Master Thread: Are measurements Everything or Nothing?

[Sgt. Ear Ache]

My daughter manages to tune her guitar by ear quite accurately and is consistent at it, in her case its a skill she has learnt, just as assessing hi fi properly by ear is a skill that is learnt through experence.

Of course. I can get pretty close myself. I can't however hear what kind of varnish is on the guitar. But you may have missed the sarcasm in what you quoted?

 

[rdenney]

My daughter manages to tune her guitar by ear quite accurately and is consistent at it, in her case its a skill she has learnt, just as assessing hi fi properly by ear is a skill that is learnt through experence.

Experience? Not so, as it turns out. Experienced listeners fared no better than inexperienced listeners in tests.

Trained? Okay, now you're talking. Various kinds of distortion can leave tell-tale artifacts, but these are often in otherwise silent or very quiet parts, or in the fade to those silences, which have to be turned up to hear them. Or something like that. I have not received that training, and frankly don't want to.

Rick "and, yes, it is training that allows one to tune a guitar by ear, too--training to recognize the beats from phase interference" Denney

 

[audio2design]

QUOTE="pkane, post: 1058647, member: 396"]
Then you may want to share these, if they are recorded, so others can confirm. If not, at least describe the method and controls you used to determine that it's really the audible difference at -120dB and not some other confounding factor that's responsible for what you've heard.
[/QUOTE]

I call BS would have required less typing.

 

[rwortman]

So long as McIntosh continues to support it.

Rick “half of whose system was made by companies that no longer exist” Denney

Mcintosh has been around a good long time and is currently backordered on everything. Probably not going anywhere anytime soon. I’m 64 years old. I don’t need them to support it forever. I’d am betting (actually) that PCM or DSD audio will not be replaced by some other digital scheme for a good long time. I don’t have anything made by companies that no longer exist although I admit that the oldest component is “only” 17 years old.

 

[MaxBuck]

You're not wrong... It's called anxiety.

What I like the most about measurements is that they give me a good yardstick for assessing whether worrying about something is a waste of time.

FOMO (fear of missing out) seems to be strong in the audiophile community.

 

[ClicketEKlack]

So if measurements are everything, there should be no perceivable difference between a Topping A90 and a Singxer SA-1, both of which measure superbly? As an owner of an A90, I admit I have a sliiiiight case of FOMO about the Singxer. Are all the pro-Singxer comparisons then simply the result of sighted bias, or is there an audible SQ improvement with a Class-A(B) amp?

 

[Blumlein 88]

So if measurements are everything, there should be no perceivable difference between a Topping A90 and a Singxer SA-1, both of which measure superbly? As an owner of an A90, I admit I have a sliiiiight case of FOMO about the Singxer. Are all the pro-Singxer comparisons then simply the result of sighted bias, or is there an audible SQ improvement with a Class-A(B) amp?

How did you carry out your comparisons? Level matched and all?

 

[rdenney]

So if measurements are everything, there should be no perceivable difference between a Topping A90 and a Singxer SA-1, both of which measure superbly? As an owner of an A90, I admit I have a sliiiiight case of FOMO about the Singxer. Are all the pro-Singxer comparisons then simply the result of sighted bias, or is there an audible SQ improvement with a Class-A(B) amp?

Sighted bias.

I’m sitting here surveying my most recent purchase, stony little ultralight Class D amplifier that runs cool enough to hide behind other stuff, but puts out more power and does so more cleanly than the two AB amps it replaced. I look at those, and I feel deep pangs of regret. Those amps (I was really effectively only using one of them) are big, heavy, and hot, and they have the look of serious purpose. But compared to that little D box, only manage about half the power, and with less forgiveness for low impedance speakers. They can’t be bridged, and into my current speakers were a dead end—there was no way to really take advantage of both of them. But they sure look serious.

And they sound great, but no better than the little box, until the limit, when they sound constrained and compressed in comparison. The class D box sounds open and effortless even when the clipping lights are flashing. There is a difference between 125 watts vs. 250 watts at the same distortion into 8 ohms, or 150 watts vs. 350 watts into my 6-ohm speakers (nominal), or 200 watts vs. 450 watts into the 4 ohms that my speakers actually present.

My right brain sure wanted those B&K amps to sound better, but my left brain knew they wouldn’t.

My left-brain prediction is a blind test would show a difference only at the limit, where the less powerful 50-pound AB amp was clipping but the 10-pound D amp was not.

But my right brain is afraid I’m missing out on something. Those B&K amps will go back to each driving a pair of (right-brain) 8-ohm Advents when I set up a new spot in my new shop.

Rick “whose on charge here?” Denney

 

[ClicketEKlack]

How did you carry out your comparisons? Level matched and all?

I've done no comparisons, only read others'.

 

[Blumlein 88]

I've done no comparisons, only read others'.

Then yes most of what you have read is likely sighted bias.

 

[rgdawsonco]

Sometime over the summer I was strolling through COSTCO, and I see stack of Denon AVR-S760H’s, priced at $439. I have little need for another Denon AVR, but I could not resist, so I snapped one up. It sat unopened for a few months. Then as (bad) luck would have it, my Denon X4500 decided to freak out, and I had to send it in to get fixed. I had been using a Denon X6200 for two channel listening elsewhere, and I had to move that to the theater where the 4500 was. So I am temporarily using the S760 for two channel.

Now, I think I can tell the difference in speakers easily. I think I can start to tell the difference in MP3’s, say below 192kbps. But I don’t think I can for sure tell the difference between any of my Denon AVR’s from near bottom to near top of the line. My two channel speakers are Klipsch Cornwall IV’s, which I really love. 75 watts is plenty for those speakers. I use a squeezebox duet over toslink or R-pi over hdmi to stream. I can’t tell the difference there either.

Clearly, the the lowly S760 would measure significantly worse than the X4500 or X6200. But it is barely, if at all, audible to me. I love reading the measurements on this site as much as anyone, but it is sort of meaningless without any understanding of what really matters. So my question is, what matters?

 

[Tom C]

The eternal question…
The answer won’t be the same for everyone, which adds to the difficulty, but things are to the point where it depends on what you’re looking for.

 

[Taddpole]

"Clearly, the the lowly S760 would measure significantly worse than the X4500 or X6200."

Basing all this on quite an assumption.

 

[Blumlein 88]

Well we've seen this sort of thread a few times. For most electronics even poor ones are below audibility. Then there are edge cases. For instance your Cornwalls likely need a pretty quiet low noise amp more so than is the norm. If you don't hear any noise with what you have at the highest volume you listen with then you are good.

I think quite often what this site shows is many expensive items perform worse than affordable gear. So if a $5000 amp measures with audible perfection, it still is a bit of a scam if a $1000 amp measures even better with more power. The better measuring amp won't sound better (unless you need the power), but why pay more for less even if both are effectively blameless in use.

 

[Blumlein 88]

Now your probably thinking my previous post didn't tell you much. So I will elaborate.

Firstly you want flat frequency response 20 hz to 20,000 hz +/- .1 db. An old adage is 85% of HiFi is frequency response. The majority of the time you hear a genuine difference between components it is due to a slight frequency response difference.

Next is noise. If you turn your volume to the max you will ever listen, you don't want to hear noise at the listening position or over your phones. For speakers if noise is 90 db below max signal you probably are fine though one can find edge cases. So 90 db SNR is almost surely good. Headphones I'm less sure where to peg it maybe lower than that.

Distortion: hmmm that one can get a bit more complicated. Generally you can't hear distortion once it gets to - 60 db or .01%, but that is with test signals. With music you probably aren't going to hear it at -40 db or 1%. So I'd say just to be sure you are okay at .01% and since you have a few pieces of gear to run thru like source, preamp, amp, I say lets make it -70 db or .003%.

And then there are the speaker's distortion. Not many can play loud at less than 1% though some can. It is mostly 2nd and 3rd harmonic with isn't very noticeable. Some headphones can do better. So it still isn't unreasonable to want -60 or -70 db distortion in the preceding electronics just to be sure.

Then lets us talk about SINAD. That is noise plus distortion. The lower the better. But it can be misleading. As a hypothetical example I could have an amp with -40 db distortion, but very low noise at -100 db. SINAD is going to be -40 db because the distortion dominates. On music that amp may sound just fine. Or another hypothetical example, I might have an amp that has -100 db distortion, and yet -40 db noise. SINAD is also -40 db because noise dominates. That amp, especially with a sensitive speaker like your Cornwalls would be unusable as you'd hear so much hiss and noise.

The good thing about SINAD is if you get numbers like -90 db or better you don't have to worry about noise or distortion because both are so low they'll not be a concern.

Oh, and speaker or headphone frequency response.............that gets even more complicated, and is the main thing that you will hear as different between different speakers and headphones. I always suggest people build a system from the speakers/or headphones backwards. Find the speaker or headphone you like and cater to its needs. The Harman work and research to develop a target for speakers using Spin-o-rama results is revolutionary. It still doesn't fully account for loudness capabilities and a few other things, but it is a big boon to choosing speakers. Their work in developing a target response for headphones is also a big step forward though still being developed.

 

[xaviescacs]

However, obfuscation, strawmen and borderline trolling...very much detected! lol...

I know it's subtle, but the difference between failing to measure anything and undetecting something (it) is there, and pointing that out it's not trolling IMHO. If we try to measure something and we can't, we don't read anything, what do we know? What knowledge have we gained? That we are unable to measure anything under those conditions, nothing more. We can believe, almost be certain, there is still something there to be measured, and keep trying, namely, trying to reject the null that there isn't anything, but we can't be sure there isn't anything, namely, we can't prove the null to be true, this is why is called null.

Logic. A rational point. Actual evidence. All things that are undetected...

Following your reasoning, anything we can imagine it's there, but undetected, right?

 

[Steve57]

Experience? Not so, as it turns out. Experienced listeners fared no better than inexperienced listeners in tests.

Trained? Okay, now you're talking. Various kinds of distortion can leave tell-tale artifacts, but these are often in otherwise silent or very quiet parts, or in the fade to those silences, which have to be turned up to hear them. Or something like that. I have not received that training, and frankly don't want to.

Rick "and, yes, it is training that allows one to tune a guitar by ear, too--training to recognize the beats from phase interference" Denney

Thats an interesting point,
Trained implies someone has been shown to do something properly.
Experence shows they have been exposed to something long term.
Neither is a guarantee of competence.
I've run a forum audio meet in the UK for the last 15 years, many of us 'regulars' consistently pick up on distortions and anomalies in the systems we and others bring, whilst quite a few don' hear anything wrong. All of us are experienced listeners.

I've had no formal training on listening, only my experience.
My view is measuring is important, but present standards do not explain the weakness in many systems.
Measurements can prove what ever you want them to.
For me its the measurements of the overall system that counts, at different voltages and with real world signals through it, and with different speaker types attached..

 

[Sgt. Ear Ache]

I know it's subtle, but the difference between failing to measure anything and undetecting something (it) is there, and pointing that out it's not trolling IMHO. If we try to measure something and we can't, we don't read anything, what do we know? What knowledge have we gained? That we are unable to measure anything under those conditions, nothing more. We can believe, almost be certain, there is still something there to be measured, and keep trying, namely, trying to reject the null that there isn't anything, but we can't be sure there isn't anything, namely, we can't prove the null to be true, this is why is called null.

Following your reasoning, anything we can imagine it's there, but undetected, right?

Following my reasoning, if somebody in the apartment next door to me is blasting their heavy metal too loudly, it's going to bug me. If somebody in their car a mile away from me is blasting their music too loudly, it won't bother me at all. I won't even know they exist. You saying "but how do you know you can't hear them?" is just a ridiculous obfuscation. We aren't having a philosophical debate about the nature of silence here. We're talking about the physical limits of audibility in so far as it affects the reproduction of recorded information. And all any of us are saying is that before we'll believe someone can do super-human things (like hearing a car stereo a mile away) we'd like to see some sort of actual evidence such can be done. If somebody tells me they can fly, and then says "and I've proven it. I just did it here in my apartment." Am I supposed to accept that? No, but if I start seeing studies from reputable sources about some people being able to fly, then maybe we're getting somewhere. So, where are the studies from reputable sources showing that people can hear the things audiophiles claim they can hear? All I see is evidence that they can't...

 

[JiiPee]

Thats an interesting point,
Trained implies someone has been shown to do something properly.
Experence shows they have been exposed to something long term.
Neither is a guarantee of competence.
I've run a forum audio meet in the UK for the last 15 years, many of us 'regulars' consistently pick up on distortions and anomalies in the systems we and others bring, whilst quite a few don' hear anything wrong. All of us are experienced listeners.

I've had no formal training on listening, only my experience.
My view is measuring is important, but present standards do not explain the weakness in many systems.
Measurements can prove what ever you want them to.
For me its the measurements of the overall system that counts, at different voltages and with real world signals through it, and with different speaker types attached..

I have couple of questions:

How were the distortions and anomalies You picked up confirmed? Through measurements, or...?

When You say that measurements can prove anything, do You mean by cheating while doing the measurements, or by cheating with the interpretation of the results? When doing properly and honestly, the measurements only provide facts (within the error margin depending on the method).

What are the system weaknesses current measurement standards fail to explain?

 

[rdenney]

I know it's subtle, but the difference between failing to measure anything and undetecting something (it) is there, and pointing that out it's not trolling IMHO. If we try to measure something and we can't, we don't read anything, what do we know? What knowledge have we gained? That we are unable to measure anything under those conditions, nothing more. We can believe, almost be certain, there is still something there to be measured, and keep trying, namely, trying to reject the null that there isn't anything, but we can't be sure there isn't anything, namely, we can't prove the null to be true, this is why is called null.

Following your reasoning, anything we can imagine it's there, but undetected, right?

If the hearer can’t be specific enough to suggest a measurement, let’s start with controlled testing to verify that the effect in question can be reliably heard, and then, even if it can, whether anyone cares. If we can’t do that, then finding an objective measurement won’t be possible. If we can, the bet is that it can probably be explained by measurements currently being made. That’s true often enough for it to be a good bet.

But this is a straw man. the experience of your club notwithstanding, the claims are nearly always that the effect is obvious to anyone with a sufficiently expensive system and who is sufficiently experienced (described as experience listening to expensive systems), including (contradictorily) the disinterested wife in the next room, and that it’s the difference between musical truth and a veil over the music.

(We need Amir to describe—yet again—the Harman training program he went through as part of his work evaluating codecs at Microsoft.)

Rick “we have to keep reminding ourselves of this” Denney