Konnichiwa,
You really think so? Why then would conduct such tests theselves? I would argue along with you that stereophile has certain interests, but I do think you are taking things more than a bit to far.
Hmmm, the editor is a quite noteworthy sound engineer, considering the level and professionalism of the measurements taken as part of Stereophile's reports I would suggest that in fact there is (or more precisely MUST BE) considerable professional expertise outside writing & entertainment.
Neither is a paper by Lipshitz, who equally has his own agenda.
I still strongly feel that both sides of the arguments should be heard and that it is far from setteled. As you noticed I have provided readily available pointers to on-line information covering both sides of the argument. Obviously you feel that you prefer the partisan approach of taking sides and trying to simply dismiss information that is is support of the "other party's" position, something I do not find all that agreeable.
And I do not consider you objections to the material presented in the linked Stewreophile articles to be valid. Have you actually even read them?
Sayonara
SY said:
You really think so? Why then would conduct such tests theselves? I would argue along with you that stereophile has certain interests, but I do think you are taking things more than a bit to far.
SY said:
Hmmm, the editor is a quite noteworthy sound engineer, considering the level and professionalism of the measurements taken as part of Stereophile's reports I would suggest that in fact there is (or more precisely MUST BE) considerable professional expertise outside writing & entertainment.
SY said:
Neither is a paper by Lipshitz, who equally has his own agenda.
I still strongly feel that both sides of the arguments should be heard and that it is far from setteled. As you noticed I have provided readily available pointers to on-line information covering both sides of the argument. Obviously you feel that you prefer the partisan approach of taking sides and trying to simply dismiss information that is is support of the "other party's" position, something I do not find all that agreeable.
And I do not consider you objections to the material presented in the linked Stewreophile articles to be valid. Have you actually even read them?
Sayonara
I am not interested in exploring all sorts of hifi-enhousiast views and opinions about Double Blind Tests. I don't think the protocol for a DBT on the subject of hifi should be that much different than the protocol for testing anything else. There should be some kind of standard DBT protocol, isnt there?
It is my intention to do this test as scientific as possible. The protocol should eliminate any kind of bias, including my own. I expect not to hear very much difference between the test components. The people who responded to my invitation till now, seem to believe they will be perfectly capable hearing the difference. Perfect! I think I should not take part in the actual test, for I might not hear a difference, even when there are some.
I am not very experienced in soldering small parts, but I do believe a ABX comparator could be beneficial in performing the tests. I'm not sure, but isn't there any ABX software available on the internet, so that I can control A, B and X using a pc?
It is my intention to do this test as scientific as possible. The protocol should eliminate any kind of bias, including my own. I expect not to hear very much difference between the test components. The people who responded to my invitation till now, seem to believe they will be perfectly capable hearing the difference. Perfect! I think I should not take part in the actual test, for I might not hear a difference, even when there are some.
I am not very experienced in soldering small parts, but I do believe a ABX comparator could be beneficial in performing the tests. I'm not sure, but isn't there any ABX software available on the internet, so that I can control A, B and X using a pc?
There's no "universal" protocol. It's like saying, "How do I design an amplifier?" You can't answer that question without making it sharper, defining exactly what it is you want to test and for what purpose.
And here the miracle happened: SY and KYW proposing similar things! Nothing sure in science, neither methods nor results. Why else would people keep on publishing?
Keyser, yes, if you want to make it scientific, you WILL have to think about things a little, and you'll have to tailor them to your questions. That means you'lll have to define your questions first. Both SY and KYW have presented you with reading according to their own preferences, you'll have to choose by yourself now. Why not read through opposing viewpoints and agendas? That should balance things out anyway. Aren't you curious to find out, through rigorous testing, to find out who got it right?
If you don't want to make it scientific, because of the tedium involved, you can always do what I suggested above: DIY a little, mod a little, compare a little, use sound files and test CD's to sharpen your ear, and yes, use your own ears, etc. Frankly, DIY helps a lot to clarify things.
Keyser, yes, if you want to make it scientific, you WILL have to think about things a little, and you'll have to tailor them to your questions. That means you'lll have to define your questions first. Both SY and KYW have presented you with reading according to their own preferences, you'll have to choose by yourself now. Why not read through opposing viewpoints and agendas? That should balance things out anyway. Aren't you curious to find out, through rigorous testing, to find out who got it right?
If you don't want to make it scientific, because of the tedium involved, you can always do what I suggested above: DIY a little, mod a little, compare a little, use sound files and test CD's to sharpen your ear, and yes, use your own ears, etc. Frankly, DIY helps a lot to clarify things.
Konnichiwa,
Much of one of the stereophile reference I gave was strictly about the statistics.
To sumarise, a .05 level of significance means that if you reject the null hypothesis there is a 95% certainty that it was rejected because of an actual difference and not because of chance. If we recject the null hypothesis even though it is actually true it is called a "Type 1" or "Type A" statistical error.
There is a second error, namely that we failed to reject the null hypothesis even though a difference was present. This is called a "Type 2" or "Type B" statistical error. There is a direct inverse relation between the risk or type 1 and type errors and a further relation to the size of the sample.
In a very large sample even a very small percentage deviation from nominal (nominal being 50%) is significant (or avoids a type 1 trror), the smaller the sample size the higher the percentage required to give a significant result (or avoid a type 1 error).
HOWEVER, the more we are certain that any rejection of the null hypothesis is not due to error the less secure we can be that we have not risked a type 2 error to an unreasonable degree. In small sample size statistical tests the risk of a type 2 error (eg we fail to reject the null hypothesis in the presence of an audible difference) becomes high enough to suggest near certainty.
I completely agree. With that you are at two points.
#1 the "double blind" element. It simply implies that both subject and the person operating the experiment are unaware of the identy of the two tested items.
#2 the statistical analysis of the results produced using any number of the available statistical menthodee.
There is. And it is actually straightforward. Note that I provided a link to a basic but sufficient explanation of Double Blind Testing (non-audio) for you in my earlier post.
To me that would imply that you first and foremost MUST test your test itself, against known audible differences. You also will be required to ensure a sufficiently large dataset is compiled that you can gain the required information using statistical methodes that minimise the risk of errors (type 1 & type 2).
I think you will find that your test is far from perfect, simply due to the actual prejudices present.
The corollary is that your other subjects also should not partake, because they may hear a difference where non exists. Any significant bias about the expectation of whats is being presented will skew the result.
You can test that extremely easily by using a definitly audible (but near the limits of conventionally aknowleged audibility) test stimulus, but suggesting to the test subjects that what they hear is in fact something where the audibility is debated and where they hold strong opinions.
Then you may have buy one. However, you can readily appoint someone to change the connections out of sight of the test subjects manually and to do so according to a preset protocol that has been carefully arranged to avoid potential bias towards suggesting a difference in the absence of one that is sufficiently audible, which can happen if you for example have a greater number of "same" or "different" pairings in you test-data in the presence of subjects with definite preconceptions.
So, if you present more "same" pairs than "different" pairs then the apparent result will support the null hypothesis if enough "all equipment sounds the same" believers are in the test subject set, EVEN IF THERE IS AN AUDIBLE difference.
Equally, if you present more "different" pais than "same" pairs then the apparent result will reject the null hypothesis if enough "all equipment sounds different" believers are in the test subject set, EVEN IF THERE IS NO AUDIBLE difference.
The other stereophile reference I provided was actually about a large size testset blind test of amplifiers which seemed to suggest audible differences existed simply because the presentation given included (unintentionally I might add) more "different" pairs and was done in front of an audience mainly consisting of "audiophiles", thus implicitly by people predisposed to hearing differences where non exist.
You may have noticed, had you checked out the links I provided that I in fact pointed you to a site that has not only such software, a number of samples that can be used to train ones hearing AND quite a bit of good advise on how to set up the test, not in the staistical or purely technical aspect, but in the practical aspect.
Of course, if you use a PC as source you are limited by it's audio capability.
I wish you luck in getteing enough participants to sign up to make the test meaningfull, statistically speaking and also in arranging a test envoinment free of peer pressure (meaning only single individulas are present for the test, not groups), free from undue stress (in my view it is essential to kep the numbers of "trials" in a single setting to no more then 5), and with a system and room setup that does not prejudice the potential audibility of differences by burrying the under tons of distortion or room related effects.
I typhically recommend wherever possible the use of Stax Electrostatic "earspeakers" with the neccesary addition of free-field frequency response correction and natural crossfeed circuitry if the recordings used for evaluation are aimed at loudspeaker reproduction (in other words non-binaural).
I find that such a test setup minimises the chance of poor room setup and /or poor speaker choice and because we are invariably down to single individuals it also minimises the chance of peer pressure.
Certainly if speakers are used they should be of a controlled directivity type, with demonstrably low levels of distortion and compression as well providing reasonable waveform preservation (which excludes emphatically Canton and B&W as well as most other generic socalled "HiFi"speakers).
Sayonara
keyser said:
Much of one of the stereophile reference I gave was strictly about the statistics.
To sumarise, a .05 level of significance means that if you reject the null hypothesis there is a 95% certainty that it was rejected because of an actual difference and not because of chance. If we recject the null hypothesis even though it is actually true it is called a "Type 1" or "Type A" statistical error.
There is a second error, namely that we failed to reject the null hypothesis even though a difference was present. This is called a "Type 2" or "Type B" statistical error. There is a direct inverse relation between the risk or type 1 and type errors and a further relation to the size of the sample.
In a very large sample even a very small percentage deviation from nominal (nominal being 50%) is significant (or avoids a type 1 trror), the smaller the sample size the higher the percentage required to give a significant result (or avoid a type 1 error).
HOWEVER, the more we are certain that any rejection of the null hypothesis is not due to error the less secure we can be that we have not risked a type 2 error to an unreasonable degree. In small sample size statistical tests the risk of a type 2 error (eg we fail to reject the null hypothesis in the presence of an audible difference) becomes high enough to suggest near certainty.
keyser said:
I completely agree. With that you are at two points.
#1 the "double blind" element. It simply implies that both subject and the person operating the experiment are unaware of the identy of the two tested items.
#2 the statistical analysis of the results produced using any number of the available statistical menthodee.
keyser said:
There is. And it is actually straightforward. Note that I provided a link to a basic but sufficient explanation of Double Blind Testing (non-audio) for you in my earlier post.
keyser said:
To me that would imply that you first and foremost MUST test your test itself, against known audible differences. You also will be required to ensure a sufficiently large dataset is compiled that you can gain the required information using statistical methodes that minimise the risk of errors (type 1 & type 2).
keyser said:
I think you will find that your test is far from perfect, simply due to the actual prejudices present.
keyser said:
The corollary is that your other subjects also should not partake, because they may hear a difference where non exists. Any significant bias about the expectation of whats is being presented will skew the result.
You can test that extremely easily by using a definitly audible (but near the limits of conventionally aknowleged audibility) test stimulus, but suggesting to the test subjects that what they hear is in fact something where the audibility is debated and where they hold strong opinions.
keyser said:
Then you may have buy one. However, you can readily appoint someone to change the connections out of sight of the test subjects manually and to do so according to a preset protocol that has been carefully arranged to avoid potential bias towards suggesting a difference in the absence of one that is sufficiently audible, which can happen if you for example have a greater number of "same" or "different" pairings in you test-data in the presence of subjects with definite preconceptions.
So, if you present more "same" pairs than "different" pairs then the apparent result will support the null hypothesis if enough "all equipment sounds the same" believers are in the test subject set, EVEN IF THERE IS AN AUDIBLE difference.
Equally, if you present more "different" pais than "same" pairs then the apparent result will reject the null hypothesis if enough "all equipment sounds different" believers are in the test subject set, EVEN IF THERE IS NO AUDIBLE difference.
The other stereophile reference I provided was actually about a large size testset blind test of amplifiers which seemed to suggest audible differences existed simply because the presentation given included (unintentionally I might add) more "different" pairs and was done in front of an audience mainly consisting of "audiophiles", thus implicitly by people predisposed to hearing differences where non exist.
keyser said:
You may have noticed, had you checked out the links I provided that I in fact pointed you to a site that has not only such software, a number of samples that can be used to train ones hearing AND quite a bit of good advise on how to set up the test, not in the staistical or purely technical aspect, but in the practical aspect.
Of course, if you use a PC as source you are limited by it's audio capability.
I wish you luck in getteing enough participants to sign up to make the test meaningfull, statistically speaking and also in arranging a test envoinment free of peer pressure (meaning only single individulas are present for the test, not groups), free from undue stress (in my view it is essential to kep the numbers of "trials" in a single setting to no more then 5), and with a system and room setup that does not prejudice the potential audibility of differences by burrying the under tons of distortion or room related effects.
I typhically recommend wherever possible the use of Stax Electrostatic "earspeakers" with the neccesary addition of free-field frequency response correction and natural crossfeed circuitry if the recordings used for evaluation are aimed at loudspeaker reproduction (in other words non-binaural).
I find that such a test setup minimises the chance of poor room setup and /or poor speaker choice and because we are invariably down to single individuals it also minimises the chance of peer pressure.
Certainly if speakers are used they should be of a controlled directivity type, with demonstrably low levels of distortion and compression as well providing reasonable waveform preservation (which excludes emphatically Canton and B&W as well as most other generic socalled "HiFi"speakers).
Sayonara
Thanks KYW!
I have read your post, and you seem to have gone about it very seriously! I cannot yet give a good comment on it, for I have not looked at all the links you provided previously (yes, you were right. I had a look at some of them only).I'll come back to it later. For me it actually came as a surprise to get this kind of useful information from you. For now thank you for the help you are offering.
keyser
I have read your post, and you seem to have gone about it very seriously! I cannot yet give a good comment on it, for I have not looked at all the links you provided previously (yes, you were right. I had a look at some of them only).I'll come back to it later. For me it actually came as a surprise to get this kind of useful information from you. For now thank you for the help you are offering.
keyser
Konnichiwa,
I have been involved in some (non too many) Audio DB Tests and even run some myself. This among other things tought me something about both the utility and problems of such tests.
I now tend to use blind testing strictly in order to establish preference with basic pre-conception that differences do exist and people are merely asked to rate their preferences. The actual test memains blind and indeed includes some "same" pairs as well. I find that people tent to be pretty consistent about what they like or not, suggesting that they (oviously) hear the differences and at least qunatify their subjective reaction to them.
No worries, I am quite interested in the actual truth. So if you are looking for it I'll be ready to contribute what little experience and knowledge I happen to have.
Sayonara
keyser said:
I have been involved in some (non too many) Audio DB Tests and even run some myself. This among other things tought me something about both the utility and problems of such tests.
I now tend to use blind testing strictly in order to establish preference with basic pre-conception that differences do exist and people are merely asked to rate their preferences. The actual test memains blind and indeed includes some "same" pairs as well. I find that people tent to be pretty consistent about what they like or not, suggesting that they (oviously) hear the differences and at least qunatify their subjective reaction to them.
keyser said:
No worries, I am quite interested in the actual truth. So if you are looking for it I'll be ready to contribute what little experience and knowledge I happen to have.
Sayonara
....Then you may have buy one. However, you can readily appoint someone to change the connections out of sight of the test subjects manually and to do so according to a preset protocol that has been carefully arranged to avoid potential bias towards suggesting a difference in the absence of one that is sufficiently audible, which can happen if you for example have a greater number of "same" or "different" pairings in you test-data in the presence of subjects with definite preconceptions....
I would second that. The only time I attended a 'controlled DBT' (by the Boston Audio Society) the amps under test had different DC offsets. It was easy to learn to ID the 'click' in switching between the two. Only I noticed, in a room full of 'experienced listeners'.
I would second that. The only time I attended a 'controlled DBT' (by the Boston Audio Society) the amps under test had different DC offsets. It was easy to learn to ID the 'click' in switching between the two. Only I noticed, in a room full of 'experienced listeners'.
God knows what I'm getting into here......
As a hobbyist and pro audio designer heres my black and white carved in stone theology!
1/ original recording, speakers and room acoustics do indeed totally dominate what we hear.
2/ equipment supports, stands, spikes, cones etc make no difference, with the exception of speaker stands and turntable supports which for obvious reasons do make a difference.
3/ Amplifiers do make a surprisingly large difference. I've designed and built loads of 'em, both commercially and, in the main, as a hobby and It never ceases to amaze me just how different they can sound, however, see point 1/ above for perspective.
4/ cables make no difference so long as, in the case of speaker cable, the resistance is nice and low. there should be a special ward in mental institutes for those who go as far as to suggest that mains cable makes a difference
5/ Components make little difference to sound in MOST cases, exceptions can be found here though, such as very low level circuitry... try using big electrolytics in a MC head amp, for example, and yes, it will spoil the sound. There are measurable attributes of electrolytics that can explain this though. 97% of the time it is the topology and layout of an amp that decides it's sound and not the make of the resistors etc.
6/ power supplies vastly larger than required make no difference.
It makes me laugh when I hear about people retro-fitting 1000va transformers to pre-amps (for example) and claiming an improvement..... Most do not seem to realise that the power almost certainly goes through a regulator stage and that PSU's any larger than that required to keep the regulator well above drop out can make no difference.
I could rant and rave for much longer but will refrain...these are my own "carved in rock" views and I have no intention of wearing out my keyboard in defending them😀
discuss amongst yourselves......
As a hobbyist and pro audio designer heres my black and white carved in stone theology!
1/ original recording, speakers and room acoustics do indeed totally dominate what we hear.
2/ equipment supports, stands, spikes, cones etc make no difference, with the exception of speaker stands and turntable supports which for obvious reasons do make a difference.
3/ Amplifiers do make a surprisingly large difference. I've designed and built loads of 'em, both commercially and, in the main, as a hobby and It never ceases to amaze me just how different they can sound, however, see point 1/ above for perspective.
4/ cables make no difference so long as, in the case of speaker cable, the resistance is nice and low. there should be a special ward in mental institutes for those who go as far as to suggest that mains cable makes a difference
5/ Components make little difference to sound in MOST cases, exceptions can be found here though, such as very low level circuitry... try using big electrolytics in a MC head amp, for example, and yes, it will spoil the sound. There are measurable attributes of electrolytics that can explain this though. 97% of the time it is the topology and layout of an amp that decides it's sound and not the make of the resistors etc.
6/ power supplies vastly larger than required make no difference.
It makes me laugh when I hear about people retro-fitting 1000va transformers to pre-amps (for example) and claiming an improvement..... Most do not seem to realise that the power almost certainly goes through a regulator stage and that PSU's any larger than that required to keep the regulator well above drop out can make no difference.
I could rant and rave for much longer but will refrain...these are my own "carved in rock" views and I have no intention of wearing out my keyboard in defending them😀
discuss amongst yourselves......
Konnichiwa,
This seems in contradiction with:
If your contention 1 is true than 3 is untrue, or the other way around. Only one of the two can be considered true.
Assuming other electronics (CD Players, Amplifiers etc.) could be illustrated to contain comparably microphonic components in the direct or shunt signal path, would you conceed the possibility of differences also for these? Hint, Electrolytic Capacitors are by comparison highly microphonic, if they have a polarisation voltage they make non too bad sensitivity capacitor microphones, but with terrible frequency response.
Hmmm. What about inductance? What about amplifier stability into capacitive loads (I agree, Amplifiers SHOULD be stable into any load, but many are not, a sideeffect of employing large amounts of negative feedback poorely)? What about RFI entering the Amplifier output stage and feedback loop from the aerial-operation of the speaker cable?
Then I guess I need to be sectioned. I know there are VERY GOOD technical reasons why the basic electrical behaviour (RLC & RFI Rejection) of a Mainscable may make a readily measurable and potentially audible difference in a system with multiple interconnected, mainspowered appliances. It is a recognised problem in measurement systems among others. The key are noiseloops.
Transformer coupled balanced gear (in essence real Pro Gear) is in effect completely impervious to this type of problem, balanced gear with electronically balanced inputs is fairly resistant to the problem but not entierly so, any unbalanced connections, especially those commonly found in domestic equipment are very problematic.
I largely agree, though I do find that resistor construction as such (eg the number of interfaces between dissimilar metals and the way these interfaces are handled) and similar issues in Capacitors can be surprisingly significant. One example would be a nichrome metal film resistor, with pressed on endcaps made from steel and with welded on copper leadouts. This gives a total of 4 Junctions to develop (parasitic) thermoelement voltages and equally currentflow dependent noise voltages.
The possibility of such improvements will depend on a lot of things. If you have poor quality mains power, with loads of harmonic distortion and DC offset a much higher power transformer may operate further away from saturation, as an example. This may result in a number of measurable differences. Equally, the larger power rating transformer may have a lower selfresonance and thusly a reduced RFI talktrhough from the mains grid. These things may indeed make differences as may other construction details of the mains transformer (presence or absence of electrostatic screens etc), though the actual differences may not be a direct result of the increased power rating of the transformer.
You do not seem to realise that regulators have well defined and limited levels of ripple rejection that tend to fall of quickly with rising frequency, just as most circuitry (except non-looped feedback circuitry) suffers reduced PSRR as frequency rises. So, if something happens to change in the harmonic spectrum of the rectified Voltage (eg reduced higher harmonics) the result may very be sufficient to be verified by measurement and to have at least the potential for audibility.
Note, I do not per se claim that any of the above things are audbile or not, but that mechanisms exist that would allow them to produce a measurable result that is well within the potentiality of audibility (eg within a -90db limit), even though audibility or not is highly debatable.
Sayonara
jez said:
This seems in contradiction with:
jez said:
If your contention 1 is true than 3 is untrue, or the other way around. Only one of the two can be considered true.
jez said:
Assuming other electronics (CD Players, Amplifiers etc.) could be illustrated to contain comparably microphonic components in the direct or shunt signal path, would you conceed the possibility of differences also for these? Hint, Electrolytic Capacitors are by comparison highly microphonic, if they have a polarisation voltage they make non too bad sensitivity capacitor microphones, but with terrible frequency response.
jez said:
Hmmm. What about inductance? What about amplifier stability into capacitive loads (I agree, Amplifiers SHOULD be stable into any load, but many are not, a sideeffect of employing large amounts of negative feedback poorely)? What about RFI entering the Amplifier output stage and feedback loop from the aerial-operation of the speaker cable?
jez said:there should be a special ward in mental institutes for those who go as far as to suggest that mains cable makes a difference
Then I guess I need to be sectioned. I know there are VERY GOOD technical reasons why the basic electrical behaviour (RLC & RFI Rejection) of a Mainscable may make a readily measurable and potentially audible difference in a system with multiple interconnected, mainspowered appliances. It is a recognised problem in measurement systems among others. The key are noiseloops.
Transformer coupled balanced gear (in essence real Pro Gear) is in effect completely impervious to this type of problem, balanced gear with electronically balanced inputs is fairly resistant to the problem but not entierly so, any unbalanced connections, especially those commonly found in domestic equipment are very problematic.
jez said:
I largely agree, though I do find that resistor construction as such (eg the number of interfaces between dissimilar metals and the way these interfaces are handled) and similar issues in Capacitors can be surprisingly significant. One example would be a nichrome metal film resistor, with pressed on endcaps made from steel and with welded on copper leadouts. This gives a total of 4 Junctions to develop (parasitic) thermoelement voltages and equally currentflow dependent noise voltages.
jez said:
The possibility of such improvements will depend on a lot of things. If you have poor quality mains power, with loads of harmonic distortion and DC offset a much higher power transformer may operate further away from saturation, as an example. This may result in a number of measurable differences. Equally, the larger power rating transformer may have a lower selfresonance and thusly a reduced RFI talktrhough from the mains grid. These things may indeed make differences as may other construction details of the mains transformer (presence or absence of electrostatic screens etc), though the actual differences may not be a direct result of the increased power rating of the transformer.
jez said:
You do not seem to realise that regulators have well defined and limited levels of ripple rejection that tend to fall of quickly with rising frequency, just as most circuitry (except non-looped feedback circuitry) suffers reduced PSRR as frequency rises. So, if something happens to change in the harmonic spectrum of the rectified Voltage (eg reduced higher harmonics) the result may very be sufficient to be verified by measurement and to have at least the potential for audibility.
Note, I do not per se claim that any of the above things are audbile or not, but that mechanisms exist that would allow them to produce a measurable result that is well within the potentiality of audibility (eg within a -90db limit), even though audibility or not is highly debatable.
Sayonara
KYW,
As I said,room acoustics etc DOMINATE what we hear, I did not say that this was to the n'th degree so that amplifier differences could not be heard...
The other point to which I take exception is the "you do not seem to understand" type remarks..... Yes I do, I am a professional design engineer, not that I'm claiming that this makes me any better than other, obviously very knowledgeable people on this site, just that yes I do understand the theory behind what I'm saying but continue to believe that any "psuedo-science" behind such things as mains cable is just that....(We could all make arguements for phenomena happening at -140dB or at sub atomic level, some, it would seem, would like to claim that cable should be aligned with the rising sun and have insulation applied by virgins on a full moon, It's all "emperors new clothes" to me!).
I guess that this thread was started to provoke the most base instincts of audiophiles and yes, I could'nt resist!, I'm not, however, going to engage in endless semantics!.
Have fun.
As I said,room acoustics etc DOMINATE what we hear, I did not say that this was to the n'th degree so that amplifier differences could not be heard...
The other point to which I take exception is the "you do not seem to understand" type remarks..... Yes I do, I am a professional design engineer, not that I'm claiming that this makes me any better than other, obviously very knowledgeable people on this site, just that yes I do understand the theory behind what I'm saying but continue to believe that any "psuedo-science" behind such things as mains cable is just that....(We could all make arguements for phenomena happening at -140dB or at sub atomic level, some, it would seem, would like to claim that cable should be aligned with the rising sun and have insulation applied by virgins on a full moon, It's all "emperors new clothes" to me!).
I guess that this thread was started to provoke the most base instincts of audiophiles and yes, I could'nt resist!, I'm not, however, going to engage in endless semantics!.
Have fun.
How selective can we be in defending our own pet theories.
From the point of view of any of the numerous rabid objectivists populating this site anyone who hears "surprisingly large difference" between amplifiers deserves to be placed in an adjoining ward.
Konnnichiwa,
Note, I did not suggest that problems where at a -140db but at -90db and depending on conditions higher. Of course, they only occur in a complete system (and not in all conditions and all combinations) and NOT if you use well designed test gear on a fully isolated (mains) circuit. Also, the more complex a system is (eg the more components) the greater chance of the problems becoming significant.
Equally, depending upon regulator IC, Audio Circuit and capacitors used in the supply significant (eg a few mV) levels of high frequency or supersonic noise can ride through to the output at levels that fit the -90db criteria easily.
Now with respect, dither at -90db has been shown to cause audible differences. Why not other signals at the same level? I feel to simply (and blindly) exclude the possibility is shortsighted, to use a "soft" word.
The problem with your view is that it in effects supports the Merchants offering "snakeoil rendered by vestal virgins during full moon", as you (and the whole crowd that is with you) refuse to seriously investigate the subject and thus to provide a position from where we can separate the actual science and the possibility of "good engineering" at the fringe (which may be regardless of any audibility or not) and thus allow the enduser to distinguish good design (or rather design apropriate to the subject) from behaviour bordering at fraud.
Take the example of mains cables again. A company called Eupen makes cables for which they illustrate significant levels of RFI rejection (though sadly not low inductance). These cables are actually made for a number of applications completely outside audio BUT they (of course) perform the same functions in the context of audio equipment. In fact, a certain company re-packages apparently standard Eupen cables in "High End" livery and sells them at a greatly increased endcustomer prices. There is supposedly a number of changes which make it different from standard Eupen cables, but this appears to not include the part number on the moulded cord....
Now if we had a modicum of science in the whole discussion, instead an argument of purely religeuous systems of belief all such practices as the one above would disappear, not because people are inherently honest, but because under more rational conditions there would not be any need for subterfuge.
Sayonara
jez said:
Note, I did not suggest that problems where at a -140db but at -90db and depending on conditions higher. Of course, they only occur in a complete system (and not in all conditions and all combinations) and NOT if you use well designed test gear on a fully isolated (mains) circuit. Also, the more complex a system is (eg the more components) the greater chance of the problems becoming significant.
Equally, depending upon regulator IC, Audio Circuit and capacitors used in the supply significant (eg a few mV) levels of high frequency or supersonic noise can ride through to the output at levels that fit the -90db criteria easily.
Now with respect, dither at -90db has been shown to cause audible differences. Why not other signals at the same level? I feel to simply (and blindly) exclude the possibility is shortsighted, to use a "soft" word.
The problem with your view is that it in effects supports the Merchants offering "snakeoil rendered by vestal virgins during full moon", as you (and the whole crowd that is with you) refuse to seriously investigate the subject and thus to provide a position from where we can separate the actual science and the possibility of "good engineering" at the fringe (which may be regardless of any audibility or not) and thus allow the enduser to distinguish good design (or rather design apropriate to the subject) from behaviour bordering at fraud.
Take the example of mains cables again. A company called Eupen makes cables for which they illustrate significant levels of RFI rejection (though sadly not low inductance). These cables are actually made for a number of applications completely outside audio BUT they (of course) perform the same functions in the context of audio equipment. In fact, a certain company re-packages apparently standard Eupen cables in "High End" livery and sells them at a greatly increased endcustomer prices. There is supposedly a number of changes which make it different from standard Eupen cables, but this appears to not include the part number on the moulded cord....
Now if we had a modicum of science in the whole discussion, instead an argument of purely religeuous systems of belief all such practices as the one above would disappear, not because people are inherently honest, but because under more rational conditions there would not be any need for subterfuge.
Sayonara
I cannot believe this thread got to 16 pages with a statement like this in the 1st post.
Ignorance must surely be bliss.
On this site some people are so occupied calling others ignorant that they never become aware of their own ignorance...
have you ever tested it in a truely unbiased way?
The mains cable could make a difference, if you were willing to go so far as to rip all the lines from your house, replace them, and have a fancy custom generator set built for your powerplant.
Good idea. I didn't go that far; just built a 50Hz sine wave generator to power my phono pre with cleanish AC. Sadly, it didn't quite work. The regenerator was still powered from the wall and guess what, the AC power cable was just as audible as before.
Then again, many here are unable to hear any difference between 'competently designed' (ha, ha) amps or anything else for that matter, so i guess we simply inhabit different planets as far as audio is concerned.
Just wondering: from all the wonderful hobbies in the world, why choose the most unsuitable?
Konnichiwa,
Actually, you don't need to do ANY of this. The noiseloops discussed are very simple and end, depending on equipment configuration, level and frequency at the nearest common earth, the nearest "Y" capacitor between N & PEN or at the very worst case the point where the house's main incomming neutral/earth are joined to the radiating oput PEN (Protected Earth Neutral[ised]) and N (Neutral) conductors are derived from.
Any of the wiring past that plays it's main role as moderatly effective aerial system (it is mosyly burried in wet ground armoured in steel sheathing.
In fact if you use any form of mains conditioner with any significant capacitance shunt elements (eg ANY OF THEM) the main noise loops are quite reliably confined to the cables after the conditioner.
All in it is a rather complex subject, but highly ameanable to measurements and deductive reasoning. No Voodoo, nothing esotheric involved and no relacement of "all wires to the generator and having a custom generator build" is required.
The custom generator may help by avoiding the introduction of a heavily distorted sindewave which usually leads to effective DC offset on the line and tends to saturate all these fancy "High End" torroidal mains transformer cores, but that is another part of engineering.
Sayonara
Tweeker said:
Actually, you don't need to do ANY of this. The noiseloops discussed are very simple and end, depending on equipment configuration, level and frequency at the nearest common earth, the nearest "Y" capacitor between N & PEN or at the very worst case the point where the house's main incomming neutral/earth are joined to the radiating oput PEN (Protected Earth Neutral[ised]) and N (Neutral) conductors are derived from.
Any of the wiring past that plays it's main role as moderatly effective aerial system (it is mosyly burried in wet ground armoured in steel sheathing.
In fact if you use any form of mains conditioner with any significant capacitance shunt elements (eg ANY OF THEM) the main noise loops are quite reliably confined to the cables after the conditioner.
All in it is a rather complex subject, but highly ameanable to measurements and deductive reasoning. No Voodoo, nothing esotheric involved and no relacement of "all wires to the generator and having a custom generator build" is required.
The custom generator may help by avoiding the introduction of a heavily distorted sindewave which usually leads to effective DC offset on the line and tends to saturate all these fancy "High End" torroidal mains transformer cores, but that is another part of engineering.
Sayonara
Kuei Yang Wang said:
I am not really sure what you want to say...
Do you mean that using a special powercord can actually make the sound better?
An acquaintance of mine makes his own cables. He hass a small business in making high-end cables. He also makes powercords. He once tried to convince me that his cables make the sound better. Ultimately it even came down to this: his cables could make the sound better, even when you use an extra mains supply unit! The degraded "signal" (power) could be made better with his cable.
KYW, you say that special powercords are not needed throughout the house. Then do you also agree that they can make the power better?
- Status
- Not open for further replies.