Home |
Search |
Today's Posts |
#41
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
In article , KH
wrote: On 9/28/2013 1:30 PM, Audio_Empire wrote: In article , Andrew Haley wrote: snip Double-blind testing works for everything else, as far as I know. I'm not going to accept any special pleading (sans really good evidence) that it may not be applicable to audio. How would you prove such a thing, anyway? I don't pretend to know. How do you prove that it DOES work for audio? Since it usually returns a null result, I'd say such overwhelmingly one-sided results indicates one of two things: either everything does sound the same (which my experience tells me is extremely unlikely), or DBTs aren't good at uncovering differences in audio gear unless they are extremely gross differences. We certainly know which of those two outcomes the "strict objectivists" believe in, but how do we prove which is the real answer? Well, you start out by demonstrating that there are objectively verifiable differences between units - through measurement - that at least approach the demonstrated lower threshold of human hearing. That would help. But you can't measure everything that could affect audio performance. How do you measure, for instance, a DAC's ability to image well? How can you measure bass performance that MIGHT be affected by such things as impedance match to downstream (or perhaps even upstream) components? If all measured artifacts or differences are, say -110db down, then a null DBT *is* expected, and a subjectively "verified" audible difference is clearly suspect. But not all differences heard in listening are a result of a frequency response reading or a S/N ratio reading. What is the "threshold of hearing", for instance, for phase anomalies which might affect image specificity? What is the "threshold of hearing" for jitter? How does that affect what the listener hears? How do you interpret what you observe on an oscilloscope about the ringing wer DAC digital filters can do) to affect the sound - if at all? The problem with your solution is that not everything seems to be easily quantifiable and many of the things that are quantifiable are not easily correlate-able to what one hears (in either a positive or a negative sense). Also, many things that people used to put a lot of stock in (like THD in amplifiers) turn out to be not very important, because it seems that even though a THD figure on an amplifier of more than 1% looks bad on paper, the ear is not really all that sensitive to it. OTOH, things like crossover notch distortion, and all types of Intermodulation distortion in amplifying devices are very noticeable, even to people who are just casual l isteners. But I digress. The point is that "the threshold of hearing" is only helpful or establishing the audibility or inaudibility of SOME performance parameters, but not ALL by a long shot. |
#42
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
In article ,
ScottW wrote: On Friday, September 27, 2013 4:04:33 PM UTC-7, Audio_Empire wrote: snip http://www.jensen-transformers.com/an/an004.pdf Bottom line. Make sure your cables have real ground wires....not just foil shield drains....or better yet, tie all your chassis grounds together in a star configuration with some lamp cord. ScottW OK, Scott. That can affect noise in some cases, I agree. But that has to be a system-wide solution, and would not be noticed by just swapping out, say, one interconnect for another. |
#43
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
In article ,
"Arny Krueger" wrote: "Audio_Empire" wrote in message ... Obviously, it easier to prove that subjectivists can hear no differences in a properly set-up DBT than it is to prove that objectivists CAN hear differences when said objectivists DON'T WANT TO hear differences. This argument fails because to be true, it demands that no person with good will and basic honesty has ever done an audio DBT. It doesn't have to have happened. It just has to be a possibility. Nothing prevents so-called subjectivists from doing DBTs .I it were just a matter of having a subjectivist do a DBT in order to obtain positive results, then such a thing would have had to happen at last once in the past 30 or more years. The problem here is that many subjectivists don't believe that DBTs are very good at picking-out subtle, but important (to a dyed-in-the-wool- audiophile) differences between audio components. In fact many so-called subjectivists have participated in DBTs and they obtained the same results as so-called objectivists. DBTs have a proven track record of converting so-called subjectivists into objectivists. Didn't convert me, and I've been privy to a number of ad hoc and formal DBTs. |
#44
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"Audio_Empire" wrote in message
... In article , KH Well, you start out by demonstrating that there are objectively verifiable differences between units - through measurement - that at least approach the demonstrated lower threshold of human hearing. Agreed. That would help. But you can't measure everything that could affect audio performance. Except you can. It's not 1955 or 1965. Ever since the late 1960s audio measurement technology has evolved rapidly yet most audiophiles are unaware the updates which I will briefly cover below. Ever since the 1980s our understanding of audio perception has evolved rapidly but again most audiophiles are unaware of the updates involved. The latter are probably better understood by young people because among other things they relate to why perceptual coding works as well as it does. But they have broad relevance to our understanding of the significance of technical performance to listening enjoyment and this does not seem to be understood as well. Audio signals have only 2 dimensions per channel. Therefore everything that can go wrong with them can be quantified with a limited number of variables. These 3 Error Variables a (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as harmonic distortion, intermodulation distortion, and frequency or phase distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering signals. Believe it or not, that is it! How do you measure, for instance, a DAC's ability to image well? First off, you look at the problem from the other way. The question you ask can be restated as: "What cases a DAC to image poorly?" The answer is some combination of the 3 Error Variables stated above. How can you measure bass performance that MIGHT be affected by such things as impedance match to downstream (or perhaps even upstream) components? The answer is some combination of the 3 Error Variables stated above. If all measured artifacts or differences are, say -110db down, then a null DBT *is* expected, and a subjectively "verified" audible difference is clearly suspect. But not all differences heard in listening are a result of a frequency response reading or a S/N ratio reading. Of course, I have pointed out that there are 3 error variables, only 2 of which are related to frequency response (linear distortion) and noise. The third error variable that needs to be considered as well is nonlinear distortion. What is the "threshold of hearing", for instance, for phase anomalies which might affect image specificity? Phase anomalies are part of linear distortion. What is the "threshold of hearing" for jitter? It is a known quantity which is an amplitude that depends on the frequency and amplitude of the desired signal and the frequency and amplitude of the modulating signal which are usually different frequencies but don't have to be. How does that affect what the listener hears? The listener reliably senses a difference between the distorted signal and a reference undistorted signal. It really does not matter how he hears it. The important thing is that he is reliably able to detect it. How do you interpret what you observe on an oscilloscope about the ringing which DAC digital filters can do) to affect the sound - if at all? Ringing is a consequence of linear distortion. The problem with your solution is that not everything seems to be easily quantifiable and many of the things that are quantifiable are not easily correlate-able to what one hears (in either a positive or a negative sense). That would appear to be a misapprehension based on a lack of knowledge of the principles of modern signals analysis and psychoacoustics. The questions that I answered would not be asked by a person who was up to date in these important fields that have been largely developed from the 1970s to the 1990s. Also, many things that people used to put a lot of stock in (like THD in amplifiers) turn out to be not very important, Harmonic distortion is not as important as it used to be because modern equipment often has vanishing amount of it. This was not always so. because it seems that even though a THD figure on an amplifier of more than 1% looks bad on paper, the ear is not really all that sensitive to it. Depending on the details of the harmonic distortion (order or nonlinearity, amount and signal frequency involved) harmonic distortion can be reliably detected below 0.1%, but not below 0.01%. There is a psychoacoustic reason for this, but it is beyond the scope of this brief article. OTOH, things like crossover notch distortion, and all types of Intermodulation distortion in amplifying devices are very noticeable, Crossover distortion is just another kind of nonlinear distortion. If you understand the basic mathematics of nonlinear distortion you know that any nonlinearity that causes IM also causes Harmonic Distortion. The point is that "the threshold of hearing" is only helpful or establishing the audibility or inaudibility of SOME performance parameters, but not ALL by a long shot. That would appear to be yet another misapprehension based on a lack of knowledge of the principles of modern signals analysis and psychoacoustics. If one wishes to make authoritative-sounding statements about audio, it helps to have kept up with audio's evolving technology. |
#45
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On 10/3/2013 5:38 AM, ScottW wrote:
On Monday, September 30, 2013 3:53:47 PM UTC-7, Audio_Empire wrote: In article , KH The point is that "the threshold of hearing" is only helpful or establishing the audibility or inaudibility of SOME performance parameters, but not ALL by a long shot. You've only just begun to explore the level of difficulty of this approach. Throw in establishing threshold under various masking stimuli...such as music. Masking makes the difference *harder* to hear, not easier. Then try to consider multi-parameter interaction on hearing threshold and quickly the problem becomes impossible to establish any correlation to actual listening. Far easier to toss the "can't hear a difference sighted" subjects from the pool. If they hear no difference sighted, that could be bias as well, or they may not be suitable. That's true. If you really want to challenge the efficacy of DBT's, then taking subjects that *reliably* hear a difference between A and B, a difference that engineering and physics says should not be acoustically distinguishable, and subjecting them to a DBT where *ALL* test parameters are *exactly* the same with the exception of blinding. That is *all* that is necessary to demonstrate that sighted listening is unreliable. You can argue DBT methodology, but the study above differs only by blinding so there's no methodological uncertainties involved. Keith |
#46
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
In article ,
"Arny Krueger" wrote: "Audio_Empire" wrote in message=20 ... =20 In article , KH =20 Well, you start out by demonstrating that there are objectively verifiable differences between units - through measurement - that at least approach the demonstrated lower threshold of human hearing. =20 Agreed. =20 That would help. But you can't measure everything that could affect aud= io=20 performance. =20 Except you can. It's not 1955 or 1965. Ever since the late 1960s audio=20 measurement technology has evolved rapidly yet most audiophiles are unawa= re=20 the updates which I will briefly cover below. Ever since the 1980s our=20 understanding of audio perception has evolved rapidly but again most=20 audiophiles are unaware of the updates involved. The latter are probably= =20 better understood by young people because among other things they relate = to=20 why perceptual coding works as well as it does. But they have broad=20 relevance to our understanding of the significance of technical performan= ce=20 to listening enjoyment and this does not seem to be understood as well. =20 Audio signals have only 2 dimensions per channel. Therefore everything t= hat=20 can go wrong with them can be quantified with a limited number of variabl= es. =20 These 3 Error Variables a =20 (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as=20 harmonic distortion, intermodulation distortion, and frequency or phase= =20 distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering=20 signals. Believe it or not, that is it! I believe it. I've seen these parameters outlined before, I didn't just fal= l off the audio milk train, you know. The point is, they don't mean much to= the average audio enthusiasts though. Again it's not so much that these te= sts exist, as it is that they aren't readily available, easily understandab= le, or easily correlated to what people hear. When I said "But you can't me= asure everything that could affect audio performance." I was responding to = Keith Hughes and the other audio enthusiasts who post here and who don't h= ave access to either the equipment or the methodologies needed to measure t= hese three major error variable categories (or their sub-categories) and m= ost likely wouldn't understand how these parameters affect what they hear, = even if they did. So, in the hobby world, we're back to square one which m= eans our ears.=20 How do you measure, for instance, a DAC's ability to image well? =20 First off, you look at the problem from the other way. The question you a= sk=20 can be restated as: "What causes a DAC to image poorly?" I don't pretend to know for sure. I don't work in the field of digital data= quantization. I have some textbooks on the subject, but they=20 don't cover stereo imaging of a digital to analog converter. I suspect that= it has to do how the converter circuits are used to recover the left and right channels, as I have noticed that Stereo DACs seem to image better than do single switched mono DACs, and that differential DACs seem t= o image best of all. But hell, I don't even know that for sure. From observ= ation, it just SEEMS that way. But we all know the pitfalls' of that kind o= f assumption.=20 =20 The answer is some combination of the 3 Error Variables stated above. Sure, but again, that's not helpful to the average enthusiast. He can't mea= sure it, he won't hear it in a DBT, and to my knowledge, no DAC manufacture= r advertises: "If you value soundstage presentation and image specificity,= don't buy our products, because they aren't good at that".=20 =20 How can you measure bass performance that MIGHT be affected by such things as impedance match to downstream (or perhaps even upstream) components? =20 The answer is some combination of the 3 Error Variables stated above. If all measured artifacts or differences are, say -110db down, then a null DBT *is* expected, and a subjectively "verified" audible differen= ce is clearly suspect. These answers aren't helpful in the least, Mr. Kruger. Because there is not= hing about this subject that the average listener can wrap his head around,= and allow him to make an informed selection. Unfortunately, audio manufact= urers know this and it allows them to make extremely (and unnecessarily) ex= pensive components that probably sound no different from other examples of = the same types of components that cost far less.=20 =20 But not all differences heard in listening are a result of a frequency response reading or a S/N ratio reading. =20 Of course, I have pointed out that there are 3 error variables, only 2 of= =20 which are related to frequency response (linear distortion) and noise. Th= e=20 third error variable that needs to be considered as well is nonlinear=20 distortion. Agreed, but again, this is just an academic exercise to most enthusiasts. T= hey don't have a lab fill of state of the art test equipment or the enginee= ring savvy to understand what it all means. Most audio enthusiasts just wa= nt the best sound for their money, and the only tools really available to t= hem are their own senses. =20 =20 =20 What is the "threshold of hearing", for instance, for phase anomalies which might affect image specificity? =20 Phase anomalies are part of linear distortion. Doesn't answer the question I asked for a quantification - but don't bother= , it was a rhetorical question anyway. =20 What is the "threshold of hearing" for jitter? =20 It is a known quantity which is an amplitude that depends on the frequenc= y=20 and amplitude of the desired signal and the frequency and amplitude of th= e=20 modulating signal which are usually different frequencies but don't have = to=20 be. Again it was rhetorical and meant to show that these numbers have little me= aning to the average audio enthusiast.=20 How does that affect what the listener hears? =20 The listener reliably senses a difference between the distorted signal an= d a=20 reference undistorted signal. It really does not matter how he hears it. = The=20 important thing is that he is reliably able to detect it. How do you interpret what you observe on an oscilloscope about the=20 ringing which DAC digital filters can do) to affect the sound - if at= =20 all? =20 Ringing is a consequence of linear distortion. Again, not the question, but also again, more or less rhetorical.=20 The problem with your solution is that not everything seems to be easily quantifiable and many of the things that are quantifiable are not easily correlate-able to what one hears (in either a positive or a negative sense). =20 That would appear to be a misapprehension based on a lack of knowledge of= =20 the principles of modern signals analysis and psychoacoustics. The questi= ons=20 that I answered would not be asked by a person who was up to date in thes= e=20 important fields that have been largely developed from the 1970s to the= =20 1990s. No, it's not, it's a real world question valid to those who have no way of = accessing modern signal analysis methodologies or the knowledge needed to u= nderstand or correlate them. In other words, virtually the=20 entire audio hobby community.=20 Also, many things that people used to put a lot of stock in (like THD in amplifiers) turn out to be not very important, =20 Harmonic distortion is not as important as it used to be because modern= =20 equipment often has vanishing amount of it. This was not always so. =20 because it seems that even though a THD figure on an amplifier of more than 1% looks bad on paper, the ear is not really all that sensitive to it. =20 Depending on the details of the harmonic distortion (order or nonlinearit= y,=20 amount and signal frequency involved) harmonic distortion can be reliably= =20 detected below 0.1%, but not below 0.01%. There is a psychoacoustic reaso= n=20 for this, but it is beyond the scope of this brief article. Actually I've seen amplifiers with almost unbelievably high levels of THD (= like 1%) even at relatively low power levels (as opposed to near clippin= g) that no one could hear when the amp was playing music. One French amp I = remember was a 70 Watt/channel amp that had 2.5% THD at 10 Watts! It sound= ed fine and for a while, was considered (by TAS) one of the best sounding a= nd most musical (and expensive) amps on the market. Yet a 1st generation Dy= naco ST-120 had a tiny amount of crossover distortion at low power levels a= nd sounded just dreadful. Dyna fixed it later, And I "fixed" mine by simply= re-biasing the drivers further into class A. (not a good fix, BTW. The cro= ssover notch was gone but the driver transistors started blowing often - ta= king the 2N3055 output devices with them).=20 OTOH, things like crossover notch distortion, and all types of=20 Intermodulation distortion in amplifying devices are very noticeable, =20 Crossover distortion is just another kind of nonlinear distortion.=20 Championing the obvious, now , are we? If you understand the basic mathematics of nonlinear distortion you know = that any=20 nonlinearity that causes IM also causes Harmonic Distortion. Of course it does.=20 |
#47
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"Audio_Empire" wrote in message
... In article , "Arny Krueger" wrote: Audio signals have only 2 dimensions per channel. Therefore everything that can go wrong with them can be quantified with a limited number of variables. These 3 Error Variables a (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as harmonic distortion, intermodulation distortion, and frequency or phase distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering signals. Believe it or not, that is it! The point is, they don't mean much to the average audio enthusiasts though. I have a limited platform from which to educate audio enthusiasts. I do what I can. Again it's not so much that these tests exist, as it is that they aren't readily available, easily understandable, or easily correlated to what people hear. Every one of the 3 items can be measured to well below the thresholds of audiblity using time-honored testing methodogies going back 30-50 years of more. Most audiophiles have the hardware and can freely download software platforms for measuring them. How available do they need to be? When I said "But you can't measure everything that could affect audio performance." I was responding to Keith Hughes and the other audio enthusiasts who post here and who don't have access to either the equipment or the methodologies needed to measure these three major error variable categories (or their sub-categories) I think they do. Got a modern PC with an audio interface? and most likely wouldn't understand how these parameters affect what they hear, even if they did. In these days of university level courses being freely available on the web, any ignorance is the fault of the ignorant. So, in the hobby world, we're back to square one which means our ears. How do you measure, for instance, a DAC's ability to image well? First off, you look at the problem from the other way. The question you ask can be restated as: "What causes a DAC to image poorly?" I don't pretend to know for sure. I dunno. Kevin gave you the right answer and you seem to be less than fully cognizant of it: KH wrote: "If all measured artifacts or differences are, say -110db down, then a null DBT *is* expected, and a subjectively "verified" audible difference is clearly suspect." Of course he's being wildly optimistic about the hearing threshold of the human ears. Reality is more like 80 dB at the best and 60 dB more typical. Audiophiles by the 100s are overcoming the most difficult measurment hurdles being acoustic measurements. |
#48
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On 10/3/2013 7:40 PM, Audio_Empire wrote:
In article , "Arny Krueger" wrote: "Audio_Empire" wrote in message ... In article , KH Well, you start out by demonstrating that there are objectively verifiable differences between units - through measurement - that at least approach the demonstrated lower threshold of human hearing. Agreed. That would help. But you can't measure everything that could affect audio performance. Except you can. It's not 1955 or 1965. Ever since the late 1960s audio measurement technology has evolved rapidly yet most audiophiles are unaware the updates which I will briefly cover below. Ever since the 1980s our understanding of audio perception has evolved rapidly but again most audiophiles are unaware of the updates involved. The latter are probably better understood by young people because among other things they relate to why perceptual coding works as well as it does. But they have broad relevance to our understanding of the significance of technical performance to listening enjoyment and this does not seem to be understood as well. Audio signals have only 2 dimensions per channel. Therefore everything that can go wrong with them can be quantified with a limited number of variables. These 3 Error Variables a (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as harmonic distortion, intermodulation distortion, and frequency or phase distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering signals. Believe it or not, that is it! I believe it. I've seen these parameters outlined before, I didn't just fall off the audio milk train, you know. The point is, they don't mean much to the average audio enthusiasts though. Again it's not so much that these tests exist, as it is that they aren't readily available, easily understandable, or easily correlated to what people hear. When I said "But you can't measure everything that could affect audio performance." I was responding to Keith Hughes and the other audio enthusiasts who post here and who don't have access to either the equipment or the methodologies needed to measure these three major error variable categories (or their sub-categories) and most likely wouldn't understand how these parameters affect what they hear, even if they did. So, in the hobby world, we're back to square one which means our ears. The point, though, is that it would be relatively easy to measure two devices, such as two DACs for which sighted "differences" have been found, show that all of these three types of distortion are, say 120db down, and show that those same listeners cannot distinguish between them under blinded conditions. Now, is this a useful thing for the "average audiophile"? I would argue yes. Not in the sense that one would want to do this type of testing for every device prior to purchase, but to understand that sighted bias can be very powerful, and cannot be "put aside" as some like to believe. This is useful to the average Joe when he's looking at two or three components in a shop at X, 3X and 13X price ranges - components that have no technical reason to sound different, and yet do when demo'd in the store. Keith |
#49
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On 10/4/2013 8:12 AM, Arny Krueger wrote:
"Audio_Empire" wrote in message ... In article , "Arny Krueger" wrote: snip I dunno. Kevin gave you the right answer and you seem to be less than fully cognizant of it: KH wrote: "If all measured artifacts or differences are, say -110db down, then a null DBT *is* expected, and a subjectively "verified" audible difference is clearly suspect." Of course he's being wildly optimistic about the hearing threshold of the human ears. Reality is more like 80 dB at the best and 60 dB more typical. Not being optimistic. Just pragmatic. At -110db, even the most rabid naysayer would have to admit his golden ears aren't that good. Keith |
#50
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"KH" wrote in message
... On 10/4/2013 8:12 AM, Arny Krueger wrote: "Audio_Empire" wrote in message ... In article , "Arny Krueger" wrote: snip I dunno. Kevin gave you the right answer and you seem to be less than fully cognizant of it: KH wrote: "If all measured artifacts or differences are, say -110db down, then a null DBT *is* expected, and a subjectively "verified" audible difference is clearly suspect." Of course he's being wildly optimistic about the hearing threshold of the human ears. Reality is more like 80 dB at the best and 60 dB more typical. Not being optimistic. Just pragmatic. At -110db, even the most rabid naysayer would have to admit his golden ears aren't that good. I feel that way about the -100 dB number, which BTW is reflected in ITU Recommendation BS 1116. It's definately an overkill number just like 44.1 KHz and 16 bits are overkill sampling for digital music. |
#51
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On 10/4/2013 8:58 AM, ScottW wrote:
On Thursday, October 3, 2013 7:37:50 PM UTC-7, KH wrote: On 10/3/2013 5:38 AM, ScottW wrote: On Monday, September 30, 2013 3:53:47 PM UTC-7, Audio_Empire wrote: In article , KH The point is that "the threshold of hearing" is only helpful or establishing the audibility or inaudibility of SOME performance parameters, but not ALL by a long shot. You've only just begun to explore the level of difficulty of this approach. Throw in establishing threshold under various masking stimuli...such as music. Masking makes the difference *harder* to hear, not easier. Exactly...it impacts the threshold. We'll establish a low threshold without it and then complain that music provides null results and conclude DBTs invalid as a concept. Go figure. But *I* was discussing subjects that purportedly *could* hear a difference sighted. Then try to consider multi-parameter interaction on hearing threshold and quickly the problem becomes impossible to establish any correlation to actual listening. Far easier to toss the "can't hear a difference sighted" subjects from the pool. If they hear no difference sighted, that could be bias as well, Great...a biased test subject whose bias is difficult to control for. Why would you want them in the pool of test subjects? Makes no sense. You know, it's pretty silly to try create disagreement by selectively parsing a reply: or they may not be suitable. That's true. What did you *think* "that's true" meant? My next paragraph stipulates that you need test only listeners that *do* hear a sighted difference. If you really want to challenge the efficacy of DBT's, then taking subjects that *reliably* hear a difference between A and B, a difference that engineering and physics says should not be acoustically distinguishable, and subjecting them to a DBT where *ALL* test parameters are *exactly* the same with the exception of blinding. That is *all* that is necessary to demonstrate that sighted listening is unreliable. Thriller....and all along I thought the question was controlling for same bias to assure blind tests were not corrupt. Silly me. Can't argue with your last sentence. If physics says there should be no audible difference between A and B, sighted listeners hear a difference, and blinded they cannot, does that mean DBT's work in all situations under all conditions? No it doesn't, but that's of precisely zero concern. It shows it works in the context in which it is of value. You're ignoring the obvious premise of using a DBT - if no difference is heard sighted, why would you *need* a DBT? You wouldn't. If the measured performance of two items (e.g., SET vs ss amp, speakers, etc.) is well above, the threshold of hearing, then the comparison is one of preference, and again a DBT is not needed. You can argue DBT methodology, but the study above differs only by blinding so there's no methodological uncertainties involved. I think we're on the same page if not communicating too well . I don't argue about DBT methodology with the exception of the valid complaint that same bias subjects can corrupt a test. But ALL subjects are biased, that's the purpose of blinding. Obviously, if a subject hears no difference sighted, he's not an appropriate subject for a DBT of that particular comparison. Not in the context in which an average audiophile would value use of a DBT. Ferstler admitted doing exactly this on RAO. He couldn't hear a difference so he just provided random responses. Easiest way to prevent this is to exclude him and people like him who don't hear a difference. Yes. That has been my point all along. *Why* they hear no difference doesn't matter. Keith |
#52
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"ScottW" wrote in message
... On Thursday, October 3, 2013 7:40:54 PM UTC-7, Audio_Empire wrote: In article , "Arny Krueger" wrote: The answer is some combination of the 3 Error Variables stated above. Sure, but again, that's not helpful to the average enthusiast. Frankly the details of the discussion don't need to be helpful to typical enthusiasts. You don't need to know all of the details to make reasonble purchase and operational decisions. Just the basic 80 dB guideline is sufficient to separate the liars from those who tell the truth. It's not helpful to a test engineer either. He lists 3 classes of parameter variation some of which still have nearly infinite possible conditions. Any test engineer knows that the parameters in analog systems can always assume a fully infinite number of values and know that that isn't a stumbling block at all. It 's too broad a brush in an effort to oversimplify. So Scott, when did you accumulate the engineering credentials that qualify you to pass judgment on things like this that you seem to admit you can't understand? Bottom line...it's extremely difficult to establish the thresholds of human hearing for all possible conditions of these parameters. Precise thresholds perhaps, but that is not needed. Again, when high end reviewers blithely claim to hear artifacts that are more than 130 dB down (e.g. certain ESS DACs) you can easily apply even the 110 dB guideline and know what is going on. If it was so easy it would have been done conclusively long ago. It was, it is just that a few people never understood the memo. Denial is more than just a river in Egypt. Also, people who claim superiority for tubed electronics whose artifacts are as little as 20 dB down have their credibility particularly challenged. |
#53
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
In article , KH
wrote: On 10/3/2013 7:40 PM, Audio_Empire wrote: In article , "Arny Krueger" wrote: "Audio_Empire" wrote in message ... In article , KH Well, you start out by demonstrating that there are objectively verifiable differences between units - through measurement - that at least approach the demonstrated lower threshold of human hearing. Agreed. That would help. But you can't measure everything that could affect audio performance. Except you can. It's not 1955 or 1965. Ever since the late 1960s audio measurement technology has evolved rapidly yet most audiophiles are unaware the updates which I will briefly cover below. Ever since the 1980s our understanding of audio perception has evolved rapidly but again most audiophiles are unaware of the updates involved. The latter are probably better understood by young people because among other things they relate to why perceptual coding works as well as it does. But they have broad relevance to our understanding of the significance of technical performance to listening enjoyment and this does not seem to be understood as well. Audio signals have only 2 dimensions per channel. Therefore everything that can go wrong with them can be quantified with a limited number of variables. These 3 Error Variables a (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as harmonic distortion, intermodulation distortion, and frequency or phase distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering signals. Believe it or not, that is it! I believe it. I've seen these parameters outlined before, I didn't just fall off the audio milk train, you know. The point is, they don't mean much to the average audio enthusiasts though. Again it's not so much that these tests exist, as it is that they aren't readily available, easily understandable, or easily correlated to what people hear. When I said "But you can't measure everything that could affect audio performance." I was responding to Keith Hughes and the other audio enthusiasts who post here and who don't have access to either the equipment or the methodologies needed to measure these three major error variable categories (or their sub-categories) and most likely wouldn't understand how these parameters affect what they hear, even if they did. So, in the hobby world, we're back to square one which means our ears. The point, though, is that it would be relatively easy to measure two devices, such as two DACs for which sighted "differences" have been found, show that all of these three types of distortion are, say 120db down, and show that those same listeners cannot distinguish between them under blinded conditions. You think that would be easy to do? It seems to me that special equipment would be needed, and while there are three major categories of distortion, there are many sub-categories within those three that would all have to be measured and correlated before an accurate objective picture of the DACs under test could be constructed. Then you'd have to set up a strict and well-controlled DBT with enough listeners to yield statistically unambiguous results. I'd say that's anything but "relatively easy", unless, of course, you are comparing that effort to jumping off the top of the Empire State building and flying like Superman. Then, you'd be right. Such a test would be "relatively easy" to implement! Now, is this a useful thing for the "average audiophile"? I would argue yes. Not in the sense that one would want to do this type of testing for every device prior to purchase, but to understand that sighted bias can be very powerful, and cannot be "put aside" as some like to believe. I don't think anyone here is advocating that sighted bias be ignored. This is useful to the average Joe when he's looking at two or three components in a shop at X, 3X and 13X price ranges - components that have no technical reason to sound different, and yet do when demo'd in the store. |
#54
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
In article ,
ScottW wrote: On Thursday, October 3, 2013 7:40:54 PM UTC-7, Audio_Empire wrote: In article , "Arny Krueger" wrote: The answer is some combination of the 3 Error Variables stated above. Sure, but again, that's not helpful to the average enthusiast. It's not helpful to a test engineer either. He lists 3 classes of parameter variation some of which still have nearly infinite possible conditions. It's too broad a brush in an effort to oversimplify. Don't you mean in an effort to look superior in this NG? I think that's a strong possibility. Bottom line...it's extremely difficult to establish the thresholds of human hearing for all possible conditions of these parameters. Add to this interaction between these sometimes dependent and sometimes independent variables and then x2 for two channels (and the interaction effect on imaging) of hopefully independent but possibly dependent interaction and then add the masking of music and the problem again becomes impossible. If it was so easy it would have been done conclusively long ago. So very true. But some believe deeply that everything sounds the same, and everything that matters can be measured. That's their mantra, their religion, their raison d'etre. It is not possible that amplifiers, preamps, disc players and DACs can sound different because to admit that would be to admit that subjectivists audio enthusiasts might be right about something. OTOH, when transistor amplifiers first came on the market, It was the subjectivist audio ommunity, led by people like Gordon Holt who screamed that transistor amps sounded terrible. It gave rise to the high-end movement, and eventually the "crazy audiophiles" concerns were vindicated. It was true that while early solid-state gear measured great, it sounded awful ( I remember that the manufacturing community called this the "transistor sound" and tried to tell the buying public that it was good thing). Eventually, the industry caught on and the design conventions that marked early solid-state gear, were abandoned and solid-state got better. Now, of course, it's state-of-the-art, and even so-called "mid-fi" gear can sound excellent. Later, of course, when CD first hit the market, the subjectivists hailed it as "perfect sound forever". Again, the "lunatic fringe" of the audiophile community tried to point out that CDs sounded terrible, especially in the upper registers. "Ear-bleedingly bright" was the term bandied about in the high-end press. "But they're perfect", said the objective camp. How can anyone dispute that? yet, time has shown us that early CDs and many early CD players did sound terrible ( I still have a lot of early discs and they still sound terrible, even on my Sony XA777ES, which is the best CD player I've ever heard - bar none). Eventually, both the production end and the playback end of the CD improved to the point where CD sound can be damned good with a good player and well recorded CD. The funny part of all of this, is that at least one objectivists who regularly posts here has said, in this very forum, that he still owns a Sony CDP-101 (the original Sony player), and that it sounds just like any modern CD player (believe me, that Sony sounded wretched - all the first generation Japanese players did and do) and he also has said that he has a Dynaco Stereo-120 amp, one of the early solid-state amps that sounded so dreadful, and he has stated that he thinks it sounds as good as any modern amp. Does this cast a certain aspersion on this audio enthusiast's ability to hear? For my part, I think so. Now, t's possible that these people really CAN'T hear any difference between components themselves, and they just assume that since they can't, neither can anyone else. But more likely, they simply WON'T hear, because they have so much of themselves wrapped-up in the proposition that if you can hear it, you can measure it, and measurements tell the entire story. I have an engineering degree, and even *I* don't believe that. But, I will say this. If, indeed, these people really CAN'T hear any differences between components , then they would almost HAVE TO rely on measurements for their opinions on audio, now, wouldn't they? |
#55
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On Friday, October 4, 2013 9:16:19 AM UTC-7, Arny Krueger wrote:
"KH" wrote in message ... On 10/4/2013 8:12 AM, Arny Krueger wrote: "Audio_Empire" wrote in message ... In article , "Arny Krueger" wrote: snip I dunno. Kevin gave you the right answer and you seem to be less than fully cognizant of it: KH wrote: "If all measured artifacts or differences are, say -110db down, then a null DBT *is* expected, and a subjectively "verified" audible difference is clearly suspect." Of course he's being wildly optimistic about the hearing threshold of the human ears. Reality is more like 80 dB at the best and 60 dB more typical. Not being optimistic. Just pragmatic. At -110db, even the most rabid naysayer would have to admit his golden ears aren't that good. I feel that way about the -100 dB number, which BTW is reflected in ITU Recommendation BS 1116. It's definately an overkill number just like 44.1 KHz and 16 bits are overkill sampling for digital music. You are joking, are you not Mr. Kruger? "... 44.1 KHz and 16 bits are overkill sampling for digital music." The tongue-in-cheek humor here is obviously the notion that 16/44.1 is overkill, when it's just barely adequate. Audio_Empire |
#56
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"Audio_Empire" wrote in message
... In article , Andrew Haley wrote: Scott wrote: High end audio community doesn't have a say so in submitting to real scientific scrutiny. What does this mean? That high-end audio enthusiats can't do things scientficially? because the priesthood will come and get them? Or for some other reason? That only real scientists can perform experiments? It means that rigorous scientific controls are rarely, if ever applied to the audiophile community and often when such things are tried, it is usually in search of some agenda, and not in search of some truth which can often be inconvenient at best. Interesting double-talk - all applications of science to audio listening tests are agenda driven. Well actually yes, and the agenda is finding the truth. The truth was found and it is alive and well, just not it seems in the high end press. But that's a business and the antipathy of business for science is well known and seen clearly in many contexts. Weight loss, anybody? ;-) No: the whole point of science is that if an experiment is done properly the results will be valid no matter who does the experiment. You don't even have to own a lab coat. All you have to do is not mess it up. In the case of audio, that's more difficult than it might seem on the surface of it. Right and after all these years the high end press still relies on sighted evaluations. Well, its a business! By the way, the man you quote, J Gordon Holt was pretty much the inventor of subjective audio reviewing and never used DBTs in his protocols. he also reported hearing differences between cables and digital playback devices. Go figure.... And he saw the light. Good for him. Now that is one absurd piece of jactitation. You have no idea of the context of that comment, nor do you know what motivated it. To assume that the man saw some "light" by making a comment that agrees with your preconceived notions is most dishonest. You sound like a religious zealot here! I think that in the scientific community we saw JGH's comments as at least being lip service. I agree that he never actually acted on it, but he did see the potential in a positive light. Andrew. |
#57
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On Monday, October 7, 2013 6:59:22 AM UTC-7, Arny Krueger wrote:
"Audio_Empire" wrote in message ... In article , Andrew Haley wrote: Scott wrote: High end audio community doesn't have a say so in submitting to real scientific scrutiny. What does this mean? That high-end audio enthusiats can't do things scientficially? because the priesthood will come and get them? Or for some other reason? That only real scientists can perform experiments? It means that rigorous scientific controls are rarely, if ever applied to the audiophile community and often when such things are tried, it is usually in search of some agenda, and not in search of some truth which can often be inconvenient at best. Interesting double-talk - all applications of science to audio listening tests are agenda driven. Were not talking about science here. We are talking about audiophiles and the debates between "objectivists" and "subjctivists." Well actually yes, and the agenda is finding the truth. No, when we are talking about the great debate in audio between the objectivists and subjectivists the "agenda" is almost always being right regardless of the truth. Both sides are guilty. The truth was found and it is alive and well, just not it seems in the high end press. But that's a business and the antipathy of business for science is well known and seen clearly in many contexts. Weight loss, anybody? ;-) No: the whole point of science is that if an experiment is done properly the results will be valid no matter who does the experiment. You don't even have to own a lab coat. All you have to do is not mess it up. In the case of audio, that's more difficult than it might seem on the surface of it. Right and after all these years the high end press still relies on sighted evaluations. Well, its a business! well yeah, I suppose if one is familiar with the history of the high end press one would understand why it is the way it is. It was an underground movement that filled a void that the then mainstream audio press didn't fill. That void being reports of actual use of audio equipment and subjective evaluations of it. That is something we all do as audiophiles. We form subjective opinions of our gear and of the recordings we play on it and we do it under sighted conditions. The high end audio press was the first to actually report that experience as part of the review process. Those magazines eventually went from underground to being the leading journals because audiophiles could actually relate to them. So yeah, audio publications are a business and it is good business to be relevant to your target audience. By the way, the man you quote, J Gordon Holt was pretty much the inventor of subjective audio reviewing and never used DBTs in his protocols. he also reported hearing differences between cables and digital playback devices. Go figure.... And he saw the light. Good for him. Now that is one absurd piece of jactitation. You have no idea of the context of that comment, nor do you know what motivated it. To assume that the man saw some "light" by making a comment that agrees with your preconceived notions is most dishonest. You sound like a religious zealot here! I think that in the scientific community we saw JGH's comments as at least being lip service. I agree that he never actually acted on it, but he did see the potential in a positive light. I am sorry but I have to ask. You mention "the scientific community" and then you say "we." So you are a scientist? Your PhD is in what branch of science? You have published what research papers in peer reviewed scientific journals? |
#58
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
In article ,
Scott wrote: On Monday, October 7, 2013 6:59:22 AM UTC-7, Arny Krueger wrote: "Audio_Empire" wrote in message ... In article , Andrew Haley wrote: Scott wrote: High end audio community doesn't have a say so in submitting to real scientific scrutiny. What does this mean? That high-end audio enthusiats can't do things scientficially? because the priesthood will come and get them? Or for some other reason? That only real scientists can perform experiments? It means that rigorous scientific controls are rarely, if ever applied to the audiophile community and often when such things are tried, it is usually in search of some agenda, and not in search of some truth which can often be inconvenient at best. Interesting double-talk - all applications of science to audio listening tests are agenda driven. Were not talking about science here. We are talking about audiophiles and the debates between "objectivists" and "subjctivists." Well actually yes, and the agenda is finding the truth. No, when we are talking about the great debate in audio between the objectivists and subjectivists the "agenda" is almost always being right regardless of the truth. Both sides are guilty. The truth was found and it is alive and well, just not it seems in the high end press. But that's a business and the antipathy of business for science is well known and seen clearly in many contexts. Weight loss, anybody? ;-) No: the whole point of science is that if an experiment is done properly the results will be valid no matter who does the experiment. You don't even have to own a lab coat. All you have to do is not mess it up. In the case of audio, that's more difficult than it might seem on the surface of it. Right and after all these years the high end press still relies on sighted evaluations. Well, its a business! well yeah, I suppose if one is familiar with the history of the high end press one would understand why it is the way it is. It was an underground movement that filled a void that the then mainstream audio press didn't fill. That void being reports of actual use of audio equipment and subjective evaluations of it. That is something we all do as audiophiles. We form subjective opinions of our gear and of the recordings we play on it and we do it under sighted conditions. The high end audio press was the first to actually report that experience as part of the review process. Those magazines eventually went from underground to being the leading journals because audiophiles could actually relate to them. So yeah, audio publications are a business and it is good business to be relevant to your target audience. Excellent point. The "high-end" came about as a backlash against early solid-state gear, which sounded terrible. The high-end press started for two reasons: first, was what I call "Julian Hirsch-ism" where tests of equipment always gave the same result. "... The Acme X7000 (amplifier, preamp, tuner, tape deck, you-name-it), like all modern (amplifiers, preamps, tuners, tape decks, you-name-'ems) has no sound of it's own and meets its specifications." This rubbed many audio enthusiasts the wrong way, because, obviously, these components did not all sound the same, and did indeed have a "sound or their own". The second reason was because these same magazines (Stereo Review, High-Fidelity, and to a lesser extent, Audio) failed to note that this new transistor gear sounded terrible. As I said before, the manufacturers hailed this terrible sound as "the transistor sound" and tried to tell the audio enthusiast and music-loving communities that this new "sound" was somehow good. It didn't take most of us but about one experience to realize that both the manufacturers and their "mouthpieces" (those aforementioned magazines) were lying through their teeth! My realization came when I was given an Allied Knight-Kit solid-state amp for Christmas one year. It sounded so terrible, that I was skeptical of solid- state electronics for audio for years. My faith wasn't restored when, after college, when my best friend and I took engineering jobs at a large Aerospace company, and he bought a Dynaco Stereo 120 and PAT-4 preamp. They too sounded lousy (but not as bad as that Knight-Kit, P-U!). I was never impressed with Stereo Review, but I will say this about High-Fidelity. In the early 'sixties, they were one classy magazine. They had this heavy matte paper cover that was really elegant and I learned so much about music from reading their articles and record reviews by such well-known writers in the field as Gene Lees and Nicholas Slomensky, that I found the magazine worthwhile in spite of heir uncritical "public-relations-firm" attitude toward their advertisers' audio products. Audio_Empire |
#59
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"Arny Krueger" wrote:
Audio signals have only 2 dimensions per channel. Therefore everything that can go wrong with them can be quantified with a limited number of variables. These 3 Error Variables a (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as harmonic distortion, intermodulation distortion, and frequency or phase distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering signals. Believe it or not, that is it! How do I evaluate psychoacoustic encoding (data reduction) using these "Error Variables"? Norbert |
#60
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
In article ,
Norbert Hahn wrote: "Arny Krueger" wrote: Audio signals have only 2 dimensions per channel. Therefore everything that can go wrong with them can be quantified with a limited number of variables. These 3 Error Variables a (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as harmonic distortion, intermodulation distortion, and frequency or phase distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering signals. Believe it or not, that is it! How do I evaluate psychoacoustic encoding (data reduction) using these "Error Variables"? Norbert I would say, off the top of my head, that those who believe that all sonic characteristics can be distilled into three error variables, would say that data reduction (psychoacoustic encoding) creates none of those variables, and as such, is totally transparent and therefore inaudible. I've actually seen some of those types actually voice that opinion. However, my experience is that it's not what data reduction TAKES AWAY from the signal that is audible, it's the non-correlated "new" information that this kind of compression ADDS to the signal which would make it a function of #3: "... noise or deterministic and random or pseudo random interfering signals." Audio_Empire |
#61
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"Norbert Hahn" wrote in message
... "Arny Krueger" wrote: Audio signals have only 2 dimensions per channel. Therefore everything that can go wrong with them can be quantified with a limited number of variables. These 3 Error Variables a (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as harmonic distortion, intermodulation distortion, and frequency or phase distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering signals. Believe it or not, that is it! How do I evaluate psychoacoustic encoding (data reduction) using these "Error Variables"? Current best practice involves using double blind listening tests. |
#62
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"Scott" wrote in message
... On Monday, October 7, 2013 6:59:22 AM UTC-7, Arny Krueger wrote: "Audio_Empire" wrote in message ... In article , Andrew Haley wrote: Scott wrote: High end audio community doesn't have a say so in submitting to real scientific scrutiny. What does this mean? That high-end audio enthusiast can't do things scientifically? because the priesthood will come and get them? Or for some other reason? That only real scientists can perform experiments? It means that rigorous scientific controls are rarely, if ever applied to the audiophile community Because they flat out refuse to participate. and often when such things are tried, it is usually in search of some agenda, and not in search of some truth which can often be inconvenient at best. Interesting double-talk - all applications of science to audio listening tests are agenda driven. Were not talking about science here. Please speak for just yourself. We are talking about audiophiles and the debates between "objectivists" and "subjectivists." I have yet to find that place in the universe where we can turn science on and off like a light switch. If you know where it is could you post the lat/longs? ;-) Well actually yes, and the agenda is finding the truth. No, when we are talking about the great debate in audio between the objectivists and subjectivists the "agenda" is almost always being right regardless of the truth. Both sides are guilty. Again, please speak for just yourself. First off, the terms objectivists and subjectivist as used in the high end press are themselves quite agenda driven. Then, if you actually check the commonly accepted definitions of the words, the mystery deepens because the high end audio press's use of them is at variance with their commonly accepted meanings in the real world. The truth was found and it is alive and well, just not it seems in the high end press. But that's a business and the antipathy of business for science is well known and seen clearly in many contexts. Weight loss, anybody? ;-) No: the whole point of science is that if an experiment is done properly the results will be valid no matter who does the experiment. You don't even have to own a lab coat. All you have to do is not mess it up. In the case of audio, that's more difficult than it might seem on the surface of it. Right and after all these years the high end press still relies on sighted evaluations. Well, its a business! well yeah, I suppose if one is familiar with the history of the high end press one would understand why it is the way it is. It was an underground movement that filled a void that the then mainstream audio press didn't fill. Right, the mainstream audio press was strikingly unable and unable to sell the Emperor's new clothes. That void being reports of actual use of audio equipment and subjective evaluations of it. I find no lack of such things in the mainstream press of the day. Remember, I was a charter subscriber of Stereophile - one of the longest-lasting "1 year subscriptions" to a periodical known to man! ;-) That is something we all do as audiophiles. We form subjective opinions of our gear and of the recordings we play on it and we do it under sighted conditions. Again please speak for yourself! The high end audio press was the first to actually report that experience as part of the review process. And those of us who were interested in the true facts of the matter found even larger lapses of accuracy than we audited the process as being used by the high end audio press. Those magazines eventually went from underground to being the leading journals because audiophiles could actually relate to them. The high end audio magazines were frequently kept on the sales counters in high end audio stores for pretty obvious reasons. So yeah, audio publications are a business and it is good business to be relevant to your target audience. Haven't seen audited circulation figures for many of them lately... By the way, the man you quote, J Gordon Holt was pretty much the inventor of subjective audio reviewing and never used DBTs in his protocols. he also reported hearing differences between cables and digital playback devices. Go figure.... And he saw the light. Good for him. Now that is one absurd piece of jactitation. You have no idea of the context of that comment, nor do you know what motivated it. To assume that the man saw some "light" by making a comment that agrees with your preconceived notions is most dishonest. You sound like a religious zealot here! I think that in the scientific community we saw JGH's comments as at least being lip service. I agree that he never actually acted on it, but he did see the potential in a positive light. I am sorry but I have to ask. You mention "the scientific community" and then you say "we." So you are a scientist? Your PhD is in what branch of science? You have published what research papers in peer reviewed scientific journals? |
#63
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"Audio_Empire" wrote in message
... Excellent point. The "high-end" came about as a backlash against early solid-state gear, which sounded terrible. If this Stereophile article isn't a pack of lies http://www.stereophile.com/content/50-years-stereophile the first generally recognized high end audio publication was Stereophile, vol 1 number 1, published Nov 2, 1962 which agrees with my recollections as a charter subscriber. The first commerical SS power amp with any kind of distribution was to the best of my knowlege the Acoustech Model 1 which was reviewed by High Fidelity in August 1963 which roughly corresponded to general retail availabiility. This BTW is the review that led to the termination of the late Julian Hirsch by High Fidelity magazine on the grounds of conflict of interest because he owned Acoustech stock. He shortly resurfaced at Stereo Review. The timing of these two facts reverse the causuality that you claim. The high-end press started for two reasons: first, was what I call "Julian Hirsch-ism" where tests of equipment always gave the same result. "... The Acme X7000 (amplifier, preamp, tuner, tape deck, you-name-it), like all modern (amplifiers, preamps, tuners, tape decks, you-name-'ems) has no sound of it's own and meets its specifications." Actually, this kind of text first showed up in Julian's Stereo Review articles, much later, again invalidating the claimed timing. During Julian's days at High Fidelity audio gear was generally audibly colored. After all, the power amps were all tubed! SS audio gear didn't generally hit the market until 1964 or later. |
#64
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
In article ,
"Arny Krueger" wrote: "Audio_Empire" wrote in message ... Excellent point. The "high-end" came about as a backlash against early solid-state gear, which sounded terrible. If this Stereophile article isn't a pack of lies http://www.stereophile.com/content/50-years-stereophile the first generally recognized high end audio publication was Stereophile, vol 1 number 1, published Nov 2, 1962 which agrees with my recollections as a charter subscriber. The first commerical SS power amp with any kind of distribution was to the best of my knowlege the Acoustech Model 1 which was reviewed by High Fidelity in August 1963 which roughly corresponded to general retail availabiility. This BTW is the review that led to the termination of the late Julian Hirsch by High Fidelity magazine on the grounds of conflict of interest because he owned Acoustech stock. He shortly resurfaced at Stereo Review. The timing of these two facts reverse the causuality that you claim. The high-end press started for two reasons: first, was what I call "Julian Hirsch-ism" where tests of equipment always gave the same result. "... The Acme X7000 (amplifier, preamp, tuner, tape deck, you-name-it), like all modern (amplifiers, preamps, tuners, tape decks, you-name-'ems) has no sound of it's own and meets its specifications." Actually, this kind of text first showed up in Julian's Stereo Review articles, much later, again invalidating the claimed timing. During Julian's days at High Fidelity audio gear was generally audibly colored. After all, the power amps were all tubed! SS audio gear didn't generally hit the market until 1964 or later. There was no timeline inferred, Mr. Kruger and Julian Hirsch was neither the first nor the only practitioner of what I called "Julian Hirschism." But you are right about most equipment being colored in those days, it basically depended on what "color" you, as the consumer/listener preferred. So when Julian and his cohort Gladden Houck insisted that the unit under test (whatever that unit happened to be - speakers excluded), "...like all modern (whatever)) has no sound of it's own..." they were either deaf or lying. Whichever it was, they were wrong as you just pointed out. But it wasn't just tubed equipment that was colored in those days, early solid-state equipment (including the earlier mentioned Acoustech Model 1, which, as well as sounding terrible also had a disconcerting habit of blowing-up!. But there were a number of SS amps that came out about then. All of them were just as colored as any tube electronics - just in a different way. I clearly remember Stewart Hegeman's Harman-Kardon Citation B power amp was also introduced in 1963 as was the Knight line of SS intergrated amps from Allied Radio. All sounded pretty bad - worse than any of the better tube amps from McIntosh and Dyna or the H-K Citation I. --- news://freenews.netfront.net/ - complaints: --- |
#65
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
In article ,
"Arny Krueger" wrote: "Norbert Hahn" wrote in message ... "Arny Krueger" wrote: Audio signals have only 2 dimensions per channel. Therefore everything that can go wrong with them can be quantified with a limited number of variables. These 3 Error Variables a (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as harmonic distortion, intermodulation distortion, and frequency or phase distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering signals. Believe it or not, that is it! How do I evaluate psychoacoustic encoding (data reduction) using these "Error Variables"? Current best practice involves using double blind listening tests. Brings up a good point, Mr. Kruger. You trot-out DBTs as being the proof of your assertions quite often. Where are they? Where, for instance, are the DBTs of data reduction and its artifacts? Where are the DBTs you keep talking about that shows that all DACs sound the same. I've searched the AES papers and I can't find them, I've done what I believe to be a duly diligent Google search for papers describing these DBTs, and their results and found little. I did find one Vanderkooy paper, on DAC sound but it doesn't say that the results were that all DACs sound the same. I would love to see some of these DBT results papers. --- news://freenews.netfront.net/ - complaints: --- |
#66
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"Audio_Empire" wrote in message
... In article , "Arny Krueger" wrote: "Norbert Hahn" wrote in message ... "Arny Krueger" wrote: Audio signals have only 2 dimensions per channel. Therefore everything that can go wrong with them can be quantified with a limited number of variables. These 3 Error Variables a (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as harmonic distortion, intermodulation distortion, and frequency or phase distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering signals. Believe it or not, that is it! How do I evaluate psychoacoustic encoding (data reduction) using these "Error Variables"? Current best practice involves using double blind listening tests. Brings up a good point, Mr. Kruger. You trot-out DBTs as being the proof of your assertions quite often. Where are they? Asked and answered. Where, for instance, are the DBTs of data reduction and its artifacts? Most commonly intermediate work products of coder developers. Where are the DBTs you keep talking about that shows that all DACs sound the same. Asked and answered. I've searched the AES papers and I can't find them, One famous example: https://secure.aes.org/forum/pubs/journal/?ID=2 "Audibility of a CD-Standard A/DA/A Loop Inserted into High-Resolution Audio Playback" Google searching provides this: https://groups.google.com/forum/#!to...nd/6rcu2FXxEU0 Which shows that I previously cited it in a response to you, and you responded to it and quoted its name. Are you having err, memory problems? ;-) |
#67
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
Audio_Empire wrote:
In article , "Arny Krueger" wrote: "Norbert Hahn" wrote in message ... "Arny Krueger" wrote: Audio signals have only 2 dimensions per channel. Therefore everything that can go wrong with them can be quantified with a limited number of variables. These 3 Error Variables a (1) Linear distortion (phase and frequency response errors) (2) Nonlinear distortion (amplitude and time base errors), measured as harmonic distortion, intermodulation distortion, and frequency or phase distortion AKA jitter. (3) and noise or deterministic and random or pseudo random interfering signals. Believe it or not, that is it! How do I evaluate psychoacoustic encoding (data reduction) using these "Error Variables"? Current best practice involves using double blind listening tests. Brings up a good point, Mr. Kruger. You trot-out DBTs as being the proof of your assertions quite often. Where are they? Where, for instance, are the DBTs of data reduction and its artifacts? Subjective Evaluation of State-of-the-Art2-Channel Audio Codecs, Soulodre et al. telos-systems.com/techtalk/00222.pdf Measuring the Characteristics of "Expert" Listeners www.mp3-tech.org/programmer/docs/96-05.pdf Subjective Evaluation of Large and Small Impairments in Audio Codecs Soulodre et al., There are probably more. Where are the DBTs you keep talking about that shows that all DACs sound the same. That's a rusty old canard. Nobody says they all sound the same, just that where they sound different there is an easily measurable explanation. It's not hard to mess it up. Andrew. |
#68
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On Tuesday, October 22, 2013 9:53:56 AM UTC-7, Andrew Haley wrote:
Audio_Empire wrote: SNIP Where are the DBTs you keep talking about that shows that all DACs sound the same. That's a rusty old canard. Nobody says they all sound the same, just that where they sound different there is an easily measurable explanation. It's not hard to mess it up. Andrew. Unfortunately, Andrew. There are people on this very forum who do assert that all DACS sound the same. They reason that this is so because they say that all measurable DAC parameters are at or below the threshold of audibility in modern designs. So yes, there is a school of belief that all modern high-end DACs are sonically transparent. and have said so on this and other forums more than once. In fact, in this very thread, on September 22, Arny Kruger answered my query about the worth of a $40,000 MSB DAC thusly: Me: Sure, good engineering will result in better sound, better reliability, and greater longevity in hi-fi gear as in any other manufactured goods, but is there really anything in an MSB Diamond Platinum DAC IV plus, for instance, to justify its $40,000+ price tag? Mr. Kruger: Or a DAC costing $4,000 or $400, or even more than some costs $40. On the 24th of September, the discussion continued: Me - responding to Mr. Kruger's reply about a $40 DAC sounding just like the more expensive DACs: Well, now you're going too far. Many DACs DO sound different (and some better) than others - even in a bias controlled test. Mr. Kruger Replies: Yes. In this point in life, THE POINT WHERE DAC CHIPS ARE SONICALLY TRANSPARENT LIES ABOUT A DOLLAR A CHANNEL OR LESS. (emphasis mine). For example, I ended up with a motherboard sound facility that produced no output, so an external card was the easiest solution. For less than $30 I obtained an audio interface that per independent tests was the equal of an audio interfact that cost me $399 in 2001. Me: However, my point is that these differences are not necessarily tied to the unit's cost. I.E. a $4000 DAC doesn't, by virtue of its cost, necessarily sound better than a $400 DAC. In my experience, however, DACs utilizing stereo D/A chips generally "sound" better (and by that I mean that there are aspects of their audio performance, such as soundstage or bass presentation that they do do differently than other designs) than do DACs that "time-share" a single D/A converter chip and those utilizing dual-differential D/As can sound better yet, but I've found no hard-and-fast rules there, either. Mr. Kruger: Your distaste for the kind of DBT that professionals use is well known. |
#69
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
Audio_Empire wrote:
On Tuesday, October 22, 2013 9:53:56 AM UTC-7, Andrew Haley wrote: Audio_Empire wrote: SNIP Where are the DBTs you keep talking about that shows that all DACs sound the same. That's a rusty old canard. Nobody says they all sound the same, just that where they sound different there is an easily measurable explanation. It's not hard to mess it up. Unfortunately, Andrew. There are people on this very forum who do assert that all DACS sound the same. They reason that this is so because they say that all measurable DAC parameters are at or below the threshold of audibility in modern designs. So yes, there is a school of belief that all modern high-end DACs are sonically transparent. and have said so on this and other forums more than once. But that's not the same as "all DACs sound the same." THere are, after all, people making tubed DACs. In fact, in this very thread, on September 22, Arny Kruger answered my query about the worth of a $40,000 MSB DAC thusly: Me: Sure, good engineering will result in better sound, better reliability, and greater longevity in hi-fi gear as in any other manufactured goods, but is there really anything in an MSB Diamond Platinum DAC IV plus, for instance, to justify its $40,000+ price tag? Mr. Kruger: Or a DAC costing $4,000 or $400, or even more than some costs $40. On the 24th of September, the discussion continued: Me - responding to Mr. Kruger's reply about a $40 DAC sounding just like the more expensive DACs: Well, now you're going too far. Many DACs DO sound different (and some better) than others - even in a bias controlled test. Mr. Kruger Replies: Yes. In this point in life, THE POINT WHERE DAC CHIPS ARE SONICALLY TRANSPARENT LIES ABOUT A DOLLAR A CHANNEL OR LESS. (emphasis mine). Well, yes. But there's more to a DAC than a DAC chip. As I said, it's easy to mess it up. I've seen some measurements of DACs with very odd-looking results. I think you're misunderstanding Arny's point. He's saying that with ordinary engineering expertise it's possible to make a transparent DAC for not very much money. He's not saying that every DAC at that price point, or any price point, is transparent. Andrew. |
#70
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
|
#71
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On Thursday, October 24, 2013 6:50:56 AM UTC-7, Andrew Haley wrote:
Audio_Empire wrote: On Tuesday, October 22, 2013 9:53:56 AM UTC-7, Andrew Haley wrote: Audio_Empire wrote: SNIP Where are the DBTs you keep talking about that shows that all DACs sound the same. That's a rusty old canard. Nobody says they all sound the same, just that where they sound different there is an easily measurable explanation. It's not hard to mess it up. Unfortunately, Andrew. There are people on this very forum who do assert that all DACS sound the same. They reason that this is so because they say that all measurable DAC parameters are at or below the threshold of audibility in modern designs. So yes, there is a school of belief that all modern high-end DACs are sonically transparent. and have said so on this and other forums more than once. But that's not the same as "all DACs sound the same." THere are, after all, people making tubed DACs. In fact, in this very thread, on September 22, Arny Kruger answered my query about the worth of a $40,000 MSB DAC thusly: Me: Sure, good engineering will result in better sound, better reliability, and greater longevity in hi-fi gear as in any other manufactured goods, but is there really anything in an MSB Diamond Platinum DAC IV plus, for instance, to justify its $40,000+ price tag? Mr. Kruger: Or a DAC costing $4,000 or $400, or even more than some costs $40. On the 24th of September, the discussion continued: Me - responding to Mr. Kruger's reply about a $40 DAC sounding just like the more expensive DACs: Well, now you're going too far. Many DACs DO sound different (and some better) than others - even in a bias controlled test. Mr. Kruger Replies: Yes. In this point in life, THE POINT WHERE DAC CHIPS ARE SONICALLY TRANSPARENT LIES ABOUT A DOLLAR A CHANNEL OR LESS. (emphasis mine). Well, yes. But there's more to a DAC than a DAC chip. As I said, it's easy to mess it up. I've seen some measurements of DACs with very odd-looking results. I think you're misunderstanding Arny's point. He's saying that with ordinary engineering expertise it's possible to make a transparent DAC for not very much money. He's not saying that every DAC at that price point, or any price point, is transparent. Andrew. It certainly seems to me that he's saying that they all sound the same. After all, in answer to my comment: "Well, now you're going too far. Many DACs DO sound different (and some better) than others - even in a bias controlled test." Mr. Kruger Replied: "Yes. In this point in life, THE POINT WHERE DAC CHIPS ARE SONICALLY TRANSPARENT LIES ABOUT A DOLLAR A CHANNEL OR LESS. (emphasis mine)." I would think that had he not meant that all DACs sound the same, he would have clarified his meaning by saying something like "I was just referring to the D/A chip, other parts of the circuitry in a DAC can, of course, affect the sound." But he didn't. This and previous comments by Mr. Kruger over the years lead me to the conclusion that this is his stand on the issue. Sure, I could be wrong, so I guess we need to hear from Mr. Kruger on this. |
#72
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"Andrew Haley" wrote in message
... Mr. Kruger Replies: Yes. In this point in life, THE POINT WHERE DAC CHIPS ARE SONICALLY TRANSPARENT LIES ABOUT A DOLLAR A CHANNEL OR LESS. (emphasis mine). Well, yes. But there's more to a DAC than a DAC chip. Agreed. It happens less and less, but it still happens. OTOH good DACs are showing up in progressively more highly integated chips. For example the AMS AS3525 chip ($10 in small quantities) includes 4 DACs, 2 ADCs and a complete programmable computer system including RAM and power management as well as sophisticated headphone and microphone amplifiers on one chip. This chip is the core of some great-sounding popular sub-$30 portable music players. It seems hard to understand how one can hook this chip up and have it work at all without its audio output being very clean. All you have to do for clean audio output is to run three separate connections from the proper chip terminals to the headphone jack. As I said, it's easy to mess it up. I've seen some measurements of DACs with very odd-looking results. For example, someone recently forwarded me some tests involving an inexpensive USB DAC that clipped several dB below FS. It's audible! I think you're misunderstanding Arny's point. He's saying that with ordinary engineering expertise it's possible to make a transparent DAC for not very much money. He's not saying that every DAC at that price point, or any price point, is transparent. The lesson I take away is that some people seem to only be able to deal with engineering concepts at the hyperbole level. |
#73
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
"Peter Irwin" wrote in message
... Arny Krueger wrote: During Julian's days at High Fidelity audio gear was generally audibly colored. After all, the power amps were all tubed! Ok. Which of the tube power amps reviewed by Julian for High Fidelity do you think you could spot in a blind test? Wikipedia says that Julian's first review for for Hi-Fi/Stereo Review was in October 1961. I don't know how much down time he had after getting fired by High Fidelity Magazine, but it was probably less than a year. Looking at http://www.roger-russell.com/magrevhf.htm prior to 1961, I see a lot of amplifiers that I suspect that could be detected in a blind test, particularly if a more difficult than average speaker load (e.g. AR-3) were involved. I'm pretty sure that MacIntosh amps from the MC75 onward were generally sonically transparent, as were more popular-priced amps such as the Dyna MKIII (with fresh tubes). With older tubes and poorly adjusted bias a Dyna MK III has been detected in an ABX test but fresh tubes and readjusted bias fixed that up. |
#74
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On Friday, October 25, 2013 6:28:46 AM UTC-7, Arny Krueger wrote:
"Andrew Haley" wrote in message ... Mr. Kruger Replies: Yes. In this point in life, THE POINT WHERE DAC CHIPS ARE SONICALLY TRANSPARENT LIES ABOUT A DOLLAR A CHANNEL OR LESS. (emphasis mine). Well, yes. But there's more to a DAC than a DAC chip. Agreed. It happens less and less, but it still happens. OTOH good DACs are showing up in progressively more highly integated chips. For example the AMS AS3525 chip ($10 in small quantities) includes 4 DACs, 2 ADCs and a complete programmable computer system including RAM and power management as well as sophisticated headphone and microphone amplifiers on one chip. This chip is the core of some great-sounding popular sub-$30 portable music players. It seems hard to understand how one can hook this chip up and have it work at all without its audio output being very clean. All you have to do for clean audio output is to run three separate connections from the proper chip terminals to the headphone jack. As I said, it's easy to mess it up. I've seen some measurements of DACs with very odd-looking results. For example, someone recently forwarded me some tests involving an inexpensive USB DAC that clipped several dB below FS. It's audible! I think you're misunderstanding Arny's point. He's saying that with ordinary engineering expertise it's possible to make a transparent DAC for not very much money. He's not saying that every DAC at that price point, or any price point, is transparent. The lesson I take away is that some people seem to only be able to deal with engineering concepts at the hyperbole level. While that's probably true, it doesn't answer the question. Have you or have you not asserted that all DACs sound the same? That's what I have gleaned from the literally dozens of posts on the subject that you have posted here over the last couple of years. If I have gotten the wrong idea, I'd like to know about it. While I find that the audible differences between DAC products are small, even subtle sometimes, they do exist, and they make most DACs unique and discernible in a double-blind listening test. |
#75
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
Audio_Empire wrote:
On Friday, October 25, 2013 6:28:46 AM UTC-7, Arny Krueger wrote: For example, someone recently forwarded me some tests involving an inexpensive USB DAC that clipped several dB below FS. It's audible! I think you're misunderstanding Arny's point. He's saying that with ordinary engineering expertise it's possible to make a transparent DAC for not very much money. He's not saying that every DAC at that price point, or any price point, is transparent. The lesson I take away is that some people seem to only be able to deal with engineering concepts at the hyperbole level. While that's probably true, it doesn't answer the question. Have you or have you not asserted that all DACs sound the same? Which part of "It's audible!" is confusing you? How can anyone possibly be any clearer? Andrew. |
#76
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On Saturday, October 26, 2013 6:57:52 AM UTC-7, Andrew Haley wrote:
Audio_Empire wrote: On Friday, October 25, 2013 6:28:46 AM UTC-7, Arny Krueger wrote: For example, someone recently forwarded me some tests involving an inexpensive USB DAC that clipped several dB below FS. It's audible! I think you're misunderstanding Arny's point. He's saying that with ordinary engineering expertise it's possible to make a transparent DAC for not very much money. He's not saying that every DAC at that price point, or any price point, is transparent. The lesson I take away is that some people seem to only be able to deal with engineering concepts at the hyperbole level. While that's probably true, it doesn't answer the question. Have you or have you not asserted that all DACs sound the same? Which part of "It's audible!" is confusing you? How can anyone possibly be any clearer? Andrew. I don't think that citing an obviously defective design is any concession to anything. Of course it's audible. Comparing two identical interconnects, where one has a defective shield in it is going to make an audible difference too, but that doesn't mean that a group of non-defective interconnects at all price points won't all sound identical. IOW, saying that a defective component is audible is a cop-out and is a non-committal position. |
#77
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
Audio_Empire wrote:
On Saturday, October 26, 2013 6:57:52 AM UTC-7, Andrew Haley wrote: Audio_Empire wrote [to Arny]: While that's probably true, it doesn't answer the question. Have you or have you not asserted that all DACs sound the same? Which part of "It's audible!" is confusing you? How can anyone possibly be any clearer? I don't think that citing an obviously defective design is any concession to anything. Of course it's audible. Comparing two identical interconnects, where one has a defective shield in it is going to make an audible difference too, but that doesn't mean that a group of non-defective interconnects at all price points won't all sound identical. IOW, saying that a defective component is audible is a cop-out and is a non-committal position. Surely, if a transparent DAC can be made for dollars, it makes sense to assert that any DAC which its not transparent is defective. I think I'd be quite happy to suggest that if two DACs can be distinguished by listening then at least one of them must be defective. Of course, I could be proved wrong by a properly- conducted blind test. Andrew. |
#78
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On Sunday, October 27, 2013 6:39:53 AM UTC-7, Andrew Haley wrote:
Audio_Empire wrote: On Saturday, October 26, 2013 6:57:52 AM UTC-7, Andrew Haley wrote: Audio_Empire wrote [to Arny]: While that's probably true, it doesn't answer the question. Have you or have you not asserted that all DACs sound the same? Which part of "It's audible!" is confusing you? How can anyone possibly be any clearer? I don't think that citing an obviously defective design is any concession to anything. Of course it's audible. Comparing two identical interconnects, where one has a defective shield in it is going to make an audible difference too, but that doesn't mean that a group of non-defective interconnects at all price points won't all sound identical. IOW, saying that a defective component is audible is a cop-out and is a non-committal position. Surely, if a transparent DAC can be made for dollars, it makes sense to assert that any DAC which its not transparent is defective. I think I'd be quite happy to suggest that if two DACs can be distinguished by listening then at least one of them must be defective. Of course, I could be proved wrong by a properly- conducted blind test. Andrew. The problem here seems to be with the term "transparent". Like I said in another post, I haven't encountered a bad DAC in years, but is that the definition of transparent? For instance, I received a DAC from a company called Gefen last week: http://www.gefen.com/kvm/gtv-192kusb...?prod_id=11499 It is similar in concept to the AudioQuest DragonFly in that it is a USB DAC. It has enhanced functionality over the DragonFly because it supports 24/192KHz where the AudioQuest only decodes 24/96 KHz. It also has an Optical output so that it can be used as a USB/SPDIF converter for use with a high-end DAC that lacks a USB input. While it sounds fine, it does not have the bottom-end of the AudioQuest (which I also have) nor does it have as good a soundstage presentation. But neither of these DACs has as good a top end or as good imaging as the Ayre QB9. Does that make them not transparent? I don't know. None of the three sounds distorted or noisy and Using a test CD and an HP 410b audio voltmeter, all three measure within ~1 dB from 10 Hz (the lowest frequency on the CD) out to 20 KHz. Yet all three sound different and the Ayre sounds best. I could live with any of them, but that certainly doesn't mean that I don't note any differences. It's like the differences between several very good speakers. Some do some things better and others do other things better, But between, say, a pair of Wilson Alexandria XLF. Magico Q7s, or Martin Logan CLS (with suitable sub-woofers), as different as they are from one another (far more different than than the differences between these three DACs), I could happily live with any of them. They are all excellent loudspeakers, I wouldn't go so far as to say that any of them was "defective", yet all produce very different musical presentations. See my point? And it may be that all current DAC chips are themselves so close to perfect as to measure virtually identically. But there is more to a commercially available DAC box than merely the D/A chip inside of it. There are power supplies, filters, analog amplifiers, etc. There are also designs that use one chip switching between right and left channels on a "time-share" basis for stereo, while other use two separate chips, one for each channel. Some designs utilizes many as four D/A chips in a stereo dual differential configuration. Surely these different design approaches are going to result in slightly different presentations. On the very top-end of the market, many DAC designs utilize proprietary D/A circuits built-up using discrete components, eschewing the use of I.C.-based D/A chips altogether. Two of thee that come to mind are the products of the British company dCS with their proprietary "ring DAC" technology and MSB with their $40,000 DAC IV "Diamond" converter using set of resin potted discrete "Ladder DAC" modules. Are we supposed to accept that these DACs with their luxury automobile price tags sound no different than a similar product costing, in some cases, orders of magnitude less than these products? I'll grant that most likely, there is nothing in these mega-buck designs that warrant, performance-wise, anyway, their price tags, but I'll guarantee that they don't sound the same as other designs, yet I doubt that anyone would say that they are less than transparent - according to the accepted definition of the term. The problem is that while I'm sure that audio groups, both formal and informal, the world over have conducted DBT tests to determine the audibility of digital to analog conversion, there seem to be no definitive studies that once can cite. Lipschitz and Vanderkooy haven't undertaken such a study, and neither have Meyer and Moran. A search of the JAES website will bring up no study that anyone has been able to cite of DAC sound except the oft- mentioned Meyer and Moran study entitled "Audibility of a CD-Standard A/DA/A Loop Inserted into High-Resolution Audio Playback" which is, in spite of some posters' protestations to the contrary, NOT a study about the audibility of different types of DACs and exhibits a totally different agenda. |
#79
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
Audio_Empire wrote:
On Sunday, October 27, 2013 6:39:53 AM UTC-7, Andrew Haley wrote: Surely, if a transparent DAC can be made for dollars, it makes sense to assert that any DAC which its not transparent is defective. I think I'd be quite happy to suggest that if two DACs can be distinguished by listening then at least one of them must be defective. Of course, I could be proved wrong by a properly- conducted blind test. Andrew. The problem here seems to be with the term "transparent". I'll define what I mean: any ADC/DAC that cannot be distinguished from a piece of wire in a listening test is transparent. [ ... ] And it may be that all current DAC chips are themselves so close to perfect as to measure virtually identically. But there is more to a commercially available DAC box than merely the D/A chip inside of it. There are power supplies, filters, analog amplifiers, etc. There are also designs that use one chip switching between right and left channels on a "time-share" basis for stereo, while other use two separate chips, one for each channel. Some designs utilizes many as four D/A chips in a stereo dual differential configuration. Surely these different design approaches are going to result in slightly different presentations. Yes, but techniques exist in the world of instrumentation to reduce to reduce artefacts way below thresholds of hearing. Audio is easy when compared with, say, measuring nanoamps in a noisy electrical environment, as in an atomic force microscope. On the very top-end of the market, many DAC designs utilize proprietary D/A circuits built-up using discrete components, eschewing the use of I.C.-based D/A chips altogether. Two of thee that come to mind are the products of the British company dCS with their proprietary "ring DAC" technology and MSB with their $40,000 DAC IV "Diamond" converter using set of resin potted discrete "Ladder DAC" modules. Are we supposed to accept that these DACs with their luxury automobile price tags sound no different than a similar product costing, in some cases, orders of magnitude less than these products? Maybe, maybe not. The only way to try is to listen. I will note, however, that discrete construction is a spectacularly bad way of making a hybrid component like a DAC: it's much harder to get the accurate matching of components that you need. The problem is that while I'm sure that audio groups, both formal and informal, the world over have conducted DBT tests to determine the audibility of digital to analog conversion, there seem to be no definitive studies that once can cite. Why should there be? It's not as if the distortion caused by converters is special. There's no reason to believe that DAC distortion will b audible at thresholds below distortion caused by any other component. Andrew. |
#80
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On Monday, October 28, 2013 5:27:08 AM UTC-7, Andrew Haley wrote:
Audio_Empire wrote: =20 On Sunday, October 27, 2013 6:39:53 AM UTC-7, Andrew Haley wrote: =20 =20 Surely, if a transparent DAC can be made for dollars, it makes sense =20 to assert that any DAC which its not transparent is defective. I =20 think I'd be quite happy to suggest that if two DACs can be =20 distinguished by listening then at least one of them must be =20 defective. Of course, I could be proved wrong by a properly- =20 conducted blind test. =20 Andrew. =20 =20 =20 The problem here seems to be with the term "transparent". =20 I'll define what I mean: any ADC/DAC that cannot be distinguished from a piece of wire in a listening test is transparent. Yes, that is the accepted definition. However, I can tell you, we aren't "t= here" yet. While that would be an impossible test to actually perform on a D/A converter, the very fact that all D/A converters sound slightly different f= rom=20 each other tells us that we aren't at that point (unless we can agree that= =20 different pieces of wire "sound' different from each other - something that I dare say that none of us is prepared to assert. =20 =20 And it may be that all current DAC chips are themselves so close to perfect as to measure virtually identically. But there is more to a commercially available DAC box than merely the D/A chip inside of it. There are power supplies, filters, analog amplifiers, etc. There are also designs that use one chip switching between right and left channels on a "time-share" basis for stereo, while other use two separate chips, one for each channel. Some designs utilizes many as four D/A chips in a stereo dual differential configuration. Surely these different design approaches are going to result in slightly different presentations. =20 Yes, but techniques exist in the world of instrumentation to reduce to reduce artefacts way below thresholds of hearing. Audio is easy when=20 compared with, say, measuring nanoamps in a noisy electrical environment, as in an atomic force microscope. Good point. And while many modern electrical components do have specs that put artifacts below the threshold of hearing (the LM49710 and equivale= nt family of op-amps come to mind here), not all devices made from such compon= ents have ALL of their artifacts banished to the realm of inaudibility -yet. I'm= also not convinced that ALL DAC designers are striving for complete transparency= , but=20 are striving for designs that they think "sound good".=20 =20 On the very top-end of the market, many DAC designs utilize=20 proprietary D/A circuits built-up using discrete components,=20 eschewing the use of I.C.-based D/A chips altogether. Two of thee that come to mind are the products of the British company dCS with their proprietary "ring DAC" technology and MSB with their $40,000 DAC IV "Diamond" converter using set of resin potted discrete "Ladder DAC" modules. Are we supposed to accept that these DACs with their luxury automobile price tags sound no different than a similar=20 product costing, in some cases, orders of magnitude less than these products? Maybe, maybe not. The only way to try is to listen. I will note, however, that discrete construction is a spectacularly bad way of making a hybrid component like a DAC: it's much harder to get the accurate matching of components that you need. As a semiconductor engineer, I will say that theoretically, at least, that is quite true. It is much easier to design and fabricate a=20 consistent monolithic circuit, where all the components of that circuit=20 share the same substrate than it is to use disparate discrete components.= =20 The integrated circuit will be more consistent from unit to unit and easier to control to a design specification. In fact, knowing what I know about circuit design, I would venture to say, that a real high-end designer would have better luck hand choosing IC parts than hand building discrete ones. That is to say. It would be easier to test a bunch of DAC chips and choosin= g to use only those few out of any wafer start, which, for some reason, test slightly better than the rest than to design such a circuit using separate= =20 transistors resistors and capacitors.. This is called "binning" and it is d= one=20 all the time for things like DSP chips, memory chips and microprocessors.= =20 As good as modern IC processing is, there is always going to be some yield= =20 variation. Many semiconductor companies bin their parts themselves, premium= =20 pricing those that test to a higher set of specifications than do the avera= ge part.=20 I think that companies who build their own DACs from scratch are playing on a naivet=E9 on the part of the audiophile who might remember when ICs for= =20 audio simply weren't very good for the application (709 op amps, anyone?).= =20 I certainly remember when a decent microphone mixer used resin-potted,=20 discrete component op-amp modules to get the low noise demanded by=20 recording studios simply because Integrated Circuit op-amps couldn't cut=20 the mustard. Then again, maybe companies like MSB and dCS make their own=20 DACs these days simply to justify their astronomical prices. Whatever the r= eason,=20 I suspect that hand selecting from a Burr-Brown or ESS SabreDAC (etc.) IC= =20 product would give at least just as good of a performance as hand-tweaked,= =20 hand-selected discrete component D/A converter modules. =20 The problem is that while I'm sure that audio groups, both formal and informal, the world over have conducted DBT tests to determine the audibility of digital to analog conversion, there seem to be no definitive studies that once can cite. Why should there be? It's not as if the distortion caused by=20 converters is special. There's no reason to believe that DAC distortion will b audible at thresholds below distortion caused=20 by any other component. I don't think that's the point. The point is that there are two equally pas= sionate=20 schools of belief with regard to DACs. Once school says that all DACs are s= o different that one needs to choose one based on how they sound in one's=20 system. The other school of thought says that they all sound alike because any criterion which would contribute to a DAC having a sound of its own is,= =20 today, in modern designs, well below the threshold of audibility. This seco= nd school maintains that there is no reason to spend more than a few dollars (= I believe $40 was mentioned) on any DAC for that reason. One of those two cam= ps=20 is wrong. Some definitive DBT results would do much to put this debate to b= ed.=20 Yet audio researchers such as Lipschitz and Vanderkooy, or Meyer and Moran = have=20 not undertaken to address this question. Indeed, a search of AES papers yie= lds no results that address this question in any direct way. That means while p= eople in the strictly objective camp can CLAIM (and some have done on this forum)= that DACs must sound the same because all artifacts have been rendered below aud= ibility,=20 but they can produce no more proof that this is the case than can the most = dyed-in- the-wool subjectivist that All DACs sound different and that these differen= ces aren't=20 subtle. I mean this isn't like the wire and cable sound "debate" where DBTs= have=20 definitively laid to rest the question of whether interconnects and speaker= cable=20 can have a "sound".=20 |
Reply |
Thread Tools | |
Display Modes | |
|
|