A Bright Shining Lie…Why ignoring an inconvenient truth is stifling system performance a blog from Roy Gregory

[microstrip]

“Greatest fool of all is the one who fools himself.”

A pity that any interesting discussion in WBF soon turns in a bloody fight. In this case both parts have good arguments, but IMHO the intrinsic allowance and even encouragement for a belligerent style by the forum is systematically killing very challenging subjects.

Anyone trying to contribute will immediately find that he is entering a mined field.

 

[tima]

Sure, I will connect the dots and make it clear. Let’s focus in the area that you have sited above for further clarification, as it should be rather obvious:

"while ... might compensate ...." if Mr. Gregory repositions the speakers to reinforce or attenuate the bass performance of his system with the new component in a new ”complementary“ arrangement, with new countributions (both constructive and destructive) from his acoustical space boundaries, the new or improved bass performance of his system are strictly the results of the speaker repositioning, in other words the loudspeaker/room’s interaction, and not an inherent quality of the new component under review.

"this does not benefit the readers at all and will lead to wrong conclusions." what Mr. Gregory is conveying to the readers is unique to his system, as the readers do not share his speakers and acoustical space. Of course, no reviewer shares the same acoustical space, and likely speakers, with the readers but with the traditional A/B comparisons, which Mr. Gregory called a stupid approach, the speaker room interaction is kept common (the same) between component A and B to allow the reviewer to make assessments on the relative qualities between components A and B under the same set of conditions. By imbedding two sets of boundary conditions, one optimized for components A and the other optimized for component B, the comparison and assessment is no longer normalized. Going beyond that, if you take the original component out of the picture all together, Mr. Gregory is attributing contribution from boundary reinforcement (both constructive and destructive)speaker/room interaction, which are unique to Mr. Gregory’s system environment, and passing them off as what can be achieved with the component under review, which the reader can be mislead to interpret as inherent qualities of the component under review and evaluation.

I honestly do not understand why the above is so hard to grasp.

Carlos, thanks very much for taking the time to write your follow-up. I understand better what you are saying. I am out of time and will come back to explore this further. But two points in brief:

To your first paragraph: I think we need to give up the idea of a component having an inherent quality, or at least the idea that we can know it.

To your second: I agree that under RG's approach the conditions for assessment are not strictly normalized. I wonder if we should consider such normalization as sacrosanct, or put differently, is there, or should there be, a standardized methodology of assessment? How would we come to that and what value are we after for review readers? At the moment I see two different approaches.

 

[ddk]

Sure, I will connect the dots and make it clear. Let’s focus in the area that you have sited above for further clarification, as it should be rather obvious:

"while ... might compensate ...." if Mr. Gregory repositions the speakers to reinforce or attenuate the bass performance of his system with the new component in a new ”complementary“ arrangement, with new countributions (both constructive and destructive) from his acoustical space boundaries, the new or improved bass performance of his system are strictly the results of the speaker repositioning, in other words the loudspeaker/room’s interaction, and not an inherent quality of the new component under review.

This is your assumption but IME not necessarily true in practice.

"this does not benefit the readers at all and will lead to wrong conclusions." what Mr. Gregory is conveying to the readers is unique to his system, as the readers do not share his speakers and acoustical space. Of course, no reviewer shares the same acoustical space, and likely speakers, with the readers but with the traditional A/B comparisons, which Mr. Gregory called a stupid approach, the speaker room interaction is kept common (the same) between component A and B to allow the reviewer to make assessments on the relative qualities between components A and B under the same set of conditions. By imbedding two sets of boundary conditions, one optimized for components A and the other optimized for component B, the comparison and assessment is no longer normalized.

Your assumption is different speaker positions for each amp isn’t always the case, you neglect the fact that the initial setup can be flawed. A higher resolution component can highlight setup flaws previously not noticed and the new speaker location can be the right one in the room for any amp and all comparisons can be done from the new location. Another reason for different speaker location can be the design of the amplifiers, many have an artificial boost somewhere in the bass or lower midrange a more linear amp certainly needs a modified setup.

Going beyond that, if you take the original component out of the picture all together, Mr. Gregory is attributing contribution from boundary reinforcement (both constructive and destructive)speaker/room interaction, which are unique to Mr. Gregory’s system environment, and passing them off as what can be achieved with the component under review, which the reader can be mislead to interpret as inherent qualities of the component under review and evaluation.

Proper setup isn’t unique to @RoyGregory and his space it’s the cornerstone of every system. How can one do justice to any component no less at a high level in a suboptimal setup? Everything can be misinterpreted sometimes a minimum degree of sophistication is required from the reader to understand the process and the author; this doesn’t absolve the reviewer if he/she lacks the experience!

david

 

[Carlos269]

This is your assumption but IME not necessarily true in practice.

Your assumption is different speaker positions for each amp isn’t always the case, you neglect the fact that the initial setup can be flawed. A higher resolution component can highlight setup flaws previously not noticed and the new speaker location can be the right one in the room for any amp and all comparisons can be done from the new location. Another reason for different speaker location can be the design of the amplifiers, many have an artificial boost somewhere in the bass or lower midrange a more linear amp certainly needs a modified setup.

Proper setup isn’t unique to @RoyGregory and his space it’s the cornerstone of every system. How can one do justice to any component no less at a high level in a suboptimal setup? Everything can be misinterpreted sometimes a minimum degree of sophistication is required from the reader to understand the process and the author; this doesn’t absolve the reviewer if he/she lacks the experience!

david

You are going to have to do better than that and explain yourself as you are throwing out hypothetical mombo-jumbo without being diligent in providing an explanation. Let’s review:

1, Why is my assumption not necessarily true in practice?

2. How do you know that a superior amp has highlighted flaws in the original set up?

3. How can you certify that you have achieved a proper set up?

Curious to see how you answer as I will take you into deep water to see if you sink or swim with your responses.

 

[ddk]

You are going to have to do better than that and explain yourself as you are throwing out hypothetical mombo-jumbo without being diligent in providing an explanation.

As I mentioned, a certain degree of sophistication is required from the reader to understand the author, your hypothetical mumbo jumbo remark shows a lack of it!

Let’s review:

1, Why is my assumption not necessarily true in practice?

Because you ignore the contribution of other components and attribute improved/different bass quality solely to new speaker position. You’re not in the same room with @RoyGregory in that moment to know what caused what? I can take this a lot further but this is enough.

2. How do you that a superior amp has highlighted flaws in the original set up?

How do you know that it hasn’t?

3. How can you certify that you have achieved a proper set up?

I know from my years of experience same as I can detect suboptimal setup but we’re not talking about me here. It’s a known fact that many if not most systems’ speaker aren’t optimized.

Curious to see how you answer as I will take you into deep water to see if you sink or swim with your responses.

You’re swimming alone in your own waters Carlos.

david

 

[Carlos269]

Because you ignore the contribution of other components and attribute improved/different bass quality solely to new speaker position. You’re not in the same room with @RoyGregory in that moment to know what caused what? I can take this a lot further but this is enough.

How do you know that it hasn’t?

I know from my years of experience same as I can detect suboptimal setup but we’re not talking about me here. It’s a known fact that many if not most systems’ speaker aren’t optimized.

You’re swimming alone in your own waters Carlos.

david

1. Because the other components have not changed. And 2. Because as I mentioned before, the speaker/room interaction is the dominant factor.

if a component is designed competently, to be linear across the frequency spectrum then it should not noticeably impact the speaker/room interaction. That’s how I know that you cannot use quality components to highlight flaws with speaker placement in the room.

If experience was all it took to identify proper set up then we would have more agreement on what it constitutes.

I am swimming in my own waters, that we can agree on.

So let me boil down this whole discussion down to its simplest form for those of you that still do not get it:

In essence the arguments comes down to this;

I state that a known reference is needed to provide context, validity, and to form a basis for a component’s evaluation and assessment.

Mr. Gregory does not believe in or exercises the reference system approach as part of his equipment evaluations.

There it is in a nutshell, you pick a side and live with it, because I’m done explaining.

 

[ddk]

1. Because the other components have not changed. And 2. Because as I mentioned before, the speaker/room interaction is the dominant factor.

It’s a factor, the rest including single component change are your assumptions there can easily be other changes that went unnoticed. Also why is the conversation limited to a single event?

if a component is designed competently, to be linear across the frequency spectrum then it should not noticeably impact the speaker/room interaction.

Many are competently designed to be nonlinear, a small bass or midrange boost can be attractive in the right context!

That’s how I know that you cannot use quality components to o highlight speaker placement in the room.

Really?

If experience was all it took to identify proper set up then we would have more agreement on what it constitutes.

I am swimming in my own waters, that we can agree on.

So let me boil down this whole discussion down to its simplest form for those of you that still do not get it:

In essence the arguments comes down to this;

I state that a known reference is needed to provide context, validity, and to form a basis for a component’s evaluation and assessment. Mr. Gregory does not believe in or exercises the reference system approach as part of his equipment evaluations.

Known reference for whom? What if the reference is flawed? Do we all agree on the same reference? It’s never that black & white what if the reference is live acoustic music and not another system?

There it is in a nutshell, you pick a side and live with it, because I’m done explaining.

You’re certainly entitled to pick your side. My point is that setup matters and it should be considered as part of the overall assessment.

david

 

[MarkusBarkus]

I think for the purpose of comparison, I would want to hold all variables except the new component, constant. That makes sense to me and is what I have done and would do...not just in audio.

However, the gray area, to me is comparison *versus* optimization. I think it would be possible to optimize a new component when adding it to an existing system, which might involve speaker position, or other changes.

Then, I think you would be comparing the "perfect" optimization of A versus the "perfect" optimization of B.

Conveying all of that in a review would be challenging, and ultimately, my variables would be different from the reviewer's system/room.

If I listened to, liked and bought a new component, I would be expecting I might change elements of my system/room at some point down the line. Sometimes for a perceived "need" sometimes just to experiment. I don't expect any review(er) could tell me that before it entered my system/room/brain.

 

[Atmasphere]

It is interesting that since you started playing around with class-D amps all of the sudden you are interested in negative feedback. You know Ralph, Halcro achieved distortion levels nearly immeasurable, but the resulting sound was sterile and amusical.

Now let’s get to the learning part, since you have developed a fascination with maximizing “gain bandwidth product” let’s have a conversation on that and see who is teaching who.

What is the slew rate of your Class D amplifiers with all this negative feedback? How many Volts per micro-second? What is the Gain-bandwidth product of these class D amps? Is it in the MHz. Is it in the GHz? 100 KHz? What is the closed-loop frequency response of your Class D amplifiers?

You know Ralph, if it is Gain Bandwidth Product that you value perhaps you need to start designing with Op-Amps. An Analog Devices LT1993 instrumentation differential amplifier will get you 7 GHz gain bandwidth product, with extremely low distortion and differential input and output.

Are you familiar with the design of high-speed oscilloscopes? Do you know what high-speed design techniques, and circuit-topologies are implement to achieve their ultra high-speed performance and ultra low distortion?

You ever implemented current feedback in your products? With nodal distributed multiplying output?

You see Ralph, while I have not done board level circuit design in over 20 years, I still retain the knowledge and understanding.

Let’s keep exploring the learning. Do you have any experience with hybrid design technology on ceramic substrates? You see Ralph, I began my electronics designing career doing space electronics for NASA.

But I’m here and ready to learn so tell me something that I have not covered during my academic years and professional career in physics & electrical engineering. I am interesting in learning. Answer the questions above and we can continue with a technical exchange.

I've always been interested in feedback; there's nothing 'sudden' going on. I've also always been interested in connecting the dots between how we perceive sound vs what we measure. For most of my life, the devices didn't exist that would really make the application of feedback practical in an audio power amplifier without causing brightness and harshness due to distortion generated by the feedback itself. You really have to get over about 30dB and that has to be at all audio frequencies before you really start to get around that problem (so while this is all control theory of course you do still have to take the human hearing perceptual rules into account). One of the issues of course is that the typical feedback node (the cathode of a tube for example) is not linear and so the feedback signal is distorted before it even can do its job. That's why you need so much feedback. In tubes this is literally impossible due to phase margin issues (oscillation) and that pesky GBP... Futterman made an OTL that boasted 60dB of feedback but that value fell off dramatically at higher frequencies, causing distortion to rise with it. We didn't want that; we wanted the distortion vs frequency to be a ruler straight line and the distortion spectra to be unaltered with frequency; hence zero feedback OTLs.

IMO the Halcro seemed to sound so sterile because of its distortion spectra- its not enough to merely have low distortion, since the ear converts all forms of distortion to a tonality; also the ear is keenly sensitive to higher ordered harmonics on account of using them to sense sound pressure. The reason an SET sounds so rich is thus also due to distortion spectra. The difference is how the 2nd and 3rd harmonics are able to (or are not able to) mask higher ordered harmonics. An SET is quite good at this despite having more higher ordered harmonics than almost any modern solid state amp. In most cases this means you lose low level detail but you gain some musicality. I like to have both, which is why we made/make OTLs- you get more detail; easily heard (and not surprisingly the distortion is lower too) and now our class D.

As you know I've been building zero feedback class A triode fully differential balanced OTLs for nearly 50 years. If you are not running feedback you need bandwidth like crazy (one of the failings of SETs, where the OPT is the bandwidth limitation but it does help that upstream there is bandwidth). We've not measured the slew rate of our class D but it is certainly higher than that of our tube amps (which have an output section risetime of about 600V/usec) as the output devices effortlessly switch at 60MHz. The class D has bandwidth (and distortion spectra) similar to a lot of lower powered SETs; it begins to roll off around 20KHz (but with about 3-4 orders of magnitude less distortion and a very nice 1st Watt). There is a lot of correction at that frequency that isn't available to an open loop design. Most of our prototypes were open loop FWIW; they demonstrated that the technology was worth pursuit.

Its true that many modern opamps have lots of GBP; very nice since as long as you don't ask more than about 20dB out of them you don't run into problems with feedback falling off at higher audio frequencies (I'll leave their design to others since you can buy them for just a few dollars...).This was not true decades ago; if you try to replace older opamps in vintage guitar pedals you find out real quick that the opamps in them were part of the 'sound' which really shouldn't happen with opamps. In theory.

We've looked into current feedback strictly for fun, since there are no speakers based on current drive its impractical. If you want to make a viable product you have to employ voltage drive since 99% of all speakers made are based on voltage drive. The other 1% are power drive and exist only in the high end audio market. Any zero feedback tube amp is a power source device (although imperfect). That 1% of speakers plus some crossover examples from the voltage camp has been a large enough market that we've made a career of it. Its nice to finally have an amp that can address the greater market without having to compromise the resulting 'sound'.

'Space electronics'... delightful

 

[Carlos269]

It’s a factor, the rest including single component change are your assumptions there can easily be other changes that went unnoticed. Also why is the conversation limited to a single event?

A review is a snapshot it is not evergreen.

Many are competently designed to be nonlinear, a small bass or midrange boost can be attractive in the right context!

And those non-linearities should be noted, documented and described in the review, not compensated for and masked.

Really?

Yes, really. Explain to me how you discern the contributions from the speaker/room interaction from those of the component’s under review and evaluation.

Known reference for whom? What if the reference is flawed? Do we all agree on the same reference? It’s never that black & white what if the reference is live acoustic music and not another system?

In order for a reviewer to gain credibility, he must be consistent and have a basis for his statements. Hence the term “reference” system.

You’re certainly entitled to pick your side.

david

The reader can pick a side and do what they will but that doesn’t preclude one from identifying and exposing the flaws and fallacy in a reviewer’s assessment process, specially if the reviewers presents it as novel, enlightening, and more sophisticated than the traditional, A/B swap in the reference system, approach.

 

[Carlos269]

I think for the purpose of comparison, I would want to hold all variables except the new component, constant. That makes sense to me and is what I have done and would do...not just in audio.

However, the gray area, to me is comparison *versus* optimization. I think it would be possible to optimize a new component when adding it to an existing system, which might involve speaker position, or other changes.

Then, I think you would be comparing the "perfect" optimization of A versus the "perfect" optimization of B.

Conveying all of that in a review would be challenging, and ultimately, my variables would be different from the reviewer's system/room.

If I listened to, liked and bought a new component, I would be expecting I might change elements of my system/room at some point down the line. Sometimes for a perceived "need" sometimes just to experiment. I don't expect any review(er) could tell me that before it entered my system/room/brain.

Well stated. Now constrain those words/thoughts to the reviewer, his role, evaluation process, and the best interest of the reader and you will be fully inline with what I have stated previously.

 

[Ron Resnick]

It is valid to believe as a matter of scientific experimentation that only one variable should be changed at a time. It is valid to believe as a matter of amplifier review methodology that speakers should be repositioned to optimize sound when comparing amplifiers.

I see the controversy here — whether speakers should be repositioned to optimize sound when comparing amplifiers — merely as a matter of review philosophy and methodology. I think either approach is fine as long as as the selected approach is disclosed in advance and explained.

In my opinion, in this context of amplifier reviewing, there is no principled, independent basis for determining objectively that one of these approaches is correct, and that the other approach is incorrect.

 

[Carlos269]

It is valid to believe as a matter of scientific experimentation that only one variable should be changed at a time. It is valid to believe as a matter of amplifier review methodology that speakers should be repositioned to optimize sound when comparing amplifiers.

I see the controversy here — whether speakers should be repositioned to optimize sound when comparing amplifiers — merely as a matter of review philosophy and methodology. I think either approach is fine as long as as the selected approach is disclosed in advance and explained.

In my opinion, In this context of amplifier reviewing, there is no principaled, independent basis for determining objectively that one of these approaches is correct and that the other approach is incorrect.

Dead wrong. For all the reasons provided earlier. Consult your technical editor and if he’s competent, he will agree with what I have stated. It is not a matter of review methodology, it is a matter of correct or wrong. There is no middle ground here. Sorry.

 

[MarkusBarkus]

...I don't expect a reviewer to hold a reference system as a static set of gear/conditions forever. If they did, that might be interesting over time, and add consistency to the reporting. But, since they are only, for the most part, human, I expect them to add/subtract and hopefully learn as time passes.

I do expect them to maintain consistency for the gear being reviewed. And certainly, if something has been changed/added/subtracted, we should know about that. I would prefer that things be held constant.

I have been fascinated by photos of rooms and posts of new power infrastructure by some reviewers, wherein we get insight into the room or the power conditions...and it's really not good.

I find myself thinking: here I am sweating every detail, dedicated lines, cable management, vibration management, and the reviewer is running off old aluminum service cable, small-gauge (in size) wire...or has multiple speakers adjacent to the subject models, rooms crammed with stuff. Wow. I'm not the room/infrastructure police, but it really impacts my impression of the fundamental, foundational basis of the review itself.

I do think that the current trend of noting additional accessories, cables, power conditioners, room size/dimensions is a positive trend. I don't know that it tells me anything specific I can deduce about the component under review, but I prefer to make that judgement for myself, if the info is made available.

 

[Atmasphere]

I have been fascinated by photos of rooms and posts of new power infrastructure by some reviewers, wherein we get insight into the room or the power conditions...and it's really not good.

We had a reviewer do a review of our MA-1 amplifier and somehow he managed to blame the amp when the aged out outlets in his apartment got significantly warm. If that were a customer I'd have advised them to get the outlets replaced as they were clearly unsafe...

 

[Al M.]

I have been fascinated by photos of rooms and posts of new power infrastructure by some reviewers, wherein we get insight into the room or the power conditions...and it's really not good.

Yup, lots of reviewers' rooms are not confidence inspiring.

 

[andromedaaudio]

I m a DIY er .
I do the reviewing myself

 

[microstrip]

(...) IMO the Halcro seemed to sound so sterile because of its distortion spectra- its not enough to merely have low distortion, since the ear converts all forms of distortion to a tonality; also the ear is keenly sensitive to higher ordered harmonics on account of using them to sense sound pressure. The reason an SET sounds so rich is thus also due to distortion spectra. The difference is how the 2nd and 3rd harmonics are able to (or are not able to) mask higher ordered harmonics. An SET is quite good at this despite having more higher ordered harmonics than almost any modern solid state amp. In most cases this means you lose low level detail but you gain some musicality. I like to have both, which is why we made/make OTLs- you get more detail; easily heard (and not surprisingly the distortion is lower too) and now our class D.
(...)

Ralph,

Perhaps you are wanting to address this question - why not simply adding a tube preamplifier with the proper harmonics to the Halcro? In such case the extremely low contributions of the Halcro should be completely masked.

I could formulate it in another way - why aren't you able to have an OTL ampllifier with the transfer function of the MP1 and the MA2? Why do we need two separate units connected in series?

 

[caesar]

...

I find myself thinking: here I am sweating every detail, dedicated lines, cable management, vibration management, and the reviewer is running off old aluminum service cable, small-gauge (in size) wire...or has multiple speakers adjacent to the subject models, rooms crammed with stuff. Wow. I'm not the room/infrastructure police, but it really impacts my impression of the fundamental, foundational basis of the review itself.
...

Excellent observation. Those guys are just selling the dream... there is always something better every new issue or post

 

[caesar]

Are we only debating amplifiers?

Has someone updated their cables to "state of the art", where the bottom end was so robust they needed to add more bass traps and the highs were so well articulated they needed to adjust the room treatments at first reflection points?