recording much farther up the frequency band?

That is, perhaps the benefit of higher frequencies is not that we actually hear them, but that having them prevents real-world digital record/playback problems in the freqency band that we can hear?

They seem to work just fine, especially with already oversampling data. Your CD player or DVD-A/SACD doesn't need to have such high sampling. They can, just no great benefit.

But, some just claim it without real evidence and when they base it on audibility of ultrasonics, then their whole premis just falls on deaf ears as their premis is just bogus.

[mtrycraft]

It is clear and apparent that this issue(just like at AES) will not be solved on this forum. For every white paper you turn up that states that any energy over 20khz is insignificant, I can turn over one that says while there is no fundimental signals audible, harmonics that contribute to establish timbre could possibly be. .

Ah, now you are hedging again about ultrasonics, "could possibly be"

Why not have some good citation to support this instead of 'could possibly be.'

Yes, psychics 'could possibly be' right.

Please invest in that referenced CD and do a DBT with it. See how well you can differentiate full CD and cut at 16kHz? It should be a snap, 10 of 10 correct responses.

[mtrycraft]

The internet is so fun. I often find it interesting that generally speaking most people will make attacks and snyde comments on computers and not face to face.

On the Internet Mrty, Skeptic and Sir Terrence all seem to be experts on digital. Unfortunately do any of you have any technical background. Pretty sure Skeptic is at least an EE and submitted some patent on something that no one else wanted.

I would say the arguements can stand on their own regardless of background but then you have audio experts like JJ on Audio Asylum who hold the same degree level as Martin Colloms who sits seemingly in opposition and is a world wide renouned expert in the field...and then people disagree wholeheartedly with him and his impressive degree.

The relative layperson is royally screwed because it looks like a bunch of people who seem to know something about science and totally dissagree on practically every point about the actual measurement and of course the testing of human subjects...well we of course know that is not the field to be conducted for an Engineer. But besides that leap out of the qualified field I'd at least expect all the EE's to agree 100% 0n all aspects of all things audio...after all is not 100% of everything measured?

It's actually rather frustrating to read because there almost seems like 3 distinct opinions just from you three and then we add the two people I mentioned who have distinct opinions from each other and also from you three. Add in the other million EE and experts and now it's starting to look like the 6 economists in a room with 6 different answers...of course any science relying on statistical evidence is in big fat trouble but that aside it's highly frustrating.

More so when references don't directly discuss EXACTLY what was being discussed to an exact tee. Hell I even read Tom Nousaine's letters to Stereophile and even he writes a bunch of weasal worded commentary on type 1 and 2 errors. It's softpeddled back and forth. Bleggh - circular statistical clap-trap.

There is no one scientific process - and it appears when my instructor said this today he was correct...if there was you three and these others would all have the exact same result.

I suppose the easiest thing for the layperson would be for you three to lay down your degrees for all of us to see. The best and most prestigious wins. After all a claim made by an expert should be able to show his/her expertise. I'm not going to hire a Doctor who does not have proof he is a doctor...so why should anyone here take advice from non audio experts. Not that we will anyway...but many companies can just throw the non-degree people's resumes in the garbage so it would be nice if I could start a similar root out process.

Most people make buying decisions on anecdotal stuff like reviews. SO if we're going to take someone who is spouting fact then that person should provide their expertise.

Damn Damn frustrating.

Yes, it is frustrating, isn't it. But credentials are irrelevant, really. Just when you thought Dr Hawksford was right about his Essex Echo, here comes Jneutron to show him his gross errors, especially for a Phd.

So, according to you one should have unquestioningly believe Hawksford, right?
I am not asking anyone to believe my posts. They should do their own research. If they want good answers, they will do good research. I have no qualms in relying on JJ and the likes. I am careful though who I listen to, regardless of what they purport to know.

[mtrycraft]

I think I'm just going to stick with what sounds good to my ears.

But you still need to be very careful with that trust. Make sure you trust your ears only, and don't let your eyes bertray it with false senses. That is a well know quantity, not speculations.

[maxg]

But you still need to be very careful with that trust. Make sure you trust your ears only, and don't let your eyes bertray it with false senses. That is a well know quantity, not speculations.

Unless you are happy that the input from your eyes will be continuous throughout your listening life in which case it matters not one jot where the increased enjoyment of your music comes from.

I wonder what proportion of my listening pleasure is derived purely from watching a 12 inch black shiny disk spin as opposed to a 5 inch silver disk, which, in most cases I cannot see?

I have heard that black CD's sound better than silver ones (I have heard the statement - never tried it) - maybe the reasoning is similar - maybe it works best with players that let you see the disk.

Of course the fact I do most of my listening with my eyes closed somewhat weakens the case...

[rb122]

But you still need to be very careful with that trust. Make sure you trust your ears only, and don't let your eyes bertray it with false senses. That is a well know quantity, not speculations.

I agree with Maxg. So long as the false sense (if it is false) is consistent and I'm not spending silly amounts of money on things of dubious-to-zero value, I'm comfortable with my protocol. The areas in which I disagree with the objectivist mainstream seem to be only in areas where even they agree that the sound is different i.e tubes vs SS and vinyl vs CD. So in those cases, I'm going with my preferences.

It seems we have reached an impasse with the technical info on this thread. That's unfortunate but at least both sides have provided us with outside info to check out. So for that, many thanks to Mtrycrafts and Sir Terrence.

[RobotCzar]

"Secondly, there is not an engineer on this planet that would not agree with me that recording at a 96khz sampling rate sounds noticeable better than at 44.1khz.(and 192khz sounds better than 96khz)"

I assume you have not polled every engineer on the planet, or even speak for the majority of them. You are speaking for those who either buy into the hype or have a vested interest in pushing the new formats. That's quite different than every one on the planet. Besides, they need to prove the hear "big" differences in a controlled test, Guess how many are volunteering to do so.

"While you cannot hear above 20khz directly, transient information in some instruments is located above 20khz. If you limit the response of a signal or sharp attack at 20khz(or even 22,050khz), the transient information will sound blurred. With a 44.1khz sampling rate, a brickwall filter MUST be used because high frequency information has to be removed at 22,050khz(which is the Nyquist frequency for 44.1khz) or aliasing will occur out of band, and within the band of human hearing. "

How the heck to do you hear things indirectly? There is NO evidence people can detect "transient information" (isn't all sound transient?) above 20 KHz. So, it doesn't sound "blurred"--we don't hear it at all. You are assuming that I brick wall filter MUST affect response in the audible band AND that such affects would be audible. There is no theoretical, measured. or tested support for those claims. Try to think critcially about what you read in magazines.

[Sir Terrence the Terrible]

Ah, now you are hedging again about ultrasonics, "could possibly be"

Why not have some good citation to support this instead of 'could possibly be.'

Yes, psychics 'could possibly be' right.

Please invest in that referenced CD and do a DBT with it. See how well you can differentiate full CD and cut at 16kHz? It should be a snap, 10 of 10 correct responses.

Umm no I am not hedging, but if thats how you want to look at it, and it makes you feel better, feel free.

I think I have provided enough information, but if I brought more you would conviently claim its bogus, with "gross" errors, so I don't see the point. I know when I have wasted my time thanks.

I have already heard what a rolled off response sounds like, don't need to pay for a free fruitless experience.

I think what is best is to just let this go. I am not going to convince you, and you are certainly not going to convince me, so this whole excercise if fruitless. Its time to move on.

[Sir Terrence the Terrible]

"Secondly, there is not an engineer on this planet that would not agree with me that recording at a 96khz sampling rate sounds noticeable better than at 44.1khz.(and 192khz sounds better than 96khz)"

I assume you have not polled every engineer on the planet, or even speak for the majority of them. You are speaking for those who either buy into the hype or have a vested interest in pushing the new formats. That's quite different than every one on the planet. Besides, they need to prove the hear "big" differences in a controlled test, Guess how many are volunteering to do so.

"While you cannot hear above 20khz directly, transient information in some instruments is located above 20khz. If you limit the response of a signal or sharp attack at 20khz(or even 22,050khz), the transient information will sound blurred. With a 44.1khz sampling rate, a brickwall filter MUST be used because high frequency information has to be removed at 22,050khz(which is the Nyquist frequency for 44.1khz) or aliasing will occur out of band, and within the band of human hearing. "

How the heck to do you hear things indirectly? There is NO evidence people can detect "transient information" (isn't all sound transient?) above 20 KHz. So, it doesn't sound "blurred"--we don't hear it at all. You are assuming that I brick wall filter MUST affect response in the audible band AND that such affects would be audible. There is no theoretical, measured. or tested support for those claims. Try to think critcially about what you read in magazines.

Sorry bud, you are too late for the party. I don't get my information from magazines so there goes your assumption. Never assume.

[Sir Terrence the Terrible]

Sir Terrence

Just want to check something. High sampling rates mean better sound is not the case necessarily at the player's end. Sterophile and other publications have said the best DAC on the planet happens to be one that is 1X oversampling unit which means no times oversampling and is the Audio Note Dac 5 - which apparently is based off the very first Sony Phillips units only Audio Note got it right and those two couldn't or wouldn't.

It was Martin Colloms who reviewed it I believe and he is certainly competant in electrical engineering.

RGA, oversampling at the players end is mtry argument, not mine. I am speaking strictly from a recording end. The designer of the DAC does make the comment on their website that he is critical of the vastly presumptuous calculations generated by the(referring to upsampling and oversampling) techniques.

[RGA]

RGA, oversampling at the players end is mtry argument, not mine. I am speaking strictly from a recording end. The designer of the DAC does make the comment on their website that he is critical of the vastly presumptuous calculations generated by the(referring to upsampling and oversampling) techniques.

Thanks for clarifying - I believe i remember you saying this before come to think of it so that is my laziness not to go back through and check.