[Steve Eddy]

Not so! At heart this is about consumerism - not being taken advantage of by unscrupulous vendors, and not being tricked into paying ten times more than necessary for audio gear. Specs are indeed relevant, and the consumer who understand this will ultimately enjoy a better audio system at a lower price than someone who does not understand.

That's all well and good if you assume that all consumers are strict utilitarians and that audio systems are nothing more than a commodity.

se

[miklorsmith]

Tubes always measure worse than SS and are preferred by many, many people.  Maybe "we" will have to come up with a global disclaimer to warn away those that may be swayed by our twisted motives.

Frankly, if we bought on specs, we'd all have receivers and Denon players.  Hey, we'd be AH!

[John Casler]

There is equipment out there that can do it. With every human voice, there is what is known as a formant. It is unique to every individual, as every one has a different size chest cavity, vocal tract, and different vocal chords. You can't disguise it by imitating another voice.  And yes, it can be isolated and measured.

Cheers

Good find.

http://www.sfu.ca/sonic-studio/handbook/Formant.html

[Daygloworange]

Thanks.

Cheers

[Daygloworange]

That's all well and good if you assume that all consumers are strict utilitarians and that audio systems are nothing more than a commodity.

Yeah, I think my gear looks bitchin' cool, dude. 

Cheers

[Ethan Winer]

John,

> If John is singing, you can "hear" that it is him.  If George or Paul is singing, then you can "easily" hear the difference. There is no measurement possible to discern between who is singing. <

Of course there is! Let's start with what difference can be heard. The main two things that vary between two singers are spectral content, which includes formants (high Q resonances) from the mouth shape, and the way notes swell and decay which is consciously controlled by the singer. One singer may tend to have a harder (faster) attack than another. Why do you think this can't be measured? If I were assigned the task of distinguishing John from Paul without listening, the first thing I'd do is run an FFT on both Wave file samples. When John makes an "ooh" vowel sound the spectrum will be different from the way Paul voices the same vowel. If one singer has an accent the oohs and aahs and ohs might change, but the formants will still be unique and valid.

> Hearing is not simply the act of receiving frequencies, amplitudes, distortions and phase relationships.  It is the process of assembling these into the perception of sound. <

I agree, but again this is irrelevant in a discussion of the ways that audio gear can alter the quality of electrical signals passing through them and how those changes can be measured.

--Ethan

[Ethan Winer]

Tubes always measure worse than SS and are preferred by many, many people.

Yes, and it's well known that adding subtle amounts of distortion can be a pleasing effect. But this is an effect, not higher fidelity! Here's a short Op-Ed piece I wrote about this for a popular recording magazine:

http://www.ethanwiner.com/edge.html

--Ethan

[ctviggen]

John,

> If John is singing, you can "hear" that it is him.  If George or Paul is singing, then you can "easily" hear the difference. There is no measurement possible to discern between who is singing. <

Of course there is! Let's start with what difference can be heard. The main two things that vary between two singers are spectral content, which includes formants (high Q resonances) from the mouth shape, and the way notes swell and decay which is consciously controlled by the singer. One singer may tend to have a harder (faster) attack than another. Why do you think this can't be measured? If I were assigned the task of distinguishing John from Paul without listening, the first thing I'd do is run an FFT on both Wave file samples. When John makes an "ooh" vowel sound the spectrum will be different from the way Paul voices the same vowel. If one singer has an accent the oohs and aahs and ohs might change, but the formants will still be unique and valid.

--Ethan

Only it's not that easy.  The formant structure is only for very small snippets of time.  I believe it's only valid for speech, and not singing.  For instance, see:

http://en.wikipedia.org/wiki/Prosody_(linguistics)

The formant structure is really good for things like voice compression, as you can examine the formant structure and build a filter that emulates the structure.  This way, you just send coefficients of the filter to the other side, which uses those coefficients.  However, singing is a different animal altogether. 

I think you'd have to use a longer time period in order to distinguish things.  I'd go on, but I'm going on vacation tomorrow and need to finalize things here.

[Steve Eddy]

Yes, and it's well known that adding subtle amounts of distortion can be a pleasing effect. But this is an effect, not higher fidelity!

Ok, but so what?

I don't know about you, but the only reason I own an audio system is ultimately for my own subjective pleasure and enjoyment. Why on earth would I want to buy something that I find gives me less pleasure and enjoyment simply because it's "higher fidelity"?

se

[Steve Eddy]

Yeah, I think my gear looks bitchin' cool, dude. 

Groooovy! 

se

[John Casler]

John,

> If John is singing, you can "hear" that it is him.  If George or Paul is singing, then you can "easily" hear the difference. There is no measurement possible to discern between who is singing. <

Of course there is! Let's start with what difference can be heard. The main two things that vary between two singers are spectral content, which includes formants (high Q resonances) from the mouth shape, and the way notes swell and decay which is consciously controlled by the singer. One singer may tend to have a harder (faster) attack than another. Why do you think this can't be measured? If I were assigned the task of distinguishing John from Paul without listening, the first thing I'd do is run an FFT on both Wave file samples. When John makes an "ooh" vowel sound the spectrum will be different from the way Paul voices the same vowel. If one singer has an accent the oohs and aahs and ohs might change, but the formants will still be unique and valid.

> Hearing is not simply the act of receiving frequencies, amplitudes, distortions and phase relationships.  It is the process of assembling these into the perception of sound. <

I agree, but again this is irrelevant in a discussion of the ways that audio gear can alter the quality of electrical signals passing through them and how those changes can be measured.

--Ethan

Hi Ethan,

Although the discussion has included "audio gear" it seems to be about "what you are hearing, versus what you are or can measure". 

And while the Formant System is interesting, and shows progress, it "IS NOT" able to approach Human Hearing.

It, with much sampling of an individual Human Voice, can tell voices apart, but in no way can it take a complex passage of two or three singers and several instruments and tell you what, and who they are they are.  And Harmonies?  Forget it.

Almost anyone with ears can do that instantly.

[Ethan Winer]

Bob,

> I think you'd have to use a longer time period in order to distinguish things. <

Okay, but who cares? The point is things that can be heard can be measured. So you use a longer gate time higher resolution FFT, and so forth.

And this is still unrelated to assessing audio fidelity of gear.

--Ethan

[Ethan Winer]

Steve,

> Ok, but so what? <

Are you really arguing that audio fidelity in gear doesn't matter? Isn't that the "fi" part of hi-fi?

I would never tell someone they shouldn't enjoy their tube amp with 5 or 10 percent distortion because it uses too little negative feedback. If they like the sound of distortion added to everything they listen to, I'm cool with it. But you can add grunge like that for a lot less than the cost of these boutique products. Two silicon diodes and a pair of well-chosen resistors can do exactly that for about 50 cents.

It seems to me the true goal of hi-fi is to hear the intentions of the recording and mixing engineers. That implies having a system that is as neutral as possible. Versus a system that adds a particular character or coloration to everything that passes through it.

--Ethan

[Dan Banquer]

It seems to me the true goal of hi-fi is to hear the intentions of the recording and mixing engineers. That implies having a system that is as neutral as possible. Versus a system that adds a particular character or coloration to everything that passes through it.

--Ethan

NOT IN HIGH END
          D.B.

[rajacat]

Our ears all have their individual hearing characteristics. How does an audio component designer compensate for all the possible variations in hearing?

Raj

[miklorsmith]

And, no system sounds like the "original"  So, which incorrect flavor do you Like?

[Dan Banquer]

Our ears all have their individual hearing characteristics. How does an audio component designer compensate for all the possible variations in hearing?

Raj

You can dial it in with one of these.

http://www.aphex.com/204.htm

              d.b.

[rajacat]

Yeah, but introducing one more component into the mix surely must introduce distortion, etc by contaminating the signal path.

[Ethan Winer]

John,

> Although the discussion has included "audio gear" it seems to be about "what you are hearing, versus what you are or can measure". <

This thread started with a post about a blatantly snake-oil device (clever little clock) winning an award. Companies that sell stuff like this, where the perception of improved sound is due entirely to the placebo effect (and comb filtering) should be ridiculed and exposed. So it wasn't me who changed the subject to questioning whether "science" is able to measure the difference between John and Paul.

--Ethan

[Ethan Winer]

NOT IN HIGH END

I'm laughing with you, brother, and crying at the same time too.