Why Some Audiophiles Fear Measurements

[Ethan Winer]

am I the only one to think that ethan was not completely serious with his figures?? heck, maybe he was?, it's just that (given the context and when it was posted) I felt it more tongue in cheek and deliberately provocative rather than an actual figure he felt true? hmm, maybe ethan had better tell exactly what he meant, given the attention that statement has attracted.

LOL, I was being serious and kidding and provocative at the same time. One problem with percentages is they're not logarithmic, which is how we hear both frequency and amplitude. Understand that 90 / 10 percent is really only 20 dB.

--Ethan

 

[Ethan Winer]

I agree the room is very important but if the basic system has real flaws the room won't fix them.

I was careful to qualify my statement with "We can ignore electronic gear and cabling etc because anything decent will be flat within 1 dB over the audible range." Key word is "decent." Other times I'll refer to "competent" gear when discounting the contribution that electronic devices make to the overall sound quality. Yes, obviously-colored gear will have an obvious affect on the sound. But to me anything colored enough to be obvious is not "high fidelity" by definition. People may like that sound, and that's fine with me! But it's not hi-fi.

Fortunately deep notches are not as audible as peaks so it works in our favor.

This is such a common misconception that I wrote a short article with audio examples to explain it:

Audibility of Narrow-Band EQ

--Ethan

 

[Ethan Winer]

Those 450 watt Monos, maybe they can handle anything, maybe they can't. Maybe they would have more than a 1% effect in a positive way, maybe not.

This is a great point. A power amplifier driven into clipping will have more than 1 percent affect on the sound.

The room and speakers make up 95%, or 90%, or 85%, whatever, the vast majority. I would say, though, that the speakers and the room must be considered as one, inseparable, in this analysis.

I agree with that too. It's very difficult to separate the room from the speakers.

--Ethan

 

[Ethan Winer]

I would keep the logarithmic scaling with power that corresponds to the way we hear.

So, Cables and electronics are 1 dB; Speakers are 10 dB; and rooms are 30 dB.

That gives a weighting of Cables and electronics 2%; Speakers 24%; Room 73%.

Personally, I don't find this subjectively different from what Ethan originally said making his argument, so far, quite robust.

I don't think the intention of these weightings should be to directly apportion money to stereo components. They are, rather, estimates of each component's relative contribution to deviations from neutrality. Engineering-wise, one should pay whatever is necessary to make the contribution of each component equal.

You totally get the spirit of what I was saying. If the ratio is 73 / 24 / 2 that's fine. At least we have a starting point to quantify what we're discussing, without relying on empty subjective terms. And you are correct that I said nothing about how people should spend their money, even though some apparently read that into my post.

--Ethan

 

[silviajulieta]

Dear Ethan: " ++++++ I explained pretty clearly how I arrived at my 90/9/1 percent figures regarding frequency response (only). " +++++

the whole problem is that that is a plain and " simplistic " way to put things, for say the least and people can agree or disagree about those %%%. These kind of %%% as I already posted and explained means almost nothing to me other that for you are " simplistic " and that's what you are enttilted for/with.

regards and enjoy the music,
Raul.

 

[Ethan Winer]

IM distortion is incredibly more complex than just that when it comes to audibility and importantly the test protocol-procedure it is applied to ... "The two tones have equal amplitude and they are one octave apart. The signal peaks are about 1.8 volts, and the distortion peaks are about .09 volts, or 5%, and the ratio of rms averaged distortion divided by the rms signal is about 4%. This distortion doesn't look so bad, but it is obviously higher and more complex than singletone distortion." ... this supports my point that measurements are meaningless unless used in their correct context

You are correct that context is very important, but I maintain that specs are even more important. I'll try to break it out logically:

10 percent IMD might be benign on two frequencies an octave apart, such as a solo flute and it's prominent second harmonic. But on a cymbal crash the low-mid tones added by the IMD will be obvious and irritating to hear. IMD on cymbals and tambourines and spoken voice is especially nasty because it adds low frequencies that are not normally present with those sources. So the obvious conclusion when buying a new amp etc is to look at the specs, and disregard anything with unusually high levels of IMD. Yes, it may not be audible on all sources, but it will certainly be audible on some sources. So it's definitely not a useless spec.

The same goes for THD and frequency response. As Ponk said to Mike a few posts back, "So these amps will make warm, sweet speakers with a bass-extension emulating low midrange hump and a harshness-reducing upper midrange dip sound wonderful, then they'll turn right around, without any eq, or any other changes in the system and make dry, treble-forward speakers with a lean bass sound wonderful too?" A coloration that sounds good on some sources may sound bad on others, and have little affect on yet others.

To me there is more to this than just the 4 parameters you feel explain audio production

As soon as you show what else there is - being excruciatingly specific! - I'll be glad to update my opinion.

it still needs to be shown how those 4 parameters correlate to what we hear

It sounds like you're talking about psychoacoustics, which is unrelated to gear parameters. Further, and I'm sure I've said it a dozen times this past week alone, what you hear one moment may not be what you hear the next. You could easily like the sound of some recording on someone's system today, but hate it tomorrow. This is inside your head and has nothing to do with what happens inside the gear. So it's not proper to bring up "how we hear" into this discussion. I'm talking about gear parameters, not the foibles of human hearing.

if I say "your 4 parameters" often it is not intended as an insult

I never took it as an insult, and I call them my four parameters too. But to all who believe there is more to assessing the fidelity of audio gear than these four parameters, please tell us specifically what more there is using quantifiable objective terms. "Resolution, transparency, soundstaging" are all useless because they mean different things to different people, and they also have no metric.

--Ethan

 

[The Smokester]

Dear Ethan: " ++++++ I explained pretty clearly how I arrived at my 90/9/1 percent figures regarding frequency response (only). " +++++

the whole problem is that that is a plain and " simplistic " way to put things, for say the least and people can agree or disagree about those %%%. These kind of %%% as I already posted and explained means almost nothing to me other that for you are " simplistic " and that's what you are enttilted for/with.

regards and enjoy the music,
Raul.

What is wrong with plain and simplistic? For each different system, the specific numbers will be different and dependent on many other things as was said by others, including the possibility that a large room with a small system listening in the near field the room may make little difference.

That's where "understanding" comes in and I know you have a lot of it when it comes to audio. Ethan has provided one idea and a framework for how to prioritize the optimization of a system. As one improves a system then the system-limiting phenomena and the numbers change. Pretty soon after improvements the room is contributing just as much deviation to the frequency response as the speakers, amps and the cables...But now all the deviations are at the 1 dB level. If one still has audio nervosa (like many of us do) then you re-evaluate where the offending deviations are coming from and go after those... Maybe it is noise so look at the power source next time around. Or maybe the "long pole in the tent" is not frequency response but distortion so examine whether to change out speakers, amps or source depending on what's cost-effective.

Many of the statements on this thread, while they are understandable and interesting to read because they relay the odessey of high end in which we share a common interest, are not directly useful in an engineering sense because they can't be translated from one system to another. Saying "my system sounds great after I installed new cables" or power turns out to be really important" does convey an idea to look into but how would one prioritize improving their own system based on such comments? These statements might be useful fs one is duplicating someone elses system, but how many of us are doing that? I have never seen two identical audiophile systems. The point is that there can be ways to usefully quantify a system and prioritize what can be done to further its development.

 

[Ethan Winer]

I explained pretty clearly how I arrived at my 90/9/1 percent figures regarding frequency response (only).

Forgive me, but where did you explain that? All I could find is you saying there's more than just frequency response, which of course is true.

--Ethan

 

[DonH50]

Don, I don't understand these equations. If there is no 2nd-order distortion there is still 6.021 dB of IMD? How can this be?

The dB terms are derived from the absolute levels in the equations. The derivation is straight-forward but tedious. In the limit, of course there is no IMD if there's no HD, so the absolute value is zero and relative terms are meaningless. That would mean it is a totally linear system. However, if there is any nonlinearity (and there is always some in a real system), then the IMD tracks the HD at a somewhat higher level given by the equations.

Another way to think of it is that as the HD terms go to (-)infinity (in dB) anything added to that is still (-)infinity.

L'Hopital is still with us...

HTH - Don

 

[kareface]

I think Ethan acknowledged that it was a simplistic method. The point is in the contexts of his comment he wasn't off base. I think the primary point of of debate is that he's excluding many other factors that will have an impact on reproduction, and people are repeatedly trying to point out the fact that he left them out. He knows other factors were left out, and he's arguing that in the context of his post the comments are accurate. So we get this cycle of people pointing out it doesn't include y factor, and him saying that y factor wasn't part of the original equation, repeat for z. In the narrow context of the original comment his numbers are largely valid. Does that necessary reflect what you hear? It's completely situational, if you have a clean enough system the room will be the all encompassing factor. If you have some glaring flaw in your system, it'll be reflected in the reproduction and might be more audible than the linear FR distortion from nodes. I still believe that using real world scenarios over 50% of the influence is acoustic. Tonality varies widely with the environment, many rooms are under treated and produce audible noise and nodes can be quite influential creating huge peaks that mask higher frequencies or drop frequencies out all together. If you think about the impact just poorly placed speakers have on reproduction you'll understand how influential the room is. Even if the speakers are placed so they image well, if they are positioned wrong relative to the environment they can sound terrible.

 

[silviajulieta]

Dear Kareface: +++++++++ " The point is in the contexts of his comment he wasn't off base. I think the primary point of of debate is that he's excluding many other factors that will have an impact on reproduction, and people are repeatedly trying to point out the fact that he left them out. " +++++++++++++


yes, that's all about and that's why is simplistic and IMHO almost useless. Thank you.

regards and enjoy the music,
Raul.

 

[microstrip]

Yes, complete data requires a lot of numbers and conditions to be reported! But the -80 figure I mentioned is mostly level-independant. That is, at a low volume you might not notice distortion even if it's only 30 dB below the music, but at high levels -80 is soft enough not to hear it. So consider -80 as a worst-case condition for "transparency."

--Ethan

Apologies for my insistence. I tried to keep it technical (I repeat myself : for me measurements should report numbers and units, but the conditions (in numbers and units) of the measurements have to be clearly specified). But your answer is even more subjective than most posts in this thread - low volume, music , high levels. Where are the W's and the Hz's?

 

[silviajulieta]

Forgive me, but where did you explain that? All I could find is you saying there's more than just frequency response, which of course is true.

--Ethan

Dear Ethan: Simplistic as your %%% statement.

regards and enjoy the music,
Raul.

 

[FrantzM]

Dear Kareface: +++++++++ " The point is in the contexts of his comment he wasn't off base. I think the primary point of of debate is that he's excluding many other factors that will have an impact on reproduction, and people are repeatedly trying to point out the fact that he left them out. " +++++++++++++


yes, that's all about and that's why is simplistic and IMHO almost useless. Thank you.

regards and enjoy the music,
Raul.

Simplistic? Yes
Useless? NO!!!

For too many years audiophiles have concentrated on what amount to a few dBs if not fraction of dB while ignoring (anf frankly being ignorant) of the effect of the room on the final sound... Thus cables, thus new amps, new preamps, new tubes, new accessories and once in a while a dose of Quantum treatment .. So it may be an exageration but it is to make a point that is beginning to gain traction... Room and speakers First and should be allocated ... well above 80% of the total budget becasue they account for that percentage ... in what you hear ??? Hmmm?

 

[kareface]

Frantz, I don't think Raul was referring to room treatments being useless, but the predictive value of the hypothesis. Too many factors are ignored to represent what is actually happening. I wouldn't say it's worthless tho, it's an interesting perspective on the issue. When you think about it, many people obsess about perfectly flat FR, and ignore many other elements to the room in the process. FR can be a useful tool, but my personal belief is it should be accompanied by a range of tests to get a full picture. By many audiophiles standards, Ethan would be correct. I've seen lots of rooms that measure reasonably flat but decay poorly. Or people who apply tons of filters to their EQs to achieve flat response only to destroy the SQ. I don't think the comment is worthless, but I also don't think it provides any predictive value.

 

[silviajulieta]

Frantz, I don't think Raul was referring to room treatments being useless, but the predictive value of the hypothesis. Too many factors are ignored to represent what is actually happening. I wouldn't say it's worthless tho, it's an interesting perspective on the issue. When you think about it, many people obsess about perfectly flat FR, and ignore many other elements to the room in the process. FR can be a useful tool, but my personal belief is it should be accompanied by a range of tests to get a full picture. By many audiophiles standards, Ethan would be correct. I've seen lots of rooms that measure reasonably flat but decay poorly. Or people who apply tons of filters to their EQs to achieve flat response only to destroy the SQ. I don't think the comment is worthless, but I also don't think it provides any predictive value.

Dear Kareface: You got my point: " but the predictive value of the hypothesis. Too many factors are ignored to represent what is actually happening. "

+++ " FR can be a useful tool, but my personal belief is it should be accompanied by a range of tests to get a full picture. " +++, absolutely right the FULL PICTURE!.

some of us are trying to " figure " and see all this subject in " wide screen " and FR only is part of it. Along this there are several " factors " where there are no " ceros " and " ones " or " db's " yet but that we think affect what we are hearing, the power quality of the electrical source in the audio system that ML point out is one of them.
Normally we audiophiles like to think not so simplistic but with almost all the factors that surround a specific subject, at least this is my overall " training " . I can think simplistic even I accept that the Ethan %% statement could be accurate for one sole factor as FR but this can't help me in a wider way.

There are only a few " numbers " that as a stand alone number could help me, example: amplifier output impedance or speaker sensitivity. I'm trying to imagine where FR could help me, I mean in which situation?.

Amplifier output impedance can help me to match the amplifier with a speaker electrical impedance: to choose which amp?

Speaker sensitivity could help to choose which amplifier power I need.

and NO these amplifier output impedance or speaker sensitivity " numbers " can't predict the quality sound but help for I make/made less/lower my mistakes when I choose audio items looking for system synergy.

Thank you again.

regards and enjoy the music,
Raul.

 

[kareface]

Raul,

This isn't an attack, but your style of response adds some degree of difficulty to understanding your point. I only say this with the intention of making it easier to convey understanding.

With that said, I understand what you're saying. I would argue that truly understanding the macro is understanding the micro. Humans have great difficulty understanding complex relationships. Many develop false or incomplete causal relations from over simplifications of complex systems that they cling to dogmatically. Viewing things from other perspectives, even ones that seem myopic, can help to define a larger picture. Accurately understanding a lot of little pictures is a better route than trying to make sense of a highly complex system as a whole. Anyways, it's not as if Ethan surreptitiously presented the context of his point.

 

[silviajulieta]

Simplistic? Yes
Useless? NO!!!

For too many years audiophiles have concentrated on what amount to a few dBs if not fraction of dB while ignoring (anf frankly being ignorant) of the effect of the room on the final sound... Thus cables, thus new amps, new preamps, new tubes, new accessories and once in a while a dose of Quantum treatment .. So it may be an exageration but it is to make a point that is beginning to gain traction... Room and speakers First and should be allocated ... well above 80% of the total budget becasue they account for that percentage ... in what you hear ??? Hmmm?

Dear FrantzM: I think that today ( I mean not because this thread. ) almost all audiophiles are aware of the room importance and room/speaker intimate relationship and one way or the other almost all are taking care about.

Yes I'm still ignorant on many subjects about room treatment but I'm in the learning process. My audio system sounds better today than few years ago thank's to changes here and there and one of those changes was room treatment. I'm convinced 100% about the critical importance of room treatment and room/speaker matching.

What I'm not convinced in deep is about those %%% like the one you posted: " well above 80% of the total budget becasue they account for that percentage ". Why 80% or 90% and not 50%? where can I read that must be 80%-90% and not 50%?

Where can I read that the importance for quality performance must be 80-90% in that regard and not 40% or 50%? why the source can't be that 50% or even more? where can we left/leave in analog source the precise/right set up that IMHO is like the room in the speaker/room duo?

Whom said and prove that numbers as %%% like the ones to follow it? as a standard?

A %%% in this regard seems to me that is a subjective-objective subject. So for me those %%% could be way different. As in the music sound reproduction where for me the frequency extremes are critical and the most important frequency ranges because are these extremes the ones that has more influence in the " color " of the whole picture, in this %%% subject by coincidence ( for me ) both extremes are the most important: analog source and room/speakers and then everything in between.

This is my point of view that obviously is different from the one each one of you already had/have on the whole subject. Can I be wrong?, yes but today is what I think till I learn something that explain to me a better way and why to " see " those things under the main %%% subject.


Regards and enjoy the music,
raul.

 

[silviajulieta]

Raul,

This isn't an attack, but your style of response adds some degree of difficulty to understanding your point. I only say this with the intention of making it easier to convey understanding.

With that said, I understand what you're saying. I would argue that truly understanding the macro is understanding the micro. Humans have great difficulty understanding complex relationships. Many develop false or incomplete causal relations from over simplifications of complex systems that they cling to dogmatically. Viewing things from other perspectives, even ones that seem myopic, can help to define a larger picture. Accurately understanding a lot of little pictures is a better route than trying to make sense of a highly complex system as a whole. Anyways, it's not as if Ethan surreptitiously presented the context of his point.

Dear Kareface: Yes you are right. Things are that is not easy for me to explain in precise way what I have on mind because my English vocabulary is limited and not good as my native language that's the Spanish.

Yes, I agree too that part by part could be easy to understand complex " relationships ". I'm under " training " here and ready to learn.

Regards and enjoy the music,
Raul.

 

[kareface]

Where can I read that the importance for quality performance must be 80-90% in that regard and not 40% or 50%? why the source can't be that 50% or even more?

It can be, but the point (much like the one you were making about room treatments) is that most audiophiles focus enough on the equipment that the influences of the hardware will either be minimal, or if it is influential, it's a audibly palatable influence. I'm happy to see many more audiophiles who are looking into proper room treatments, but I would wager the majority of self proclaimed audiophiles don't pay attention to or put enough emphasis on acoustics. I would add of the audiophile that get to the point where a several thousand dollar upgrade will provide little notable benefit to the reproduction, less than a third have antiquate (if any) acoustical treatment. I see a trend of greater emphasis on acoustics, but it isn't nearly as prevalent as it should be given the influence acoustics has on reproduction.