Kuei Yang Wang, I think you are just as much a "selfrightuous preacher" as I am. You are convinced there are audible differences. I strongly think there are (in almoste every case) no audible differences. Some people that participate in this thread seem to support your point of view, others seem more supportive to mine.
The question, to DB or not to DB probably leads us to a dead end.
I am still wondering, how many people on diyaudio.com belong to the so-called objectivists, and how many count themselves to subjectivists.
Considering the difference between objectivists and subjectivists, it is actually quite strange subjectivists don't mind to be called subjectivists. The name already says that for example the differences they think to hear between amps are based on their personal opinion, rather than reason or fact.
We tested 8 times (I know, not quite enough). I would have liked to do more tests, but my friend got fed up he wasn't really sure each time. He got 5 out of 8 right. Even if this score represented a true, small audible difference...
Who would pay thousands of dollars more for this small difference, if spending half of that amount of money extra on the speakers, could buy you a speaker that betters the ones you have now on every aspect?
Even if there is some audible difference between equipment, I still stand behind the point I tried te make in my first post:
Most important are the quality of the recorded material, the speakers, and the way they interact with the room.
The question, to DB or not to DB probably leads us to a dead end.
I am still wondering, how many people on diyaudio.com belong to the so-called objectivists, and how many count themselves to subjectivists.
Considering the difference between objectivists and subjectivists, it is actually quite strange subjectivists don't mind to be called subjectivists. The name already says that for example the differences they think to hear between amps are based on their personal opinion, rather than reason or fact.
This is true IMPOV. A few weeks ago, I tested the "quality" of my high-end friends' golden ear. The difference between his big buck, many great reviews, highly regarded valve amp and my el cheapo $200 PA-amp should be audible, right?
We tested 8 times (I know, not quite enough). I would have liked to do more tests, but my friend got fed up he wasn't really sure each time. He got 5 out of 8 right. Even if this score represented a true, small audible difference...
Who would pay thousands of dollars more for this small difference, if spending half of that amount of money extra on the speakers, could buy you a speaker that betters the ones you have now on every aspect?
Even if there is some audible difference between equipment, I still stand behind the point I tried te make in my first post:
Most important are the quality of the recorded material, the speakers, and the way they interact with the room.
Hi all
It's very convenient to put everyone into boxes.
-objectivists and subjectivists?-
Most people are probably a mix, don't you think?
Do you see the world in black and white?
Of course those things are very important.
Other things are very important , too.
And the more of your equipment you build youself,
the more you start to figure out what is important.
- TO YOU -
Happy building
cheers
It's very convenient to put everyone into boxes.
-objectivists and subjectivists?-
Most people are probably a mix, don't you think?
Do you see the world in black and white?
keyser said:
Of course those things are very important.
Other things are very important , too.
And the more of your equipment you build youself,
the more you start to figure out what is important.
- TO YOU -
Happy building
cheers
Konnichiwa,
Actually, I know there are audible differences. But that is actually besides the point.
I mainly object to the way ABX/DB testing is applied, implemented and analysed by it's proponents within the Audio Objectivist camp. That objection is one strictly based on the fact that their tests are implemented in an ill advised manner at least and in many cases actively and deliberatly (if that is the result of a concious choice or just the subconcious be left open to question) fraudulent.
Now you cite such tests as your sole reson why what you claim should be given consideration. Based on that I must suggest that your position should be ignored.
In the whole discussion so far (and you may remember the same discussion taking place years back between us on a different board) you have not changed or adjuste dyour arguments the least.
You can think strongly that the moon is made of green cheese. That means DIDLY SQUAT.
The Facts are:
1) There are strong indications (call them Audio Folklore) that audible differences appear to exist where according to simplistic electronic theory they should not. But there are easily predictable MEASUREABLE differences in full accordance with sensible levels of electronic theory (eg where the simplification has not been taken so far as to yield unrealistic and meaningless results). Of course, electronic theory can only predict the electronic behaviour and not if the resultant changes to the signal are audible and to what degree. But as any Folklorist will tell you - if there is smoke there is usually a fire smoldering somewhere.
2) The majority of the popularly quoted ABX & DB Tests in the context of High End Audio have been repeatedly and quite authoratively criticised on grounds of statistical methode and various psychological factors. These criticisms where made availabel to said group of vocal experimenteors, have never been illustrated to have been non-factual and must therefore be considered as valid. The usual group I like to call ABX Mafia has refused, in the light of said criticim of their methodes to alter them. They therefore must accept the charge of deliberatly caring out pseudo-scientific tests which are prejudiced and will with good reliability return the results desired by this group, disregardless of reality. Therefore said tests should be discarded as deliberately distorted on ideological grounds by anyone who approaches things from a scientific and not from a religious viewpoint.
Quite frankly, the sum of 1 & 2 leaves you with a convenient belief that happens to have no rational foundation. Like so many billions of people you are welcome to hold such irrational believes, take comfort from them and anything you like, but if you attemt to pass them off as fact anywhere I'm around we'll go around the same roundabout again. You are welcome to try.
If this where an election for the presidency of the United States, this may of some relevance. But I am not trying win opinion polls or votes. I enquire into the actual nature of things and prefer to know exactly to believing on questionable or even counterfactual and irrational grounds.
Not neccesarily.
As said, DB tests like so many other types of test have clear requirements in order to give usefull results. Failure to comply with such requirements means you have to accept that the given test (yours or anyone elses) is worthless and a waste of time, plus that it contributed nothing to the sum knowledge on the subject. That's all.
I do not particulary care as of course both objectivists and subjectivists are equally and completely wrong. Both are irrational, pseudo religious systems who give their adherents something reality does not support, namely certainty that things are as they believe. Any such irrational will sooner or later lead to a run-in with reality.
Any form of strongly held opinion or belief is counterproductive when researching a subject, as the belief will usually obscure facts that would invalidate it.
Pray tell, what else EXCEPT my own experience, opinion, experinece and knowledge should I be concerend with, for my own use?
You use DB testing very much in a way that INVARIABLY makes them completely invalid. Here is why.
1) The subject in the test is aware of what is being tested and has strong opinions on what the test should reveal. As a result his reactions will be determined more by expectation than by what is being heard.
2) The number of datapoints in the set is way too small to allow any semi relaible conclusion. In fact, I have not even at hand any readily done calculations to illustrate how likely it is that a modest audible difference is being overlooked.
If we use a 16 Trial test and we consider that 12 correct identiciations constitute a significant (significance .05 or higher) indication of an audible difference, the we would, depending upon the degree of actual, audible difference have a likelyhood of between 55 - 83% that this result was not actually due to no difference being heard but due to a Type 2 / B Statistical Error. With a 8 Trial test this likelyhood would even larger.
If we where to even out the risks of Type 1 and Type 2 Errors in our analysis we would note that the minimum number of datapoints in a given set must be well above 100 in order to allow an equally low chance of Type 1/A and Type 2/B errors and that ONLY AND STRICTLY if the provision of objection 1 can be overcome by suitable design of the test.
So, basically you completely your own time and that of your aquaintance and nothing in your result is of any consequence to anything. You may consider a ritual believers go through in order to strengther their beliefs, regardless of reality.
I agree that a large part of the result depends on using competently engineered speakers, in this context competent implies the following requirements:
1) Low distortion
2) Low compression
3) Well controlled directivity (if variable then with narrowing dispersion at high frequencies and no abrupt directivity changes.
4) Accurate reproduction of Impulses
5) Even frequency balance in room, under normal reverbrant field condition (anechonic on-axis Frequency response is irelevant)
Practially ALL HiFi Speakers, be they "High End" and expensive or "Consumer Fi" and chap are subject to design principles and approaches which severely prejudice performance on pretty much ALL of the above accounts. In fact, in most cases the above is only fulfilled in large format studio moitors (NOT in so-called nbearfiled monitors) and even then only in a minority of them.
Surprising the speakers preferred by the most "Ultra-Subjectivist" group in High End audio (the Ultra-Fidelity crowd) tend to be quite decent on all the above counts. Could it it be that they hear more differences simply because they have the better speakers?
I dount I will have given any more reason for thought or to re-evaluate your erroneous position, but I felt it needed to be said in order to offer a reasnoable balance to the views you promulgate.
Sayonara
keyser said:
Actually, I know there are audible differences. But that is actually besides the point.
I mainly object to the way ABX/DB testing is applied, implemented and analysed by it's proponents within the Audio Objectivist camp. That objection is one strictly based on the fact that their tests are implemented in an ill advised manner at least and in many cases actively and deliberatly (if that is the result of a concious choice or just the subconcious be left open to question) fraudulent.
Now you cite such tests as your sole reson why what you claim should be given consideration. Based on that I must suggest that your position should be ignored.
In the whole discussion so far (and you may remember the same discussion taking place years back between us on a different board) you have not changed or adjuste dyour arguments the least.
keyser said:
You can think strongly that the moon is made of green cheese. That means DIDLY SQUAT.
The Facts are:
1) There are strong indications (call them Audio Folklore) that audible differences appear to exist where according to simplistic electronic theory they should not. But there are easily predictable MEASUREABLE differences in full accordance with sensible levels of electronic theory (eg where the simplification has not been taken so far as to yield unrealistic and meaningless results). Of course, electronic theory can only predict the electronic behaviour and not if the resultant changes to the signal are audible and to what degree. But as any Folklorist will tell you - if there is smoke there is usually a fire smoldering somewhere.
2) The majority of the popularly quoted ABX & DB Tests in the context of High End Audio have been repeatedly and quite authoratively criticised on grounds of statistical methode and various psychological factors. These criticisms where made availabel to said group of vocal experimenteors, have never been illustrated to have been non-factual and must therefore be considered as valid. The usual group I like to call ABX Mafia has refused, in the light of said criticim of their methodes to alter them. They therefore must accept the charge of deliberatly caring out pseudo-scientific tests which are prejudiced and will with good reliability return the results desired by this group, disregardless of reality. Therefore said tests should be discarded as deliberately distorted on ideological grounds by anyone who approaches things from a scientific and not from a religious viewpoint.
Quite frankly, the sum of 1 & 2 leaves you with a convenient belief that happens to have no rational foundation. Like so many billions of people you are welcome to hold such irrational believes, take comfort from them and anything you like, but if you attemt to pass them off as fact anywhere I'm around we'll go around the same roundabout again. You are welcome to try.
keyser said:
If this where an election for the presidency of the United States, this may of some relevance. But I am not trying win opinion polls or votes. I enquire into the actual nature of things and prefer to know exactly to believing on questionable or even counterfactual and irrational grounds.
keyser said:
Not neccesarily.
As said, DB tests like so many other types of test have clear requirements in order to give usefull results. Failure to comply with such requirements means you have to accept that the given test (yours or anyone elses) is worthless and a waste of time, plus that it contributed nothing to the sum knowledge on the subject. That's all.
keyser said:
I do not particulary care as of course both objectivists and subjectivists are equally and completely wrong. Both are irrational, pseudo religious systems who give their adherents something reality does not support, namely certainty that things are as they believe. Any such irrational will sooner or later lead to a run-in with reality.
Any form of strongly held opinion or belief is counterproductive when researching a subject, as the belief will usually obscure facts that would invalidate it.
keyser said:
Pray tell, what else EXCEPT my own experience, opinion, experinece and knowledge should I be concerend with, for my own use?
keyser said:
You use DB testing very much in a way that INVARIABLY makes them completely invalid. Here is why.
1) The subject in the test is aware of what is being tested and has strong opinions on what the test should reveal. As a result his reactions will be determined more by expectation than by what is being heard.
2) The number of datapoints in the set is way too small to allow any semi relaible conclusion. In fact, I have not even at hand any readily done calculations to illustrate how likely it is that a modest audible difference is being overlooked.
If we use a 16 Trial test and we consider that 12 correct identiciations constitute a significant (significance .05 or higher) indication of an audible difference, the we would, depending upon the degree of actual, audible difference have a likelyhood of between 55 - 83% that this result was not actually due to no difference being heard but due to a Type 2 / B Statistical Error. With a 8 Trial test this likelyhood would even larger.
If we where to even out the risks of Type 1 and Type 2 Errors in our analysis we would note that the minimum number of datapoints in a given set must be well above 100 in order to allow an equally low chance of Type 1/A and Type 2/B errors and that ONLY AND STRICTLY if the provision of objection 1 can be overcome by suitable design of the test.
So, basically you completely your own time and that of your aquaintance and nothing in your result is of any consequence to anything. You may consider a ritual believers go through in order to strengther their beliefs, regardless of reality.
keyser said:
I agree that a large part of the result depends on using competently engineered speakers, in this context competent implies the following requirements:
1) Low distortion
2) Low compression
3) Well controlled directivity (if variable then with narrowing dispersion at high frequencies and no abrupt directivity changes.
4) Accurate reproduction of Impulses
5) Even frequency balance in room, under normal reverbrant field condition (anechonic on-axis Frequency response is irelevant)
Practially ALL HiFi Speakers, be they "High End" and expensive or "Consumer Fi" and chap are subject to design principles and approaches which severely prejudice performance on pretty much ALL of the above accounts. In fact, in most cases the above is only fulfilled in large format studio moitors (NOT in so-called nbearfiled monitors) and even then only in a minority of them.
Surprising the speakers preferred by the most "Ultra-Subjectivist" group in High End audio (the Ultra-Fidelity crowd) tend to be quite decent on all the above counts. Could it it be that they hear more differences simply because they have the better speakers?
I dount I will have given any more reason for thought or to re-evaluate your erroneous position, but I felt it needed to be said in order to offer a reasnoable balance to the views you promulgate.
Sayonara
What could possibly be wrong with the ABX testing? I'd like to point out that the only ones that implement ABX or DB testing in such an "ill-advised...diliberately fraudulent" manner are the audio subjectivists. Basically the only scientific method used by the subjectivist community is cost.
Just read an article in Stereophile. The reviewers have pretty much made up their mind when they see the price tag of the product. They take so long to get to the point. An article in Sterophile reminds me of a shrewd used car salesman. Every review starts out the same way, telling you how carefully it was packaged. Higher cost products just have more detailed reviews that's all.
Sony amplifiers are quite state of the art I'll have you know. You're bashing them because they don't cost enough. A Sony ES reciever with class D amplifier will hold its own quite comfortably in the hi-fi-lie arena if it were ever given a fair chance. It's quite a fine product in the hi-fi-reality arena.
I have yet to hear an amplifier with 1% distortion that I've liked. Distortion is exactly what the word means, it's an undesired
change in the waveform of the signal. It may make the sound more pleasant to the ear, but it's not a true representation of the source. The goal of hi-fi is to reproduce to original material as faithfully as possible, and 1% distortion is not very faithful by any standards in than last half a century. Clearly you don't play a musical instrument or have ever been to an orchestra if you think 1% distortion sounds good.
Konnichiwa,
As with any other test:
1) The implementation
2) The interpretation
An ABX test is in effect a specific to measure, or to quantify a given effect. ABX/DB Testing uses a specific methode, namely the statistical.
Statistics are as prone to misinterpretation than any other measurement. I'll not bore you with examples either way, if you took statistics modules at university you will be well aware of the issues.
That depends on the definition of "Art". They are competent implementations of principles desrived from common prejudices and usually in direct opposition to how and what humans hear.
I am not bashing anything. I merely used an example and Sony is a nice, fat target. In a software discussion I would have used invariably Microsoft for my example.
Could you please be clear as to what amplifiers you have heard that 1% Distortion and under what conditions and how you verified that distortion you heard was EXACTLY 1% when you heard it?
Please note that I very specifically qualified the measurements stated.
Hmmm. I really hope you do not use electrodynamic speakers or headphone for listening. Because they have hugely more distortion at listening levels than even "high distortion by design" amplifiers, such SE Triodes.
Moreover, what if you have an amplifier that distorts, but produces distortion such that it is FULLY masked by the ears mechanisms and the brains DSP system? In other words the distortion is readily measurable, but it is reliably inaudible.
Is it still an "undesirable" change in signal waveform? And what if, in order to reduce the large amount of inaudible distortion you would make the measured THD much lower but by re-distributing the distortion energy across a larger number of higher harmonics the endresult would now actually produce distortion that is MORE audible?
What do you object to? Measured distortion or audible distortion? Me, i object to the latter.
Funny, the vast majority of speakers will have much higher levels of distortion, but these you ignore conveniently.
Actually, I don't play any instrument right now, but have played several over time. I have also spend a lot of time as sound engineer both live and recording. And I regulary go to classical concerts at the Barbican and RFH.
And I do not contend that "1% Distortion" sounds good or not, I contend that saying "1% THD" is compeletly meanngless and of no consequence or value, unless much more about the context is stated, such as the SPL present (with rising loudness of the signal the human auditory system will mask higher levels of distortion), the frequency and of course the specific spectrum of the distortion.
It is easy to demonstrate for example that 0.1% THD can be much more objectionable than 9.6% THD. If you don't believe it, take the Geddes Test here:
http://gedlee.com/distortion_perception.htm
I leave it here, once you actually have studied what you have been talking about and have made sure that you are aware of at least the absolute basics of statistics, physio & psychoacoustics as well as the specifics of distortion I will be happy to continue discussing Stistics, ABX/DB Testing and distortion with you.
Sayonara
hummhoom said:
As with any other test:
1) The implementation
2) The interpretation
An ABX test is in effect a specific to measure, or to quantify a given effect. ABX/DB Testing uses a specific methode, namely the statistical.
Statistics are as prone to misinterpretation than any other measurement. I'll not bore you with examples either way, if you took statistics modules at university you will be well aware of the issues.
hummhoom said:
That depends on the definition of "Art". They are competent implementations of principles desrived from common prejudices and usually in direct opposition to how and what humans hear.
hummhoom said:
I am not bashing anything. I merely used an example and Sony is a nice, fat target. In a software discussion I would have used invariably Microsoft for my example.
hummhoom said:
Could you please be clear as to what amplifiers you have heard that 1% Distortion and under what conditions and how you verified that distortion you heard was EXACTLY 1% when you heard it?
Please note that I very specifically qualified the measurements stated.
hummhoom said:
Hmmm. I really hope you do not use electrodynamic speakers or headphone for listening. Because they have hugely more distortion at listening levels than even "high distortion by design" amplifiers, such SE Triodes.
Moreover, what if you have an amplifier that distorts, but produces distortion such that it is FULLY masked by the ears mechanisms and the brains DSP system? In other words the distortion is readily measurable, but it is reliably inaudible.
Is it still an "undesirable" change in signal waveform? And what if, in order to reduce the large amount of inaudible distortion you would make the measured THD much lower but by re-distributing the distortion energy across a larger number of higher harmonics the endresult would now actually produce distortion that is MORE audible?
What do you object to? Measured distortion or audible distortion? Me, i object to the latter.
hummhoom said:
Funny, the vast majority of speakers will have much higher levels of distortion, but these you ignore conveniently.
hummhoom said:
Actually, I don't play any instrument right now, but have played several over time. I have also spend a lot of time as sound engineer both live and recording. And I regulary go to classical concerts at the Barbican and RFH.
And I do not contend that "1% Distortion" sounds good or not, I contend that saying "1% THD" is compeletly meanngless and of no consequence or value, unless much more about the context is stated, such as the SPL present (with rising loudness of the signal the human auditory system will mask higher levels of distortion), the frequency and of course the specific spectrum of the distortion.
It is easy to demonstrate for example that 0.1% THD can be much more objectionable than 9.6% THD. If you don't believe it, take the Geddes Test here:
http://gedlee.com/distortion_perception.htm
I leave it here, once you actually have studied what you have been talking about and have made sure that you are aware of at least the absolute basics of statistics, physio & psychoacoustics as well as the specifics of distortion I will be happy to continue discussing Stistics, ABX/DB Testing and distortion with you.
Sayonara
I hesitate to throw my $ 0.02 in this discourse but how can I resist so much fun
All sides have good points but don't refer to the same contexts, and the subject has so many complexities that no one can ever give clear answers. In any case most problems come from the generalization of a few peronal experiences to all situations everywhere and for everyone.
If no differences between electronic devices existed, why would so many serious, earnest designers, many well respected, conduct earnest discussions about amp topologies etc here on these forums? Obviously, *some* people on *some* kinds of equipment hear (and measure!) *some* differences *sometimes*.
I won't even go into the audibility problem - what kind of distortion appears more audible or more objectionable. Again, a vast field of study.
I'll just enlarge the point on statistics.
The statistical problem really exists: sample sizes of less than 15 to 20 really can't give you a reliable statistic. Any statistic can only test for rejection of the null hypothesis. The null hypothesis states that no differences exist. If we can reject this hypothesis with a good significance level, we "prove" that differences exist. If we can't reject the null hypothesis, we *don't* "prove" ipso facto that differences don't exist. We merely stay inconclusive. Point to KYW.
We can, however, infer from such a result that even if we could not reject the null hypothesis in spite of a resonable sample size (say, 10 - on the lower edge, but quite resonable already), that only small differences may have existed, and not hugely objectionable ones. Conversely, a "significant" difference does not ipso facto mean a *huge* difference. You may hear or measure differences with a high level of significance - but a low absolute value. The differences may exist, but may not matter.
Typical statistical tests only appy under conditions of
- normality
- random sampling (no autocorrelation)
Normality means that the population must follow the bell curve. In this particular case, not only do we test a device, we implicitly test people's perceptions. Sample size constraints apply to both both devices and people! You must conduct the test using many amps from the same series! Just look at the data sheets for op amps, and their standard deviations! Not even to mention transistors.
Random sampling means , not only must you select people, cables, locations, sources, speakers, and rooms, from a large, uncorrelated population... but you also can't bunch a *time* series into a parameric statistic. When you switch amps 10 times in a row, you effectively do a time series. You got data point x just after data point x-1 and just before data point x+1. Therefore you can't lump these data points together into a standard statistical test. And a sampling of 50 A/B switches done sequentially and by 2 people on a single day using 2 amps only with one set of speakers in one and the same room represents a violation of all the above.
Still, in spite of all that, the musical enjoyment matters more than all this to most people, I hope at least. And here I must say, if you can't hear the difference between two components I wouldn't bother either to look for more or pay more. Buy the cheaper one, and enjoy. People enjoy music from cheap boom boxes too. Perfect playback means nothing if you don't enjoy it. And if you enjoy it, playback quality doesn't matter all that much anymore.
All sides have good points but don't refer to the same contexts, and the subject has so many complexities that no one can ever give clear answers. In any case most problems come from the generalization of a few peronal experiences to all situations everywhere and for everyone.
If no differences between electronic devices existed, why would so many serious, earnest designers, many well respected, conduct earnest discussions about amp topologies etc here on these forums? Obviously, *some* people on *some* kinds of equipment hear (and measure!) *some* differences *sometimes*.
I won't even go into the audibility problem - what kind of distortion appears more audible or more objectionable. Again, a vast field of study.
I'll just enlarge the point on statistics.
The statistical problem really exists: sample sizes of less than 15 to 20 really can't give you a reliable statistic. Any statistic can only test for rejection of the null hypothesis. The null hypothesis states that no differences exist. If we can reject this hypothesis with a good significance level, we "prove" that differences exist. If we can't reject the null hypothesis, we *don't* "prove" ipso facto that differences don't exist. We merely stay inconclusive. Point to KYW.
We can, however, infer from such a result that even if we could not reject the null hypothesis in spite of a resonable sample size (say, 10 - on the lower edge, but quite resonable already), that only small differences may have existed, and not hugely objectionable ones. Conversely, a "significant" difference does not ipso facto mean a *huge* difference. You may hear or measure differences with a high level of significance - but a low absolute value. The differences may exist, but may not matter.
Typical statistical tests only appy under conditions of
- normality
- random sampling (no autocorrelation)
Normality means that the population must follow the bell curve. In this particular case, not only do we test a device, we implicitly test people's perceptions. Sample size constraints apply to both both devices and people! You must conduct the test using many amps from the same series! Just look at the data sheets for op amps, and their standard deviations! Not even to mention transistors.
Random sampling means , not only must you select people, cables, locations, sources, speakers, and rooms, from a large, uncorrelated population... but you also can't bunch a *time* series into a parameric statistic. When you switch amps 10 times in a row, you effectively do a time series. You got data point x just after data point x-1 and just before data point x+1. Therefore you can't lump these data points together into a standard statistical test. And a sampling of 50 A/B switches done sequentially and by 2 people on a single day using 2 amps only with one set of speakers in one and the same room represents a violation of all the above.
Still, in spite of all that, the musical enjoyment matters more than all this to most people, I hope at least. And here I must say, if you can't hear the difference between two components I wouldn't bother either to look for more or pay more. Buy the cheaper one, and enjoy. People enjoy music from cheap boom boxes too. Perfect playback means nothing if you don't enjoy it. And if you enjoy it, playback quality doesn't matter all that much anymore.
So you specficied the power level. I hate to break it to you but 100W and 20W isn't a great deal of SPL difference. I, as well as everyone and their baby is aware of the change in sensitivity to distortion with SPL. I've listened to a wide variety of equipment from the lowly Sony up Mark Levinson and CAL, even an MJZ Quadrant class A with a bias current that would make a Sumo or Pass class A look like energy misers.
I'd like to point out again that the botched inplementations and interpretations come mostly from the hi-fi-lie community.
The high distortion of speakers doesn't justify the use of a high distortion amplifier. It's all the more reason to lower amplifier distortion.
I'd like to point out again that the botched inplementations and interpretations come mostly from the hi-fi-lie community.
The high distortion of speakers doesn't justify the use of a high distortion amplifier. It's all the more reason to lower amplifier distortion.
Konnichiwa,
I hate to break it to you, but I was quite well aware of the difference between 20W & 100W (the rule is 10 log(P1/P2) BTW), which is 7db.
I choose the figures very deliberatly.
You may forgive me for saying, but you failed to mention it.
You could have said for example "I, as well as everyone and their baby is aware of the change in sensitivity to distortion with SPL - but I am argue with you about 1% distortion without qualification anyway", but as it so happens, you did not.
Well, I have slept with many girls, some black, some white, some asian, many mixed. I found them all interetsing to the highest degree, but I married one, not many.
So, what are you trying to say with all that meaningless verbiage?
Ahh, you agree after all. The ABX Mafia is to blame.
Good we agree and can dismiss thge topic and can proceed to more productive issues. (Sorry, but my definition of a lie is to knowingly clai what is known to be divergent from reality - the ABX Mafia is guilty as charged for about the last 20 Years or so).
Why not? How about this simple scenario. I'll be High Tech for arguments sake.
Let us say I took a DSP and a digital amplifier. I provied the combo with a sytem capable of measuring the distortion of the whole system. I now allowed a DSP algorythm to compensate the Speakers distortion, using a highly linear measurement mirophone and highly intelligent (read semi - ai) software. The result would be a system where the amplifier would have measurably VERY HIGH levels of distortion, yet the system overall would be largely free from distortion.
But of course, you could not possibly aspire to such a system, because after all, the amplifier would have a lot of distortion and distortion is BAD.
Sayonara
(BTW, the above amplifier/speaker system is something I proposed on the "Bass" e-mail list in the mid 1990's, primarily to address room issues, but the point is that any given distortion can be to a reasonasble degree offset by suitable pre-distortion.)
hummhoom said:
I hate to break it to you, but I was quite well aware of the difference between 20W & 100W (the rule is 10 log(P1/P2) BTW), which is 7db.
I choose the figures very deliberatly.
hummhoom said:
You may forgive me for saying, but you failed to mention it.
You could have said for example "I, as well as everyone and their baby is aware of the change in sensitivity to distortion with SPL - but I am argue with you about 1% distortion without qualification anyway", but as it so happens, you did not.
hummhoom said:
Well, I have slept with many girls, some black, some white, some asian, many mixed. I found them all interetsing to the highest degree, but I married one, not many.
So, what are you trying to say with all that meaningless verbiage?
hummhoom said:
Ahh, you agree after all. The ABX Mafia is to blame.
Good we agree and can dismiss thge topic and can proceed to more productive issues. (Sorry, but my definition of a lie is to knowingly clai what is known to be divergent from reality - the ABX Mafia is guilty as charged for about the last 20 Years or so).
hummhoom said:
Why not? How about this simple scenario. I'll be High Tech for arguments sake.
Let us say I took a DSP and a digital amplifier. I provied the combo with a sytem capable of measuring the distortion of the whole system. I now allowed a DSP algorythm to compensate the Speakers distortion, using a highly linear measurement mirophone and highly intelligent (read semi - ai) software. The result would be a system where the amplifier would have measurably VERY HIGH levels of distortion, yet the system overall would be largely free from distortion.
But of course, you could not possibly aspire to such a system, because after all, the amplifier would have a lot of distortion and distortion is BAD.
Sayonara
(BTW, the above amplifier/speaker system is something I proposed on the "Bass" e-mail list in the mid 1990's, primarily to address room issues, but the point is that any given distortion can be to a reasonasble degree offset by suitable pre-distortion.)
Keyser, if crank-winding is what you like, you can have lots of fun reading James Randi's website, where the degree of creativity in rationalization exhibited by people failing controlled tests of various sorts well exceeds that of the high end audio community. The sorts of excuses about why claimed audible phenomena vanish when the nameplate is covered are pretty repetitious- there's been nothing new in that regard for as long as I've been doing audio (about 35 years). The new frontiers for crankdom are in areas like health and energy, where there's always something new, novel, and fraudulent to defend.
If you're trying to get a better understanding of protocols, statistics, and sensitivities of proper controlled testing, you'd do better to read the papers of the experts, starting with Stanley Lipshitz, Richard Greiner, and Floyd Toole. It will prove more educational than sorting through the mish-mash of fuzzy "logic" you'll get from reading audio bulletin boards.
If you're trying to get a better understanding of protocols, statistics, and sensitivities of proper controlled testing, you'd do better to read the papers of the experts, starting with Stanley Lipshitz, Richard Greiner, and Floyd Toole. It will prove more educational than sorting through the mish-mash of fuzzy "logic" you'll get from reading audio bulletin boards.
Konnichiwa,
Nope. I merely pointed out that if you wish to conduct a test to give you the answer to a specific question then you must make sure your test is apropriate and able to answer the question. Like if you wish to determine an amplifiers frequency response it would be ill advised to use a THD meter.
A very large number. For example distortion measurements require an analysis according to spectral content, noiseloading can reveal non-harmonic distortion. TIM is readily aknowleged to be of concern, PIM less so which makes it however non the less important. There is much documentation on the whole subject for the last 5 ot 6 Decades, why not apply yourself to them?
Sayonara
keyser said:
Nope. I merely pointed out that if you wish to conduct a test to give you the answer to a specific question then you must make sure your test is apropriate and able to answer the question. Like if you wish to determine an amplifiers frequency response it would be ill advised to use a THD meter.
keyser said:
A very large number. For example distortion measurements require an analysis according to spectral content, noiseloading can reveal non-harmonic distortion. TIM is readily aknowleged to be of concern, PIM less so which makes it however non the less important. There is much documentation on the whole subject for the last 5 ot 6 Decades, why not apply yourself to them?
Sayonara
Bottom line of this subject, in audio some elements of the chain have found fairly satisfactorily solutions to the ears of most people. Amplifiers seem to qualify here, since we have this discussion: to some people most amps sound the same, to some people they don't. Speakers obviously don't qualify - almost eeryone agrees that they sound different from model to model. Sadly we have to test amps through speakers - ultimately the amp should drive a speaker, not test equipment - and so we have a problem: we test a more perfect machine through a less perfect one.
To SY: in every field of sensory perception you find the kind of discussion we have here: cooking, photography, musical performance (by artist, not by electronics). In all of these some people don't perceive, or don't mind, the minute differences from one to the other "interpretation of the theme". And mixed in all of that you do certainly have fraudulent claims, rationalization of non existing improvements etc.
But to go from there to claim that no differences axist in any and all configurations for all times, because "the problem has found a satisfactory universal solution" or along these lines, goes a bit far doesn't it? You can't just throw out the kid with the bath water and claim that the food tastes the ame in all restaurants, that all car tires perform equally well if well engineered and inflated to manufacturer's specs etc.
And I wouldn't call all these competent engineers discussing amp topologies on these boards here, fraudulent baiters for the poor unwashed masses.
To SY: in every field of sensory perception you find the kind of discussion we have here: cooking, photography, musical performance (by artist, not by electronics). In all of these some people don't perceive, or don't mind, the minute differences from one to the other "interpretation of the theme". And mixed in all of that you do certainly have fraudulent claims, rationalization of non existing improvements etc.
But to go from there to claim that no differences axist in any and all configurations for all times, because "the problem has found a satisfactory universal solution" or along these lines, goes a bit far doesn't it? You can't just throw out the kid with the bath water and claim that the food tastes the ame in all restaurants, that all car tires perform equally well if well engineered and inflated to manufacturer's specs etc.
And I wouldn't call all these competent engineers discussing amp topologies on these boards here, fraudulent baiters for the poor unwashed masses.
That's a rather mixed bag list, encompassing some things that are well outside of sensory issues. But certainly in fields of sensory research that I've worked in (haptic and organoleptic), one wouldn't even think about claiming any sort of significance in uncontrolled experiments.
Please provide a precise citation for a statement like this. I've never seen or heard of such a thing, except in the straw-man arguments of the goo-goo tweak squadrons.
Of course not.
Personal experience such as "I tested amp A vs amp B and found no difference" makes for good, solid common sense every day decisions. Telling other people on a forum makes for interesting input.
But statements worded in the fashion of wide ranging universal applicability, such as "high end gear sounds no better than off the shelf consumer gear" need much, much more careful experimentation, setup, and controls.
I think most unresolved arguments in these forums boil down to people extrapolating their personal, valid and sound (sic) experiences, to some kind of universal truth statement.
Not to speak of the abuse of words such as "significant".
Actually, my mistake, I reworded from the tone of the thread starter. Nobody used this exact phrase. Posts 1 and 22 contain statements of similar vein, to my understanding.
I have to admit I also believe most electronics and CD players, and cables, etc, show pretty minor differences. But I would never claim that they don't matter, or that, as in post # 1, just speaker, recording and room make a difference to the sound (post #1, last paragraph).
Konnichiwa,
I was tempted to write a long missive on the nature of the whole "audio subjectivist" vs. "audio objectivists" as a primarily ideological debate on the positions and mindsets often mislabeled "conservative" and "liberal".
But having just finished a long sunday afternoons work on the chassis of my next Amp I think I'll just go, get a few beers and kick back instead.
In order to still make my point, instead of the long missive I shall recommend a nice bit of reading, namely "The ILLUMINATUS! Trilogy" by Robert Anton Wilson & Robert Shea around 30 years ago....
Oh yes, and to all wanting to continue to argue with me, read the tagline!
Aum, Sayonara
I was tempted to write a long missive on the nature of the whole "audio subjectivist" vs. "audio objectivists" as a primarily ideological debate on the positions and mindsets often mislabeled "conservative" and "liberal".
But having just finished a long sunday afternoons work on the chassis of my next Amp I think I'll just go, get a few beers and kick back instead.
In order to still make my point, instead of the long missive I shall recommend a nice bit of reading, namely "The ILLUMINATUS! Trilogy" by Robert Anton Wilson & Robert Shea around 30 years ago....
Oh yes, and to all wanting to continue to argue with me, read the tagline!
Aum, Sayonara
Konnichiwa,
Maybe. While we are at a reading list for users of human society at large, I often quote R.A.W. from Illuminatus in the fictional bookelett "Never whistle while your pi55ing", like here:
And now for something completely different....
Sayonara
MBK said:R.A.W. also has a piece on General Semantics on his site, maybe one should make it required reading for forum users
Maybe. While we are at a reading list for users of human society at large, I often quote R.A.W. from Illuminatus in the fictional bookelett "Never whistle while your pi55ing", like here:
And now for something completely different....
Sayonara
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.