Originally posted by Jayfior I don't hear any difference. I can't hear over 13000khz. If anyone can I'd like to have them have a doctor check their ears to find out if they can hear that high frequency.
I don't have the studies to post, but have read them, they tend to say that even though adult humans normally only hear to about 15k at most, when you EQ things in the extreme upper ranges, they can be noticed in mixes. This isn't an answer to the original question, but lends itself as a example that changes in frequencies in upper ranges above the actual hearing frequencies, CAN BE HEARD. I will try to find a couple of the articles, if I can.
Reid
Originally posted by treidm
I don't have the studies to post, but have read them, they tend to say that even though adult humans normally only hear to about 15k at most, when you EQ things in the extreme upper ranges, they can be noticed in mixes. This isn't an answer to the original question, but lends itself as a example that changes in frequencies in upper ranges above the actual hearing frequencies, CAN BE HEARD. I will try to find a couple of the articles, if I can.
Reid
This is because EQ causes phase shift of all frequencies, and we can hear this phase shift within frequencies that we can hear.
You can hear an EQ bump of 3 dB at 50kHz even if there isn't any material above 20kHz. This is easily provable and has been shown many times over such that some of the best EQ algorithms have high frequency controls well above the Nyquist limit.
This is a situation where it is not the addition of higher frequencies that is audible, but rather the mathematical process. Certainly we don't have to record at 96kS/s in order to benefit from this mathematical process if we choose to.
My point in previous post, was just to say, what we think we can hear, may not be the total picture, possibly we may perceive things we don't recognize immediately as something we hear. These sampling rates we are discussing, are simply that, aren't they? We are not getting the complete analog signal in digital music, it is merely lots of samples taken and converted into 1's and 0's, isn't it? So the higher the sampling rate, the closer we get to the true analog signal. I've never heard rates at 96k, only 32, 44.1, and 48. Seems logical that the closer you can get to the true analog signal, though, would be the closer to reality. It could be that equipment, speakers etc. have not progressed far enough yet to utilize the harmonics of the higher sampling rates, yet. As in computers, programs that can truly utilize 32 bit, are just coming into their own, for a while 32 bit was there, but we weren't seeing any real difference in programs or graphics. What is the scientific truths of yesterday are not neccessarily the truths of today, and so what we believe to be true today, may not be so, in the future. What we are discussing today as the possible overkill in sampling rates, will surely one day be considered inferior! We will not reach perfection folks, unless by the grace of God we go to heaven, but as humans we strive to get as close as we can. Reid
[QUOTE]Originally posted by treidm My point in previous post, was just to say, what we think we can hear, may not be the total picture, possibly we may perceive things we don't recognize immediately as something we hear.
It would then not be difficult to put together a scientific study to show that we can indeed hear some things that our simple textbooks don't say that we can. Example, if we can perceive in any form material above 20kHz then it would be possible to put together a very specific A/B/X test to prove that fact. As of yet I have never seen any indication that we humans can hear anything above 20kHz. Remember, the ear is really a digital device. As such, it is not surprising that it has the opportunity to be band limited. It is human nature that likes to think that there is a possibility that there is always more, but as of yet (and I've been researching this for quite some time) there is no evidence to indicate that. I've researched this with audiologists, neurologists, and audio specialists and in their communities there is no evidence that we can hear above 20kHz.
These sampling rates we are discussing, are simply that, aren't they? We are not getting the complete analog signal in digital music, it is merely lots of samples taken and converted into 1's and 0's, isn't it?
...but run through a reconstruction filter that is a phase aligned FIR filter capable of "redrawing" the analog waveform through those points exactly as it went in (based on none-too-simple mathematical laws). Therefore it is too simplistic to think of it in the elementary view of digital sampling as simply being samples of specific voltages at specific moments in time. This very intuitive assessment of the situation undermines the remainder of the work that is done to actually get that basic contraption to work properly.
So the higher the sampling rate, the closer we get to the true analog signal.
Not necessarily. This is again an intuitive but incorrect conclusion. There are actually reasons that higher sampling rates will be WORSE for the audio (more mathematical error due to rounding of coefficients in filters, more HF material passing through a system that has high inter-mod distortion, etc.). To propose that higher sampling rates will somehow be beneficial is to put forth that the Nyquist theory is actually incorrect in regards to being able to capture a waveform accurately below the Nyquist frequency. If the ear can't hear above 20kHz and the Nyquist frequency is indeed correct then why do we need higher sampling frequencies?
I've never heard rates at 96k, only 32, 44.1, and 48. Seems logical that the closer you can get to the true analog signal, though, would be the closer to reality.
This is indeed an area where simple, intuitive logic can be deceiving without a complex understanding of the amazing amount of math that goes into a digital audio system. It's not nearly as simple as meets the eye.
It could be that equipment, speakers etc. have not progressed far enough yet to utilize the harmonics of the higher sampling rates, yet.
I have both microphones and monitors that are spec'd flat up to 40kHz and I have recording equipment that is capable of recording up to 96kS/s. There are definitely advantages of transducers being able to reproduce higher frequencies (just as we talked about with the EQ's earlier in this post) but that has nothing to do with being able to benefit, somehow, from higher frequency listening. It has everything to do with listening to material in-band in phase.
As in computers, programs that can truly utilize 32 bit, are just coming into their own, for a while 32 bit was there, but we weren't seeing any real difference in programs or graphics. What is the scientific truths of yesterday are not neccessarily the truths of today, and so what we believe to be true today, may not be so, in the future. What we are discussing today as the possible overkill in sampling rates, will surely one day be considered inferior! We will not reach perfection folks, unless by the grace of God we go to heaven, but as humans we strive to get as close as we can. Reid
I think that you'll find that what we're talking about here isn't as simple as Newtonian physics. It's actually very complex, and with an understanding of it's mathematical complexity it is very possible to make the conclusion that, based on a specific understanding of the ear's capabilities, we can mathematically conclude EXACTLY what is necessary to record and playback audio to meet those needs, including dynamic range, frequency response, sensitivity to noise in certain bands, bandpass modulations, etc. If we can understand the boundaries of the ear in all of these capacities then we can most assuredly design a system that is capable of performing exactly to that standard without fault, no? As we understand the ear now (and we have a pretty good understanding of all of the parameters I've listed above, between Fletcher-Munson curve and other studies we've got the ear mapped out very well) we have a system that is capable of mathematically meeting those needs. Someone will have to prove that either A. the math is wrong, or B. our understanding of the ear is wrong before there is reason to change the current paradigm. With as much marketing as is going on right now about higher sampling frequencies, nobody stepping up in either of these two capacities indicates that the current system is actually NOT defective, but does not live up to the marketing needs of a society who falls prey to the apparently very obvious conclusion that more is always better. I've been studying this stuff for a while and done several listening tests myself. I'm waiting for anything to indicate that the marketing engine is right and that the real brains behind the operations are wrong.
Nika:
All I can say is , sorry ma'am, I didn't mean to get you upset. I really wasn't talking about the theory of gravity? I don't know where you got that impression, but I will try to stop this by saying, I'm sure you are right, and god bless all. Reid
One minor point of clarification is that my name is Nika, short for Nicholas. My father's name is "Nick" and they didn't want to get us confused so they named me "Nika" since birth.
As for the comment about gravity, I think a lot of people are taught about digital sampling that "so long as you sample twice in each waveform you can accurately capture the frequency" which is a fairly simple thing to see. You can draw it out on paper and very visually see that if you can sample once in each half cycle that you should be able to capture frequencies accurately.
I think a lot of people even wonder why it took until the 20th century for anyone to discover this. Isn't it obvious? This simplistic explanation to me is about as simple as Newtonian physics, and this is why I said that it's NOT that simple.
Nyquist didn't say that "so long as you sample twice in each waveform you can accurately capture the frequency of the waveform." He said that you can capture EVERYTHING about the waveform, from the frequency to the amplitude to the phase. Now THIS is where it gets very non-intuitive, but it is true: while only sampling two points per highest frequency you wish to capture in a waveform it is actually possible to completely redraw the waveform with complete detail, including amplitude, phase, and frequency, with no error. It's not very obvious why this is the case, but it has to do with some pretty lofty mathematics and a principle that tells us that it's actually only possible to redraw a band limited and filtered sine wave through two given points in one way.
I'm just pointing out that the simplistic version of the Nyquist theorem that we're given in college, combined with the quick deductions that we make of it is indeed as simple as Newtonian physics. Nyquist, on the other hand, drew much more consequential mathematical conclusions than that, and I was merely pointing out how complex all of that is.
Does that explain any better?
Don't worry, I'm not upset. It's just an area I've become very interested in.
Sorry about the name thing, I thought it was a female name.
It seems as though you are trying to show how much you know, and how little I do, which is probably true in this area, but as a moderator, you need to use more tact in how you respond to the everyday joe that comes to forums. It seems like no matter how much we think we know, and we are at the pinacle of what ever it is, change will happen. Many times what we thought was so concretely right, has been replaced by new theorems and technology. I've just never been one to say, always or never, if I can avoid it. Things can be so subjective. I go to many other forums, and in general, most are quite friendly. I don't buy alot from Sweetwater, so if I stop, it won't matter, and I'm sure I wouldn't be missed. I like to learn what I can, and voice my opinions when I can. I've seen posts from sound engineers, debating this issue. I certainly can't respond with the information and expertise that I've seen, used by them on other forums. But I have seen a contrary opinion to yours, and they were from currently employed sound engineers. I wasn't trying to get you wound up, just talking, but everything seems so black and white, and you are right, kinda attitude, with you. Again, I'm not an engineer, just a recording musician, but I think, maybe I should reconsider using Sweetwater as a source. I train younger people in my trade from time to time, and it's not easy to teach without making the apprentice feel dumb, or sounding like you know it all. It takes tact and patience. I don't feel to comfortable voicing my opinion now, I'm sure it will be picked apart, so see all you other guys, and God Bless.
Reid
Thank you for your email. I can understand your perpsective.
First, about the name thing, don't sweat it. You can imagine that the mistake is made often. Nika is indeed a female name in many cultures.
As for the sampling frequency thing, your post makes sense. Please allow me to give you my perspective so you know where I'm coming from.
While it certainly appears that the past few posts have been a personal discussion between you and I (hello anybody???.......echo.......echo.......) this site is published in a public manner and many people read it. I, too, am a member of many other forums and it's remarkable how many people read the forums without responding. Things that are written are viewed by MANY people (I think that the forums at Sweetwater get several hundred thousand hits a month) and unfortunately noone is really around to edit for correctness as in magazines and newspapers. It is very common for people to walk away with very incorrect information by reading forums where unqualified people have presented what seem like logical and intuitive arguments that are incorrect. It seems like forums have great potential to disseminate mis-information as much as they have the potential to disseminate valid and correct information.
I, personally, find this to be unfortunate. I am not so much interested in the discussion for the sake of you, Reid. I am more interested in setting the record straight for the benefit of the many other people who will be reading this thread.
The area of higher sampling frequencies happens to be amongst the biggest marketing blitzes I know of in recent times, and along with that is a substantial amount of misinformation - MUCH OF WHICH is perpetuated by the manufacturers themselves. The forums are one of the areas where the information seems to spin out of control and I can only do my best with the short time I have to put in the equal/opposite perspective where I can. I think it would be unfortunate if you or many other people purchased higher sampling frequency equipment because of the very logical arguments you have presented. I apologize if you don't care for the manner in which I presented this case. For what it's worth, I was adamantly arguing your case not all that long ago and then a few whiz bang engineers explained to me in no uncertain terms how I was incorrect. I followed it up with a lot of research that, when the marketing was removed from the situation, substantiated their case in a fairly undeniable fashion.
Now don't get me wrong - there is indeed some validity to the notion of higher sampling frequency recording, and I'd be happy to go into some of that if we want to.
I may see you around. I'm out there quite a bit.
Cheers!
Nika.
P.S. I thought you might enjoy reading this thread. I'm far from being what I'd call an "expert" in a lot of this stuff. When people point that out to me comprehensively is when I happen to learn the most.
I, personally, find this to be unfortunate. I am not so much interested in the discussion for the sake of you, Reid. I am more interested in setting the record straight for the benefit of the many other people who will be reading this thread.
If you have absolute data, and want to teach or set the record straight, good. Your method though, is more like a cop, putting up a sign that said, "what is the speed limit here?" Then he pulls you over and says, "You are speeding and I'm giving you a ticket." You posed a question, I assumed to get information, or opinions on the issue, not as a sneaky way to get your agenda out, right? If you wish to teach or share knowledge, you don't have to ask questions, wait for some average guy, giving his opinion, and then show what you know, you can just state things on the forum, you are a moderator, right? I just don't like your method, that's all!
Reid
Hmm. I see your perspective. I didn't have anything to do with the question, though. I'm just another guy who saw the question when it was posted. I'm not "ALLMIGHTY SWEETWATER". I'm just a guy who works there and who has been asked to moderate the Microphone forum (thus I'm identified as a "moderator" whereever on here I post).
I see your point, though. It does indeed look a little odd.
Nika:
I just saw Sweetwater sound as the starter of topic and you a moderator by your name. Felt like I got ambushed. Hmmmm I looked closer and see that DrGroove is the actual moderator of this room. I'm new to Sweetwater forums, I just bought some stuff, and the sales rep., told me about the forums. I just hope, the moderators will try to be gentle and not to condescending when discussing subjects, and if they feel they have indisputable facts on a subject, just state them. But I jumped the gun, as you are not the moderator of this room, and for that I apologize, you have as much right to come into a room, and state whatever you want.
Reid
Re: Can you or can you not hear the difference between 48k and 96k? (06-08-01)
Nika:
Don't sweat it! I over reacted. But think of this for just a moment, this was the original question...
Originally posted by SweetwaterSound Can you or can you not hear the difference between 48k and 96k?
Now if the answer is indisputable, but because of question, you draw some in to say, "I can hear a difference at 96", then you tell him why he can't, you have basically called him a liar. I just think the question is setting some forum users up, if, you feel you have unquestionable proof, state it, don't pose a question to get agenda out.
Reid
I think that the question is indeed a bit vague. Can you hear the difference between 48k and 96k on what? On really bad converters that have deplorable filters that roll off and phase shift signals well into the audible range? Or are we talking about being able to hear the difference in ideal obtainable situations (really nice converters, monitoring system, etc.)
I think that the bottom line is that ANYONE could hear the difference if the converters in question are poor enough (and there are some other situations as well) but if the question is read as "is there an INHERENT audible difference (as in with ANY converter) between the sample rates" then the answer is "no".
I'll be the first to admit that there CAN be a difference, but as the components get better the difference goes away. Therefore there is not an "inherent" difference. The difference is provided by poor analog sections and poor filtering. These are not an artifact of lower sampling, but lesser quality manufacturing. There's a fine line between those.
I get what you are saying, and have a better understanding from all the previous posts. The question though, seemed to be an empirical one, based on observation or experience, without regard to the system or theory. We gotta stop meeting this way.
Reid
Reply With Quote
Reid
Bookmarks