Euphoric index of Audio Performance

[microstrip]

(...) My opinion is that most if not all audible differences are measurable, especially given that most high-grade instruments resolve much better than our ears, but setting up the right measurement can be very painful if not impractical (at least for the vast majority of us). (...)

Don,
It is out of the thread subject but such a reasonable comment can not be ignored . Measurable means returning a number with some significance and with a casual connection with the effect. If you find a difference that you correlate with sound it should have a systematic variation - next time you measure the same value it should result in a similar sound difference. Just getting different numbers is not enough.
Also one of the big problems I see in audio measurements is reducing an enormous set of data to a few useful parameters.
Another problem is prediction. Measurements should tell something (for those who accept that amplifiers sound different, surely.) Can we say how a Krell, a Classe Audio or a Levinson sound just looking at the measurements?

 

[Phelonious Ponk]

Tim,
It seems me you are out of the scope of the situation that has been described initially. We are looking for situations as described by Amir, in which the perception of the sound is so extraordinary that the usual rules do not apply. You are just describing a classical situation of amplifier saturation, progressive clipping and headroom limits.

As far as I understand the Euphoric Audio Multiplier should only apply to small differences - if it is applied to large differences it would overload the system!

BTW, in my experience these situations can happen in a wide range of prices, not only in super systems. Also the EAM is not related to the expectation bias!

My post was a direct response to Amir's comment about the inspiration for the overall idea:

Giving context, my original post was prompted by someone getting a Macintosh amp with double the power relative to his emotiva and came and commented about the improvements. Then someone post that he only had a "3db" improvement and such a thing should have never triggered a big "wow" moment. Question is, was that objection valid? We know there were two difference: different amp and doubling of the power.

...but that doesn't mean I didn't get out of scope. Scope creep is sort of a way of life with me. As is thread drift. I prefer a dynamic environment to a rigidly defined one.

Tim

 

[DonH50]

Don,
It is out of the thread subject but such a reasonable comment can not be ignored . Measurable means returning a number with some significance and with a casual connection with the effect. If you find a difference that you correlate with sound it should have a systematic variation - next time you measure the same value it should result in a similar sound difference. Just getting different numbers is not enough.
Also one of the big problems I see in audio measurements is reducing an enormous set of data to a few useful parameters.
Another problem is prediction. Measurements should tell something (for those who accept that amplifiers sound different, surely.) Can we say how a Krell, a Classe Audio or a Levinson sound just looking at the measurements?

I'll just say, agreed. To your last statement, I once upon a time went on a quest to "prove" I could measure differences among such amps and was largely successful. It was not easy and required a quantifying a number of parameters under various test conditions (as well as the support of an interested store owner and customer pool). It ultimately was not useful because people hear what they want, question every measurement, and repeating such a rigorous set of measurements would be painful (and impractical for me now). I cherish the experience, however, since I learned that people were hearing things we had not previously mesaured, and figuring out how to quantify such things was very painful.

 

[Phelonious Ponk]

Point: At midrange it takes 10 dB extra to power to provide a perceived doubling in volume. 3 dB is detectable by most people (the threshold appears to be around 1 dB or so).

Yeah, pretty bad brain fart on my part; doubling the power = 3db, not 3db = doubling the volume. Intellectual dyslexia. I know better, but I digress. The point stands, though. If you were running one amp close to its headroom limit on very dynamic music and you switched to one with twice the power, even if in all other ways they were identical, you could hear a lot more improvement than the measurable 3db more volume. It could happen. And that would be the kind of euphoric index I would like: The kind that makes sense.

Tim

 

[DonH50]

Yes, that is very true. At the clipping point, IIRC from intercept points, adding 3 dB additional headroom would reduce the 3HD by 9 dB or so relative to the peaks, a significant change. Positively giddy! I have a derivation someplace that might be an interesting post over in the technical arena, err, "area"...

 

[mep]

I think this thread is hopelessly doomed as far as anyone coming to agreement on anything. The nature of audiophiles is to argue about everything. We make the dysfunctional politicians on both sides of the house look like kissing cousins by comparison.

I will say that all recordings that have significant dynamic range always provide a wow factor for me. Recordings with really good dynamic range almost always are thrilling to listen to. And both the power meters and VU meters always tell the truth about what is going on with your source material and there is actually pretty good correlation between them visually. If I have a tape playing that has really good dynamic range, my VU meters will constantly be swinging from -20 to +3. If I look at the power meters on my amps, they will fall to zero and then climb to a high level (I have never buried the power meter) as the amp is tracking the dynamic range of the tape. +3 on the VU meter and the amp is putting out lots of power for the duration of that signal. VU meters showing -20, the meter isn’t going to move off zero on the amps. The same holds true for playing LPs with lots of dynamic range (and I know that some of the digital people with no LP playback capability don’t believe that LPs have much dynamic range, but it’s not true). When I’m playing an LP and I can hear how good the dynamic range is, one look at the power meters provides confirmation.

The same holds true in reverse. Recordings that were made at O VU and don’t vary from that all sound loud, but they have no dynamic range which I think everyone who has heard these all too familiar recordings can attest to. And the meters again won’t lie. If I convert a compressed LP or CD to tape, it will show up as O VU on the tape meters. If I just play it straight from my music server or turntable, you will see the meters on the amp basically not move.

In summary, give me recordings with as much dynamic range as possible and there will be some wow moments to savor.

 

[Phelonious Ponk]

I think this thread is hopelessly doomed as far as anyone coming to agreement on anything. The nature of audiophiles is to argue about everything. We make the dysfunctional politicians on both sides of the house look like kissing cousins by comparison.

I will say that all recordings that have significant dynamic range always provide a wow factor for me. Recordings with really good dynamic range almost always are thrilling to listen to. And both the power meters and VU meters always tell the truth about what is going on with your source material and there is actually pretty good correlation between them visually. If I have a tape playing that has really good dynamic range, my VU meters will constantly be swinging from -20 to +3. If I look at the power meters on my amps, they will fall to zero and then climb to a high level (I have never buried the power meter) as the amp is tracking the dynamic range of the tape. +3 on the VU meter and the amp is putting out lots of power for the duration of that signal. VU meters showing -20, the meter isn’t going to move off zero on the amps. The same holds true for playing LPs with lots of dynamic range (and I know that some of the digital people with no LP playback capability don’t believe that LPs have much dynamic range, but it’s not true). When I’m playing an LP and I can hear how good the dynamic range is, one look at the power meters provides confirmation.

The same holds true in reverse. Recordings that were made at O VU and don’t vary from that all sound loud, but they have no dynamic range which I think everyone who has heard these all too familiar recordings can attest to. And the meters again won’t lie. If I convert a compressed LP or CD to tape, it will show up as O VU on the tape meters. If I just play it straight from my music server or turntable, you will see the meters on the amp basically not move.

In summary, give me recordings with as much dynamic range as possible and there will be some wow moments to savor.

Yep. Dynamic range is good. I agree...well, not so sure about the vinyl part... See? The thread isn't hopelessly doomed after all!

Tim

 

[mep]

I think that dynamic range plays a very large part in what makes recorded music sound real (and we can argue over what "real" means and sounds like I suppose). And having said that, in addition to dynamic range, you need to have a high level of density of information. For the here and now and in my experience, nothing surpasses tape for density of information along with the capability of providing what sounds like an incredible amount of dynamic range. I would say that the "average" audiophile who has not been exposed to the sound of a high-quality 15 ips 2 track tape of music they know and love would all have a "wow" moment when the tape starts.

I found it funny when someone else on this forum commented that it wasn't fair that people played tapes at audio shows because it makes everything else sound like crap. This guy brought some of his prized digital recordings to have played back at the show and was embarrassed at how bad they sounded after hearing a tape being played in the same room. Bruce said it was like bringing a knife to a gun fight.

 

[microstrip]

Giving context, my original post was prompted by someone getting a Macintosh amp with double the power relative to his emotiva and came and commented about the improvements. Then someone post that he only had a "3db" improvement and such a thing should have never triggered a big "wow" moment. Question is, was that objection valid? We know there were two difference: different amp and doubling of the power.

Amir,
Your first very challenging post was about EAM. But if we now start mixing an amplifier with double power (surely a large difference in any sense) with a different brand one (a small difference to some people) , both used in undefined conditions, we can be sure we will never arrive anywhere about EAM, as Mep was wisely suggesting.

 

[microstrip]

(...) I will say that all recordings that have significant dynamic range always provide a wow factor for me. Recordings with really good dynamic range almost always are thrilling to listen to. And both the power meters and VU meters always tell the truth about what is going on with your source material and there is actually pretty good correlation between them visually. If I have a tape playing that has really good dynamic range, my VU meters will constantly be swinging from -20 to +3. If I look at the power meters on my amps, they will fall to zero and then climb to a high level (I have never buried the power meter) as the amp is tracking the dynamic range of the tape. +3 on the VU meter and the amp is putting out lots of power for the duration of that signal. VU meters showing -20, the meter isn’t going to move off zero on the amps. The same holds true for playing LPs with lots of dynamic range (and I know that some of the digital people with no LP playback capability don’t believe that LPs have much dynamic range, but it’s not true). When I’m playing an LP and I can hear how good the dynamic range is, one look at the power meters provides confirmation.

The same holds true in reverse. Recordings that were made at O VU and don’t vary from that all sound loud, but they have no dynamic range which I think everyone who has heard these all too familiar recordings can attest to. And the meters again won’t lie. If I convert a compressed LP or CD to tape, it will show up as O VU on the tape meters. If I just play it straight from my music server or turntable, you will see the meters on the amp basically not move (...) .

Mep,
VU meters have a typical 30dB range - IMHO you can not use them to look for dynamic range in recordings. Also many of them are peak programme meters with response factors optimized for different applications, but not to show dynamic range. BTW, I have no doubt that digital has better measured dynamic range than analog, what we should address is the very difficult to analyze perception of dynamic range of the different media by listeners. And here I agree with you - most of the time vinyl seems more "dynamic" than CD (in the abusive sense).

 

[amirm]

Amir,
Your first very challenging post was about EAM. But if we now start mixing an amplifier with double power (surely a large difference in any sense) with a different brand one (a small difference to some people) , both used in undefined conditions, we can be sure we will never arrive anywhere about EAM, as Mep was wisely suggesting.

But I don't think everyone here agrees that upping the power at that level was a big change. Numerically it was doubling the power. Yet, most people claimed that it shouldn't have made hardly any difference. In other words, they are practicing EAM just the same, but applying a ratio less than 1 to it.

 

[fas42]

PS: What might be more powerful would be a discussion around what categories of system changes make big differences, which ones are incremental, but measurable, repeatable and verifyable, and which ones, while they seem to loom large for some, measure very small and are lost on many. If we could get there, then we could begin to at least undestand what things need to be scrutinized and which ones are given.

I agree wholeheartedly with this thought: might be worth kicking off a new thread focusing on such ...

Frank

 

[fas42]

I'll just say, agreed. To your last statement, I once upon a time went on a quest to "prove" I could measure differences among such amps and was largely successful. It was not easy and required a quantifying a number of parameters under various test conditions (as well as the support of an interested store owner and customer pool). It ultimately was not useful because people hear what they want, question every measurement, and repeating such a rigorous set of measurements would be painful (and impractical for me now). I cherish the experience, however, since I learned that people were hearing things we had not previously mesaured, and figuring out how to quantify such things was very painful.

Now this is getting somewhere, Don! I personally would be fascinated in what parameters, and how you tested and trends in the results obtained. Care to start a thread somewhere talking about this a bit more ...?

FRank

 

[fas42]

I found it funny when someone else on this forum commented that it wasn't fair that people played tapes at audio shows because it makes everything else sound like crap. This guy brought some of his prized digital recordings to have played back at the show and was embarrassed at how bad they sounded after hearing a tape being played in the same room. Bruce said it was like bringing a knife to a gun fight.

That's exactly what the problem with digital playback is. One tiny thing not right and it sounds like crap! Tape, on the other hand, you could drop the machine on a concrete floor, dunk the tape in a pool of water, and use bits of fencing wire to hook it all together, and the playback would still sound very pleasant. Different media, different problems ...

Frank

 

[DonH50]

Now this is getting somewhere, Don! I personally would be fascinated in what parameters, and how you tested and trends in the results obtained. Care to start a thread somewhere talking about this a bit more ...?

FRank

No, I'd be slaughtered. I have none of the data anymore, and it was 25+ years ago whilst I was in college. I also built a couple of distortion generators, one using MOSFETs and another BJT-based to emphasize different harmonic spectra. I have to believe others have done much better and more recent work. Some interesting outcomes were that THD hardly mattered below 1 to 3 % or so, but IMD of 1 % and less was readily apparent. People actually thought higher distortion sounded better in some cases... The bass "speed", i.e. rise time of the amp and speaker's output, hardly mattered but the "tail" after (often a slowly-decaying ring) was quite audible (which was better correlated to the effective damping factor of a given amp/speaker combination over frequency). Noise floor mattered in the quiet passages but SNR of 60 dB was about where most people stopped considering it a factor; the noise shape was more important than the amount. It was much harder to discern problems with music or even multiple tones (above perhaps five tones) than with 1 - 3 test tones. Long-term power supply charge storage mattered more than short-term for dyanmic peaks. And so forth. All of which will be (probably rightly so) challenged by numerous folk, especially since the data effectively exists only in my mind at this point... I believe what I remember, but certainly do not expect anybody else to believe the same.

 

[mep]

Mep,
VU meters have a typical 30dB range - IMHO you can not use them to look for dynamic range in recordings. Also many of them are peak programme meters with response factors optimized for different applications, but not to show dynamic range. BTW, I have no doubt that digital has better measured dynamic range than analog, what we should address is the very difficult to analyze perception of dynamic range of the different media by listeners. And here I agree with you - most of the time vinyl seems more "dynamic" than CD (in the abusive sense).

Those VU meters are a damn good indication of whether you have really good dynamic range in a recording or if you don't. So I do think they have something to tell you. If the meters are glued to 0 VU, you are in compression hell. If they swing from -20 to +3, there will be smile on your face.

 

[JackD201]

Simply put back to you, some people hear the roar of a mountain lion in the gentle purr of a kitten.

If I understood correctly, Amir is proposing that perhaps there is an emotional multiplier, this well-named euphonic effect, that can make very small differences seem large. Of course there is, but it varies wildly from person to person. The problem, and perhaps it's only mine, is that putting the word "index" in there, seems to suggest that these wildly varying experiences could be categorized and compared in a way that would impart some kind of meaning that would translate from one who hears the lion to one who hears the kitten. That's not going to happen. I wish I could "Give the human mind and experience some credit here," But there are at least a couple of people in this community and many in the broader audiophile community who hear sabre-toothed tigers when, to my ears, the kitten is stone dead. Euphonic index has a very nice ring to it, but regardless of what you call it, I'm going to need some verification.

YMMV

Tim

Tim,

As both of us have marketing backgrounds, I do believe you and I object more about the WAY the experiences are conveyed more than the possibility that the experiences may be valid. Being blasted by NEW! BETTER! BREAKTHROUGH! or whatever gimmicky slogan or tagline for products that have been barely changed on a daily basis make us cringe.

One thing is for sure though. There's no accounting for tastes.

Here's something that I have come across quite often that has a lot to do with how a target population tends to associate a sound with a category of products leading to very polarized reactions. SETs!!!!!!!

The general description of SETs are that they are rolled off at both extremes but have a glorious midrange. They are supposed to be dynamically limited but the midrange makes up for all that, so much so that any limitations they are supposed to have are inconsequential. Now why would SETs be categorized this way? Mainly because the greater majority of folks whove listened to SETs have done so using only moderately efficient loudspeakers. In this scenario they are pretty much accurate. Plop even a 3 watt 2a3 in a system with 105dB of efficiency and flat impedance, multi-amp them, and suddenly they no longer fit the general stereotype. Some SET lovers don't like their own amps in such systems. So we get on to tastes and Amir's EAM.

Some people enjoy the LACK of driver control! The resulting FR non-linearities, and the added distortions of SETs can be appealing (which is scientifically proven BTW) and thus be an EAM even if these are a result of equipment interaction more than they are inherent qualities of the amps themselves!

That's just one example and in audio there are so many!

Jack

 

[NorthStar]

Can't mix emotions from music with electronics.

 

[fas42]

May I be so bold to suggest a variation of the EAM, which I shall call the 95 factor. In simple terms, if a reproduction mechanism was perfect in every conceivable way, then it would have a rating of 100, out of 100 say. Nothing can be so "good", so let's drop back to a fairly arbitrary figure of 95 out of 100, representing a system's ability to reach a level of performance which allows for at least moments of a suspension of disbelief, that is, the playback sounds "real". Now if a system regularly hits and breaks through the 95 barrier, then it will have high levels of wow factor, EAM will be well above 1. Some systems may only achieve a 95 on certain recordings, or when certain elements in the listening environment are in alignment. But the big struggle for everyone in the game is to try and hit the 95 as much as possible.

Rough as guts audio may be said to have a 50 factor, something that a newbie has assembled from standard bits might be around the 80 level. Most people tuning in here are above the 90 point, and are working hard at, and many times succeeding in punching through the "magic" 95 number. The trick, of course, is to always be running at 95 or above, but it's very hard work, the closer you get to 95 and higher the harder it is, and hence all the tears and disappointment, etc.

My personal assessment of the 95 factor is consistency in getting the tweeter next to the ear disappearing. People may sneer and laugh, but so far it has been a failproof way of testing for the 95 factor, and exceptional EAM ...

Frank

 

[terryj]

nah, it's only the 1,140 th time you have told us.

no doubt not the last either I'd hazard.

1,141.

Correct.