john curl said:
Prof. Ed Cherry has backed Bob's view of PIMD, and concluded that "the only nonlinearity which is likely to contribute to PIMD in a typical amplifier is the collector-base capacitance of the second-stage transistor."
He also cited boundary conditions in which a non-feedback amp can have PIMD, and one in which revoval of feedback converts AIMD into PIMD.
Cherry, "Amplitude and Phase of Intermodulation Distortion", JAES, May 1983.
Martin Colloms' Article
Martin Colloms’ article makes for interesting reading and provides a lot of food for thought.
It was interesting to see him opening with reference to a blind listening test he did in the mid 1970’s, as those seem to be anathema to many today. I’d like to see more controlled listening tests. I like the idea that he threw in a tube amp as a ringer, and it was interesting to see that its perceived sound quality was higher. His observation that “the association between measured performance and [perceived] sound quality is uncertain” seems reasonable to me as long as you add the word “perceived” as I did to his sentence. Of course, we also have the variable that solid state amplifiers were not necessarily all that good in the mid-seventies as well. With all of these variables, it seems a bit of a stretch to attribute the better perceived sound of the 25 wpc tube amplifier to its coincidental use of less negative feedback.
Most no-feedback and tube power amplifiers have enough output impedance to cause a significant frequency response dependence on the speaker load, and I think many dismiss this variable too quickly. A recently-reviewed high-end loudspeaker had an impedance swing from 3 ohms to a peak of 40 ohms in the midrange. Connect this to an amplifier with an output impedance of 0.4 ohms (corresponding to a damping factor of 20) and you get a peak-to-peak frequency response variation of over 1 dB. It is amusing to see people look for and speculate on un-measurable consequences of negative feedback while ignoring such an obvious variable.
It is not inconceivable to me that some loudspeakers sound better with a little bit of impedance in series with them (indeed, note that some people advocate current drive of loudspeakers). In this regard, I found his last sentence to be quite on-target: “And if speakers were designed to have smoother impedance curves, to be relatively uncritical of amplifier or cable matching, and to offer higher sensitivity, we would have greater freedom to examine the feedback question and the validity of the low-power, short-path, zero-feedback approach to power amplification.
The speculation that low open-loop bandwidth causes NFB to somehow darken the sound is seductively intuitive. Colloms properly points out that low open loop bandwidth is usually associated with high NFB, although this need not always be the case. After all, we are filtering the signal in the forward path to a frequency of perhaps less than 1 kHz and then “fixing up” the frequency response with the NFB. Kind of makes sense, doesn’t it? Maybe there is something to this that we don’t know about, but we do know that it is not something that has been technically postulated or measured. What I mean by this is that TIM and PIM, for example, have been blamed on low open-loop bandwidth, but these technical explanations have been shown to be wrong both in theory and by measurement of real amplifiers. If there is something here, it is something else that we have not yet gotten our arms around. If it can be measured, I don’t think we’ve found the measurement that reveals it yet.
One thing worth observing is that in amplifiers with low open loop bandwidth, the forward-propagating drive signal emanating from the input stage is usually more like the derivative of the program signal than the program signal itself, and its phase relationship may differ by nearly 90 degrees from what it would be if the open loop bandwidth were wide. Bear in mind that in a properly designed NFB amplifier, this does not mean that the leading edge peaks are larger, but rather that the later, lower-frequency portions of the error signal are smaller.
Put another way, the relationship between the program signal and feedback signal as they mix together in the forward path is different at certain locations, depending on whether the open loop bandwidth is wide or narrow. While the effects of this seem not to be measurable for the amplifier as a whole, it is a difference that is going on inside. It is also interesting to note that the signal coming off of an LP is largely in the differentiated domain due to the RIAA equalization.
Cheers,
Bob
Martin Colloms’ article makes for interesting reading and provides a lot of food for thought.
It was interesting to see him opening with reference to a blind listening test he did in the mid 1970’s, as those seem to be anathema to many today. I’d like to see more controlled listening tests. I like the idea that he threw in a tube amp as a ringer, and it was interesting to see that its perceived sound quality was higher. His observation that “the association between measured performance and [perceived] sound quality is uncertain” seems reasonable to me as long as you add the word “perceived” as I did to his sentence. Of course, we also have the variable that solid state amplifiers were not necessarily all that good in the mid-seventies as well. With all of these variables, it seems a bit of a stretch to attribute the better perceived sound of the 25 wpc tube amplifier to its coincidental use of less negative feedback.
Most no-feedback and tube power amplifiers have enough output impedance to cause a significant frequency response dependence on the speaker load, and I think many dismiss this variable too quickly. A recently-reviewed high-end loudspeaker had an impedance swing from 3 ohms to a peak of 40 ohms in the midrange. Connect this to an amplifier with an output impedance of 0.4 ohms (corresponding to a damping factor of 20) and you get a peak-to-peak frequency response variation of over 1 dB. It is amusing to see people look for and speculate on un-measurable consequences of negative feedback while ignoring such an obvious variable.
It is not inconceivable to me that some loudspeakers sound better with a little bit of impedance in series with them (indeed, note that some people advocate current drive of loudspeakers). In this regard, I found his last sentence to be quite on-target: “And if speakers were designed to have smoother impedance curves, to be relatively uncritical of amplifier or cable matching, and to offer higher sensitivity, we would have greater freedom to examine the feedback question and the validity of the low-power, short-path, zero-feedback approach to power amplification.
The speculation that low open-loop bandwidth causes NFB to somehow darken the sound is seductively intuitive. Colloms properly points out that low open loop bandwidth is usually associated with high NFB, although this need not always be the case. After all, we are filtering the signal in the forward path to a frequency of perhaps less than 1 kHz and then “fixing up” the frequency response with the NFB. Kind of makes sense, doesn’t it? Maybe there is something to this that we don’t know about, but we do know that it is not something that has been technically postulated or measured. What I mean by this is that TIM and PIM, for example, have been blamed on low open-loop bandwidth, but these technical explanations have been shown to be wrong both in theory and by measurement of real amplifiers. If there is something here, it is something else that we have not yet gotten our arms around. If it can be measured, I don’t think we’ve found the measurement that reveals it yet.
One thing worth observing is that in amplifiers with low open loop bandwidth, the forward-propagating drive signal emanating from the input stage is usually more like the derivative of the program signal than the program signal itself, and its phase relationship may differ by nearly 90 degrees from what it would be if the open loop bandwidth were wide. Bear in mind that in a properly designed NFB amplifier, this does not mean that the leading edge peaks are larger, but rather that the later, lower-frequency portions of the error signal are smaller.
Put another way, the relationship between the program signal and feedback signal as they mix together in the forward path is different at certain locations, depending on whether the open loop bandwidth is wide or narrow. While the effects of this seem not to be measurable for the amplifier as a whole, it is a difference that is going on inside. It is also interesting to note that the signal coming off of an LP is largely in the differentiated domain due to the RIAA equalization.
Cheers,
Bob
MC didn't bother to determine if "better" meant "output more closely resembling input" or "caused EQ changes that we liked." Or even, "fattened the sound by adding second harmonic distortion." But determining that would risk undermining the raison d'etre of the people who paid him.
Could the tube amp have been distinguished from the others in that test if levels were matched and EQ was adjusted? I know where I'd bet...
Could the tube amp have been distinguished from the others in that test if levels were matched and EQ was adjusted? I know where I'd bet...
SY said:
I was wondering about the level matching issue as well. The logistics of matching levels is quite a mess for 18 amplifiers. I somehow doubt that he built an 18-position transfer switch with level control in each leg. Did the one that was perceived as best simply have the highest gain?
I was at Bob's presentation at RMAF 2006, and it was very hard to tell the tube and solid-state amps apart in the level-matched setup he built. I couldn't reach any sort of conclusive result. I'm thinking that maybe, given some high-quality recordings that were also very repetitive, and lots of time, I might be able to do it.
Since these guys are in the business of selling advertising, it's pretty hard to not take what they say with a grain of salt.
TIM, PIM and IIM
Here is a brief summary of TIM, PIM and IIM distortion mechanisms that some may find useful.
TIM
Transient Intermodulation distortion (TIM) is a high frequency distortion brought on by the rate of change of the signal. So-called “hard” TIM is synonymous with slew rate limiting. Soft TIM is associated with the onset of slew rate limiting, but is indistinguishable from other forms of high-frequency nonlinearity. As such, THD-20 is usually very effective in measuring it. Indeed, in most cases it is not possible for an amplifier to create TIM without creating substantial amounts of THD-20. TIM is usually created by input stage overload under transient conditions. Negative feedback has wrongly been blamed for TIM. However, it is fair to say that designs with no NFB force the design of an input stage with high dynamic range (it has to handle the full amplitude of the input signal) and the consequent inability to create TIM due to transient error signal overload. Bear in mind that the error signal is usually a differentiated version of the input signal, but its peak is no larger than the input signal itself; with NFB, the lower frequencies are SMALLER, and the input stage is less taxed.
PIM
Phase Intermodulation distortion (PIM) results from the change in signal phase as a function of signal amplitude. Barrie Gilbert was technically correct in his paper as far as it went, but he oversimplified the situation. He studied the conversion of Amplitude Intermodulation distortion (AIM) to PIM via NFB. PIM is simply related to the change in the amplifier closed loop rolloff pole with signal, regardless of the underlying cause. If there is AIM compression, open loop gain decreases, the closed loop pole comes in, and there is more in-band phase shift due to that pole. PIM will thus result. However, in such a case, the amplifier will exhibit quite high levels of conventional IM (AIM) as well.
Real amplifiers have other sources of PIM in the open loop that, like other distortions, are reduced by NFB. As such, amplifiers without any negative feedback can exhibit significant PIM.
IIM
Interface Intermodulation Distortion (IIM) is created when the output impedance of the power amplifier changes as a function of signal current. It does not matter how the output impedance is established or what mechanism causes it to change with signal.
IIM was originally blamed on amplifiers that relied on negative feedback to establish a low closed-loop output impedance, but even amplifiers without negative feedback exhibit IIM, since their output impedance can change as a function of signal due to output stage crossover nonlinearities or output transistor beta variations, just to name two possible causes. In many cases, the use of negative feedback actually reduces IIM. For a given amount of negative feedback, it is true that a lower open loop output impedance will produce a better result.
All of these distortions can and have been measured. Although negative feedback has been blamed for these distortions, this generalization is wrong. It is easy to disprove a generalization with a counter-example. The MOSFET power amplifier I published uses large amounts of negative feedback and has a fairly low open loop bandwidth. Yet it has vanishingly small amounts of all of these distortions. These distortions are real, and many amplifiers have them, but it is not the fault of negative feedback but rather the particular design.
Cheers,
Bob
Here is a brief summary of TIM, PIM and IIM distortion mechanisms that some may find useful.
TIM
Transient Intermodulation distortion (TIM) is a high frequency distortion brought on by the rate of change of the signal. So-called “hard” TIM is synonymous with slew rate limiting. Soft TIM is associated with the onset of slew rate limiting, but is indistinguishable from other forms of high-frequency nonlinearity. As such, THD-20 is usually very effective in measuring it. Indeed, in most cases it is not possible for an amplifier to create TIM without creating substantial amounts of THD-20. TIM is usually created by input stage overload under transient conditions. Negative feedback has wrongly been blamed for TIM. However, it is fair to say that designs with no NFB force the design of an input stage with high dynamic range (it has to handle the full amplitude of the input signal) and the consequent inability to create TIM due to transient error signal overload. Bear in mind that the error signal is usually a differentiated version of the input signal, but its peak is no larger than the input signal itself; with NFB, the lower frequencies are SMALLER, and the input stage is less taxed.
PIM
Phase Intermodulation distortion (PIM) results from the change in signal phase as a function of signal amplitude. Barrie Gilbert was technically correct in his paper as far as it went, but he oversimplified the situation. He studied the conversion of Amplitude Intermodulation distortion (AIM) to PIM via NFB. PIM is simply related to the change in the amplifier closed loop rolloff pole with signal, regardless of the underlying cause. If there is AIM compression, open loop gain decreases, the closed loop pole comes in, and there is more in-band phase shift due to that pole. PIM will thus result. However, in such a case, the amplifier will exhibit quite high levels of conventional IM (AIM) as well.
Real amplifiers have other sources of PIM in the open loop that, like other distortions, are reduced by NFB. As such, amplifiers without any negative feedback can exhibit significant PIM.
IIM
Interface Intermodulation Distortion (IIM) is created when the output impedance of the power amplifier changes as a function of signal current. It does not matter how the output impedance is established or what mechanism causes it to change with signal.
IIM was originally blamed on amplifiers that relied on negative feedback to establish a low closed-loop output impedance, but even amplifiers without negative feedback exhibit IIM, since their output impedance can change as a function of signal due to output stage crossover nonlinearities or output transistor beta variations, just to name two possible causes. In many cases, the use of negative feedback actually reduces IIM. For a given amount of negative feedback, it is true that a lower open loop output impedance will produce a better result.
All of these distortions can and have been measured. Although negative feedback has been blamed for these distortions, this generalization is wrong. It is easy to disprove a generalization with a counter-example. The MOSFET power amplifier I published uses large amounts of negative feedback and has a fairly low open loop bandwidth. Yet it has vanishingly small amounts of all of these distortions. These distortions are real, and many amplifiers have them, but it is not the fault of negative feedback but rather the particular design.
Cheers,
Bob
SY said:
In the Cary example alone he spoke of interactively turning down the negative feedback, it's hard to imagine doing this with no gain change. Speaking of Cary, it's an example I have used before. It has ample amounts of any kind of distortion you care to pick, so discussing it in the same context as .01% vs. .0001% of this or that distortion is kind of pointless.
On second harmonics, I once was at an Absolute Sound listening session where people were cooing over some Javanese gamelan through a 60W tube amp and power hungry speakers, even though the distortion was almost at the level of the fundamental. The gamelan made it so easy to recognize this I didn't need any meters.
john curl said:
Hi John, I guess you are still referring to that long letter to the editor of audio that you guys sent in that never got published. I had forgotten about that letter and have never been able to find it in my attic, and you were kind enough to send me a copy of my own fairly detailed response to the letter that I wrote back to you guys.
I seem to recall in one of the other threads that someone here agreed to put up both letters on a site if we would supply them. I thought that was a good idea and a fair way to put the whole thing to rest. Since you are a signatory to that letter, and I was a recipient, I don't see why you cannot just put it up. I'll be happy to supply my response, even absent your original letter (it would probably make a little more sense in the context of your original, however). I'd really be interested in seeing your letter again.
I'm perfectly happy to forget about the whole thing and not do anything. However, in the spirit of fairness, if you are not willing to share it, I think you should stop talking about a letter whose contents you are unwilling to reveal. I don't think you rebutted me at all, but the fair thing to do is to let others see the exchange for themselves.
In any case, for anyone here who is interested in seeing my reply to that letter, I will happily provide it by email if you email me a request for it at bob@cordellaudio.com.
Cheers,
Bob
andy_c said:
Are you saying you didn't read it all the way through to find out?I didn't see exactly that statement in there, but I did see something similar.
Andy, I meant the idea that with feedback you have a delayed version of the input going back into the input being delayed again, you get the point. These 'intuitve' arguements aimed at explaining something technical to lay people are dangerous. Next they point out that with no feedback you don't have these 'echos'. MC didn't say this exactly but the inference of the fedback signal being a delayed version of the input was there.
BTW in the same vein I think calling DA a 'memory' effect by Walt and others was (without clarification) an unfortunate choice of words.
scott wurcer said:
I had to look up "Droste", but as soon as I did, I realized what you were saying. His feedback analogy was something along the lines of the mirrors on opposite walls of a barber shop, giving what appears to be reflections going on to infinity.
For some reason, this reminds me of a radio show I used to listen to, called The Old Time Radio Show. They played old programs like The Shadow. At the beginning of every show, they played a skit by Fibber McGee and Molly. In this skit, Fibber McGee claims that the radio waves of old broadcasts are "still bouncing around out there", and that because of this, you could "not only tune your radio to different frequencies, but also different times". This was the perfect introduction to the show! Fibber McGee's somewhat plausible but totally wrong description reminds me of some of these explanations by audio reviewers to the lay readers. The reviewer often doesn't understand himself what he's trying to explain, so it becomes the blind leading the blind.
Folks,
I've just learned that my mailbox on my website is full; there may be a problem as the host just recently changed some of their software. Please email me at audiohead7@yahoo.com. I apologize for the inconvenience.
Cheers,
Bob
I've just learned that my mailbox on my website is full; there may be a problem as the host just recently changed some of their software. Please email me at audiohead7@yahoo.com. I apologize for the inconvenience.
Cheers,
Bob
Re: Martin Colloms' Article
Mr Colloms article referenfed here brings again to the limelight a recurring issue reviewers usually give ample coverage, and that feeds them with limitless inspiration.
Toghether with designers equally inclined to seek those features, they talk about flat and uninvolving renderings from such or that amplifier or technology, against rich and emotional experiences provided by different ones.
This argument is suficiently powerfull to lead skilled, strongly engineering oriented people like Bob to allow for an "X factor" as something to recon.
Not discounting its true existence - that is, the real existence of subtle phenomena not fully understood - some sanity checks are in order if for no other purpose at least to organize the search.
I have often attended live performances which for different reasons were flat and uninvolving. In fact genuinely rich and emotional performances are more the exception than the rule. Sure it may be said that outstanding performances by outstanding artists must belong to this category, but has somebody had the chance to witness the recording session live and compare it with the final product?
In fact, I can discriminate fairly easily good performances of fine works no matter how bad the reproduction chain is, something probably most will agree they can do also. Most of us are here because we have a taste for music after all.
What I mean is one should not take at face value the "involving" feature, for it may eventually lead to unknowingly prefer a reproduction system that in fact is transforming an original source in a different end product under uncontroled conditions and probably that is not what was intended.
There are broadly 2 camps here. On the one hand those who affirm there are hidden components in music that are not measurable and will never be.
On the other side, still another possible classification, those who belive there are "X factors" out there we still have not learned to measure, and those who discount their existence and attribute all this fuss to self deception, market forces or whatever.
For the first main group, there is little hope for agreement as much as there is little hope for an agnostic to "unconvert" a believer, be it religion, afterlife, alien abductions or whatever. Since there is no way to falsify a belief, there is no rational road to disprove an act of faith.
At most it should be noted that if that were the case - that there are hidden components in music that are not measurable and never will be - then it should be the only realm in technology where this happens, being the frontiers of the unknown by now both at the Planck scale, or before the first millisecond after the big bang, which is the kind of sanity check I was talking about. No space for a major force like electromagnetism to be discovered, rest assured.
Are there features we have not learned to characterize and measure?
Wish I knew. There is strong evidence we have both the theoretical tools to handle audio signals, and fine instrumentation to dissect them. But then which are the thresholds of audibility for different aspects of sound?
We can isolate absolute sensitivity thresholds, and have learned a lot about amplitude and time masking, yet we are fully aware we can easily spot features of a sound source which are really hard to extract by measurement, particularly with respect to location, something that has much to do with reverberation and binaural perception, so in the end it is probably a safe bet to admit there is more to learn about measuring an amplifier performance before deciding whether it is good or not.
Sorry for the long post
Rodolfo
Bob Cordell said:
Mr Colloms article referenfed here brings again to the limelight a recurring issue reviewers usually give ample coverage, and that feeds them with limitless inspiration.
Toghether with designers equally inclined to seek those features, they talk about flat and uninvolving renderings from such or that amplifier or technology, against rich and emotional experiences provided by different ones.
This argument is suficiently powerfull to lead skilled, strongly engineering oriented people like Bob to allow for an "X factor" as something to recon.
Not discounting its true existence - that is, the real existence of subtle phenomena not fully understood - some sanity checks are in order if for no other purpose at least to organize the search.
I have often attended live performances which for different reasons were flat and uninvolving. In fact genuinely rich and emotional performances are more the exception than the rule. Sure it may be said that outstanding performances by outstanding artists must belong to this category, but has somebody had the chance to witness the recording session live and compare it with the final product?
In fact, I can discriminate fairly easily good performances of fine works no matter how bad the reproduction chain is, something probably most will agree they can do also. Most of us are here because we have a taste for music after all.
What I mean is one should not take at face value the "involving" feature, for it may eventually lead to unknowingly prefer a reproduction system that in fact is transforming an original source in a different end product under uncontroled conditions and probably that is not what was intended.
There are broadly 2 camps here. On the one hand those who affirm there are hidden components in music that are not measurable and will never be.
On the other side, still another possible classification, those who belive there are "X factors" out there we still have not learned to measure, and those who discount their existence and attribute all this fuss to self deception, market forces or whatever.
For the first main group, there is little hope for agreement as much as there is little hope for an agnostic to "unconvert" a believer, be it religion, afterlife, alien abductions or whatever. Since there is no way to falsify a belief, there is no rational road to disprove an act of faith.
At most it should be noted that if that were the case - that there are hidden components in music that are not measurable and never will be - then it should be the only realm in technology where this happens, being the frontiers of the unknown by now both at the Planck scale, or before the first millisecond after the big bang, which is the kind of sanity check I was talking about. No space for a major force like electromagnetism to be discovered, rest assured.
Are there features we have not learned to characterize and measure?
Wish I knew. There is strong evidence we have both the theoretical tools to handle audio signals, and fine instrumentation to dissect them. But then which are the thresholds of audibility for different aspects of sound?
We can isolate absolute sensitivity thresholds, and have learned a lot about amplitude and time masking, yet we are fully aware we can easily spot features of a sound source which are really hard to extract by measurement, particularly with respect to location, something that has much to do with reverberation and binaural perception, so in the end it is probably a safe bet to admit there is more to learn about measuring an amplifier performance before deciding whether it is good or not.
Sorry for the long post
Rodolfo
Re: Re: Martin Colloms' Article
Not at all, a pleasure to read!
Two comments, if I may.
I think we accept the existence of audible differences too easy, too uncritical. Of course there are audible differences between audio reproduction systems. Differences, that a statistically significant proportion of listeners will confirm. And such differences should give us enough reason to research what is going on. But too often we accept any report, however poorly backed-up, as yet another proof of audible differences. That can only confuse the issues and send us on a wild goose chase. Not all 'reports' on audible differences are equal; we need some method to weed out the anecdotes and identify the genuine valid ones.
Then the comparison to 'original' or 'live' performances. During a large part of the hi-fi history that was a valid goal: if only we could reach the frequency response to reproduce all frequencies in a live performance equally! I would humbly suggest that we left this behind us decades ago. Recording and reproducing music has become a form of art in itself. It is impossible to reproduce an opera or a hard-rock live concert in any meaningfull way. Pro live: atmosphere, being there, the beautiful venue, seeing your admired artists in the flesh, and all these things that give you a feeling of something special. Pro recorded/reproduced: MUCH better sound stage, better detail, better balanced instruments, no coughing, no people walking in front of you.
We should get away from living in the past and insisting on trying to copy live performances. There is a work of art on that CD or DVD. Our endeavour is to try to reproduce it as faithfull as possible. Add nothing, take nothing away.
Jan Didden
ingrast said:
Not at all, a pleasure to read!
Two comments, if I may.
I think we accept the existence of audible differences too easy, too uncritical. Of course there are audible differences between audio reproduction systems. Differences, that a statistically significant proportion of listeners will confirm. And such differences should give us enough reason to research what is going on. But too often we accept any report, however poorly backed-up, as yet another proof of audible differences. That can only confuse the issues and send us on a wild goose chase. Not all 'reports' on audible differences are equal; we need some method to weed out the anecdotes and identify the genuine valid ones.
Then the comparison to 'original' or 'live' performances. During a large part of the hi-fi history that was a valid goal: if only we could reach the frequency response to reproduce all frequencies in a live performance equally! I would humbly suggest that we left this behind us decades ago. Recording and reproducing music has become a form of art in itself. It is impossible to reproduce an opera or a hard-rock live concert in any meaningfull way. Pro live: atmosphere, being there, the beautiful venue, seeing your admired artists in the flesh, and all these things that give you a feeling of something special. Pro recorded/reproduced: MUCH better sound stage, better detail, better balanced instruments, no coughing, no people walking in front of you.
We should get away from living in the past and insisting on trying to copy live performances. There is a work of art on that CD or DVD. Our endeavour is to try to reproduce it as faithfull as possible. Add nothing, take nothing away.
Jan Didden
Re: Re: Re: Martin Colloms' Article
Jan, you have brought here a pivotal issue on which I fully agree.
Then, recognizing this, whe should therefore better define both the role and the means to gauge the performance of the amplifier - or more broadly the end reproduction system designer.
Are we to seek for a satisfying listening experience based on entirely subjective aspects and then for which particular listener group - there are many and varied - or are we to define objective standards and disregard whatever issues not possible to measure?
Rodolfo
janneman said:
Jan, you have brought here a pivotal issue on which I fully agree.
Then, recognizing this, whe should therefore better define both the role and the means to gauge the performance of the amplifier - or more broadly the end reproduction system designer.
Are we to seek for a satisfying listening experience based on entirely subjective aspects and then for which particular listener group - there are many and varied - or are we to define objective standards and disregard whatever issues not possible to measure?
Rodolfo
Re: Re: Re: Re: Martin Colloms' Article
Rodolfo,
I haven't heard anybody say that the ideal of an amp that is transparent, in that it only transmits whatever gets to its input (but amplified of course) is not valid. What the differences are about is how to make sure that is the case, what measurements to do to verify that. Again and again we hear that measurements cannot discern audible differences. But what audible differences?
It would be a huge step forward if we would take this business - and ourselves - serious. As long as we read: "I swapped zero fb amp for a global fb amp, the difference was huge, so fb must be really bad", or "amp xx had WBT connectors, amp yy only had Neutriks and sounded really bad, so Neutriks clearly destroy the sound field", we are not taking ourselves serious. We are making a fool of ourselves if there ever was one.
There are precious few serious investigations and tests of whether amps really sound different. And you can only do that if you concentrate on the sound, and nothing else. That leads inexorably to (double) blind testing, how abhorrent the idea may be to some.
Not that I have hopes for a change of minds; too much hinges on, too much money is involved in the current state of audio. Except for a few excentric fools nobody in the industry is interested to know what differences are audible, under what conditions, anyway.
Jan Didden
ingrast said:
Rodolfo,
I haven't heard anybody say that the ideal of an amp that is transparent, in that it only transmits whatever gets to its input (but amplified of course) is not valid. What the differences are about is how to make sure that is the case, what measurements to do to verify that. Again and again we hear that measurements cannot discern audible differences. But what audible differences?
It would be a huge step forward if we would take this business - and ourselves - serious. As long as we read: "I swapped zero fb amp for a global fb amp, the difference was huge, so fb must be really bad", or "amp xx had WBT connectors, amp yy only had Neutriks and sounded really bad, so Neutriks clearly destroy the sound field", we are not taking ourselves serious. We are making a fool of ourselves if there ever was one.
There are precious few serious investigations and tests of whether amps really sound different. And you can only do that if you concentrate on the sound, and nothing else. That leads inexorably to (double) blind testing, how abhorrent the idea may be to some.
Not that I have hopes for a change of minds; too much hinges on, too much money is involved in the current state of audio. Except for a few excentric fools
nobody in the industry is interested to know what differences are audible, under what conditions, anyway.Jan Didden