Jump to content

What Measurements can we hear? Can we measure sound quality?


Ar9Jim

Recommended Posts

  • 1 month later...

The measurement of 'quality' requires a solid, agreed upon definition of what 'quality' is. That is not possible. We can measure certain aspects of quality but because quality itself lacks certain definition we cannot agree what such a measurement would be.

 

In the best selling philosophy book of all time, Zen And The Art of Motorcycle Maintenance, author Robert Pirsig posits that 'quality' is the interaction between subject (sound) and object (the listener). That interaction is necessarily different for every listener simply because no two listeners are exactly alike. Therefore 'quality' remains an abstract without an agreed upon interaction. It lacks control except by the individual.

  • Thank You 6
Link to comment
Share on other sites

Our guide for quality does have a baseline in audio - fidelity. For example, if we knew for certain that we owned a power amp that was ideally faithful in reproducing the audio signal into a speaker without distortion or coloration, then we would be listening to the highest quality component we could ask for.

 

The reality is that power amps are not ideal, they distort into a complex load, like a loudspeaker, in various degrees. The 'tube sound' for instance is, like I have stated many times before, is the tendency for the amp to resist changes in output impedance due to complex resistance changes in the speaker. Pure SS amps struggle with this, and can clip quite easily. Tube amps struggle with high output impedance, thus causing microphonic effects and speaker cone ringing.

 

There's a lot to know about audio. There's a lot of science, experiments and engineering lessons that we all have access to. What I don't like to see is people sticking to their own way of thinking simply because it's what they are used to thinking about, then offering up a cop-out like 'we all have our own opinion' or 'everyone hears differently'. While those are true to some extent, that shouldn't be the end of the discussion. How can you ever know for certain your way of thinking is infallible? Isn't it reasonable to support your views with objective, verifiable measurements and formulas? We should strive for knowledge with the resource of this forum, not simply yak to each other how we think audio should be. Because really, if there's no sure way to arrive at quality, and fidelity, and therefore ultimate carefree gear, then we're just wasting time talking about it all.

 

"It should sound like it isn't there" - BillD

 

When you no longer think about your components when you listen to music, you have arrived.

  • Thank You 10
  • That Rocks 1
Link to comment
Share on other sites

17 minutes ago, Nahash5150 said:

Our guide for quality does have a baseline in audio - fidelity. For example, if we knew for certain that we owned a power amp that was ideally faithful in reproducing the audio signal into a speaker without distortion or coloration, then we would be listening to the highest quality component we could ask for.

 

The reality is that power amps are not ideal, they distort into a complex load, like a loudspeaker, in various degrees. The 'tube sound' for instance is, like I have stated many times before, is the tendency for the amp to resist changes in output impedance due to complex resistance changes in the speaker. Pure SS amps struggle with this, and can clip quite easily. Tube amps struggle with high output impedance, thus causing microphonic effects and speaker cone ringing.

 

There's a lot to know about audio. There's a lot of science, experiments and engineering lessons that we all have access to. What I don't like to see is people sticking to their own way of thinking simply because it's what they are used to thinking about, then offering up a cop-out like 'we all have our own opinion' or 'everyone hears differently'. While those are true to some extent, that shouldn't be the end of the discussion. How can you ever know for certain your way of thinking is infallible? Isn't it reasonable to support your views with objective, verifiable measurements and formulas? We should strive for knowledge with the resource of this forum, not simply yak to each other how we think audio should be. Because really, if there's no sure way to arrive at quality, and fidelity, and therefore ultimate carefree gear, then we're just wasting time talking about it all.

 

"It should sound like it isn't there" - BillD

 

When you no longer think about your components when you listen to music, you have arrived.

When I don't know how to explain my observations with measurements, I post them here in hopes that someone can enlighten me.  Often I am tutored by members here and find objective ways to describe my experience. Sometimes even the brilliant people here are still searching for objective explanations. Sometimes I have to rely on my senses in lieu of currently unknown objective methods of measurement. In those cases I state my own opinion and it isn't a cop out.  I welcome constructive, objective comments. I haven't seen any indication that anyone here claims to be infallible or blindly refuses to apply objective methods when they are known.  

  • Thank You 3
Link to comment
Share on other sites

28 minutes ago, iamjohngalt said:

When I don't know how to explain my observations with measurements, I post them here in hopes that someone can enlighten me.  Often I am tutored by members here and find objective ways to describe my experience. Sometimes even the brilliant people here are still searching for objective explanations. Sometimes I have to rely on my senses in lieu of currently unknown objective methods of measurement. In those cases I state my own opinion and it isn't a cop out.  I welcome constructive, objective comments.

 

If what I said doesn't apply to you, then there's no reason to defend yourself. 'What I don't like to see...' is the key phrase in my post.

 

Quote

I haven't seen any indication that anyone here claims to be infallible or blindly refuses to apply objective methods when they are known.  

 

Good, I'm glad you haven't. :D

 

@UncleMeat I don't think anyone has ever disputed that transistors and tubes work differently. But Bob Carver already proved that a SS amp can be designed to sound like a tube amp. It was one of his claims to fame. He did it with null testing as well, so it's pretty well established that an amp as a whole, not just transistor vs tube, can emulate other topologies.

 

Link to comment
Share on other sites

Vacuum tube amps, due to their high output impedance making them current sources, are affected drastically by the impedance curve of the load/loudspeaker. Solid state amplifiers, especially the more massively parallel ones with large numbers of output devices, appear as being close to a pure voltage source. They are nearly impervious to the changing load impedances of a loudspeaker. Some early SS amplifiers of poor design also behaved in unexpected ways, such as breaking into oscillation, with certain loads.

  • Thank You 1
Link to comment
Share on other sites

On 3/6/2018 at 4:36 PM, basspig said:

Vacuum tube amps, due to their high output impedance making them current sources, are affected drastically by the impedance curve of the load/loudspeaker. Solid state amplifiers, especially the more massively parallel ones with large numbers of output devices, appear as being close to a pure voltage source. They are nearly impervious to the changing load impedances of a loudspeaker. Some early SS amplifiers of poor design also behaved in unexpected ways, such as breaking into oscillation, with certain loads.

 

Actually, the opposite is true. Current sources are used specifically for varying impedances and like to see a near short for the best performance. A voltage source struggles into low impedance because of the voltage divider effect. The problem with the tube amp current source is that it has a specific impedance reflection, depending on the load (an ideal current source has infinite internal resistance). This can cause some amplitude distortion effects. Voltage sources, via transistors, perform rather poorly into a changing load, which is why well designed feedback in necessary to tame such problems as Miller Effect, output capacitance and Early Effect, and most especially clipping due to current saturation.

  • Thank You 3
Link to comment
Share on other sites

  • 11 months later...

I think you have it backwards. A voltage source will (theoretically) try to deliver infinite current into a dead short, whereas a current source will try to maintain the current flow at the same value in to a range of load impedances.

 

An ideal voltage source has zero internal resistance so that changes in external load resistance will not change the voltage supplied.

 

An ideal current source has infinite resistance so that changes in external load resistance will not change the current supplied.

 

A well designed transistor amplifier will be relatively unaffected by the load, within the limits of its design. Such an amplifier has some commonalities with the pass bank of a regulated power supply. Of course, some feedback is necessary to maintain the regulation of a DC supply, and the same holds true for a circuit used in audio amplification.

Link to comment
Share on other sites

1 hour ago, basspig said:

I think you have it backwards. A voltage source will (theoretically) try to deliver infinite current into a dead short, whereas a current source will try to maintain the current flow at the same value in to a range of load impedances.

 

An ideal voltage source has zero internal resistance so that changes in external load resistance will not change the voltage supplied.

 

An ideal current source has infinite resistance so that changes in external load resistance will not change the current supplied.

 

A well designed transistor amplifier will be relatively unaffected by the load, within the limits of its design. Such an amplifier has some commonalities with the pass bank of a regulated power supply. Of course, some feedback is necessary to maintain the regulation of a DC supply, and the same holds true for a circuit used in audio amplification.

 

Apples and oranges here...

 

Voltage sources in practice struggle with low impedance precisely because of the current demand. This is why coupling devices with voltage sources like to see a high impedance (x10 ideally) so as to not cause distortion due to the voltage divider effect, as well as wasted heat, caused by undue current drain. Low impedance speakers tend to kill transistor amplifiers - this isn't my idea, it's a fact in all of the audio world. Tube amplifiers, because of their high output impedance, and thus, more virtual current sourcing ability, tend to not mind low impedance loads. Again, this is well known. High sensitivity speakers, for example, tend to be a very difficult, near short loads, and very often require a coupling transformer in typologies such as PA systems and live concerts when driven by transistor amplifiers.

 

Voltage source amplifiers rated down to 2 ohm continuous are quite rare. However, quite common for classic tube amplifiers.

 

My response was to your initial comment:

 

On 3/6/2018 at 4:36 PM, basspig said:

Vacuum tube amps, due to their high output impedance making them current sources, are affected drastically by the impedance curve of the load/loudspeaker

 

That's actually not true. One of the benefits of high output impedance is indeed useful for tracking a varying load impedance. This is why differential amplifiers benefit from current sourced tail current - they are close to immune to the varying changes in the transistors' internal impedances and load current. Current sources are used primary in all of electronics for precisely the problem of varying load impedances, especially non-linear impedances and high frequency impedance changes (like noise).

 

An op-amp is as close to an ideal voltage source amplifer that we have. However, they have severe current sourcing limitations due to power consumption - often no more than a few milli-amps. It's relatively easy, thanks to physics, to create a voltage. Not so easy to source current though (think alkaline batteries and such). Voltage sources are simply not known to be stable into low impedance. A great amount of materials and components and design needs to be employed to reach a near ideal voltage source for high power applications. In fact, it's the challenge of all challenges when designing an amplifier (think mag coils, tracking downconverter, class D, etc).

 

Feedback is necessary for linear output in a mulitport, multistage system. Has nothing to do with DC supply regulation, unless you're talking about power supply feedback, which should not be confused with signal feedback.

 

On 3/6/2018 at 4:36 PM, basspig said:

Solid state amplifiers, especially the more massively parallel ones with large numbers of output devices, appear as being close to a pure voltage source. They are nearly impervious to the changing load impedances of a loudspeaker.

 

I beg to differ. @RichP714 posted an article on transconductance that should help.

 

 

I talk to Bob about this all the time. Pure voltage source amplifiers need to be extremely expensive to handle a loudspeaker load at high power. They are wildly inefficient as well (total waste of power). The mag coil power supply design as well as current feedback transfer function design aid a voltage source amplifier with incredible results. It's no mistake that these designs imitate some of the virtues of tube amplifier driving characteristics, especially their higher output impedance.

 

Suffice to say, a high power voltage source amplifier requires an immense, wasteful power supply to remain linear, and that is after we get over the many shortcomings of emitter followers driving a very complex reactive load. "Driving your car with chopsticks" is such a good analogy.

  • Thank You 1
Link to comment
Share on other sites

I disagree with much of what Greg says.  With voltage negative feedback, solid state amps perform close to ideal voltage sources.  When they run out of linear region, distortion goes up, and they become something else.  You can construct an amp that monitors load current and use that as feedback if you want a constant current source type of amp.  That design can also run out of linear region and no longer be an amp.  It depends on series or parallel speaker resonances.  Amp efficiency is an implementation issue.

 

I always thought all I had to do was get a great tube amp, and I would be in sonic bliss.  However, with the introduction of solid state amps, it was discovered that speaker systems had to change to match SS amp characteristics.  Speaker systems of tube and SS eras match amps of each era.  Much of this is how speaker impedance behaves at resonances.

 

  • Thank You 1
Link to comment
Share on other sites

3 minutes ago, jvandyke_texas said:

I disagree with much of what Greg says.  With voltage negative feedback, solid state amps perform close to ideal voltage sources.  When they run out of linear region, distortion goes up, and they become something else.  You can construct an amp that monitors load current and use that as feedback if you want a constant current source type of amp.  That design can also run out of linear region and no longer be an amp.  It depends on series or parallel speaker resonances.  Amp efficiency is an implementation issue.

 

I always thought all I had to do was get a great tube amp, and I would be in sonic bliss.  However, with the introduction of solid state amps, it was discovered that speaker systems had to change to match SS amp characteristics.  Speaker systems of tube and SS eras match amps of each era.  Much of this is how speaker impedance behaves at resonances.

 

 

My point was not that tube amps are superior, because we all know that tube amps as a design have a lot of flaws. However, that does not mean they don't have good/useful characteristics.

Link to comment
Share on other sites

It's clear that humans are not necessarily interested in accurate sonic reproduction.  If we were, tone controls would be adjusted to compensate for room acoustics at flat response.  Bass would never be turned up to 8 and Basspig would have a 250W stereo, not 20kW with an auxiliary breaker box.  There would be no Gundry dip.  It's not hard to extrapolate that there's something about tube powered amps people like that's not necessarily related to sonic accuracy.

  • Thank You 2
Link to comment
Share on other sites

20 hours ago, jvandyke_texas said:

It's clear that humans are not necessarily interested in accurate sonic reproduction.  If we were, tone controls would be adjusted to compensate for room acoustics at flat response.  Bass would never be turned up to 8 and Basspig would have a 250W stereo, not 20kW with an auxiliary breaker box.  There would be no Gundry dip.  It's not hard to extrapolate that there's something about tube powered amps people like that's not necessarily related to sonic accuracy.

 

Agreed.

 

However, pure voltage sources are not a knock-down drag-out solution to the loudspeaker. There is no prevailing opinion or evidence that I am aware of to suggest that such a design is practically obtainable (not to say that brilliant things can not be done with transistors and feedback...but to say black and white, ideal voltage is the answer is not a design undertaking I'd shoot for). Voltage source amplifiers have a long history of problems (hear hear Douglas Self).

 

Requiring 20kw of power for blissful listening experience is only in support of my point - you could do the same with 20watts of tube amps.

 

Kidding...  :D

 

  • Thank You 1
Link to comment
Share on other sites

I read briefly through the whitepaper and I get your point, but I disagree that a good voltage source amplifier is large and expensive. And with class G/H designs, efficiency is pretty good nowadays. Class A is the pig in the room as far as efficiency, and for that, high power output is going to be costly in terms of cooling and power consumption.

 

It's been known for over 50 years that speakers designed during the tube era sound 'dead' on SS amps. It's as you say, because the tube amp will increase its voltage output at the resonance of the speaker and maintain almost constant power to the load that is varying in impedance across the frequency band.

 

Modern speakers are designed to work with modern amplifiers though. I was an all tube guy for many years. I held out until the late 1970s with my home brew tube monoblocks. Nothing sounded fatter, played as loud, or seemed as tolerant of overdrive as my tube amps. My first SS amp was a Fisher TX200. It crackled when it ran out of gas. Years later, someone loaned me a 40wpc Dynaco SS amp. Same thing. Seemed to lack bottom end. In the late 1970s, I picked up a Marantz 2270. For sure, this should have more bass. It didn't. It sounded grainy, thin and brittle and when driven hard, it would oscillate and that would manifest as audible clicks or chirps in the tweeters.

 

A year later, I brute-forced it and got a PL D-500. That broke the barrier, but it wasn't perfect. It wasn't until 1984, when the MOSFET amplifiers came along, and would finally replicate the clip tolerance of my tube amps, while offering much wider bandwidth and power.

 

Now those Lowther speakers are interesting, but they come with a great many (and ridiculous) compromises. The size of the horns, just to get down to 50Hz, and the problems of intermod using a single driver across the whole band. Well, they're probably okay if playing string quartet music, but pipe organ, or marching band music? I don't think so.

 

My experience with seeking out amplifiers for reproducing low frequencies has taught me that for power efficiency and woofer control with exceedingly low distortion, a transistor amp is the way to go. Nowadays, I run industrial amplifiers which can work all day into 1Ω loads without strain and damping factors in excess of 2000. And I do get Nelson's point about overdamping. I did lose some bass by going to the much higher gauge wire. But the bass I have now is more accurate without the boom.

 

If I were to try to recreate my present system with vacuum tubes, I would need 15 tons of air conditioning capacity and a direct feed from Niagara Falls. I simply could not even afford to turn on the system for one hour.

 

The present system is clean, quiet, cool-running and relatively maintenance-free. I could never see myself going back to vacuum tubes.

  • Thank You 2
Link to comment
Share on other sites

Do I have something backwards here, and don't fake it if you don't know.

My thinking was at resonance, the speaker cone has maximum excursion, and you are getting much sound produced.  So you don't want to dump energy into that and create more peak in the response if it's treble.  But I thought in the old ported enclosure woofers, they would detune resonance and damping so that it would extend bass response with the Q peak beyond lower cut off.  So you probably do want to dump energy there.  That just isn't going to work well with a SS amp if resonance is parallel and high impedance.

 

Maybe contemporary ported boxes are designed to have series resonance to do the same trick.  Is that what's in the Infinity Kappa 9 amp killers?

 

Link to comment
Share on other sites

49 minutes ago, jvandyke_texas said:

Do I have something backwards here, and don't fake it if you don't know.

My thinking was at resonance, the speaker cone has maximum excursion, and you are getting much sound produced.  So you don't want to dump energy into that and create more peak in the response if it's treble.  But I thought in the old ported enclosure woofers, they would detune resonance and damping so that it would extend bass response with the Q peak beyond lower cut off.  So you probably do want to dump energy there.  That just isn't going to work well with a SS amp if resonance is parallel and high impedance.

 

Maybe contemporary ported boxes are designed to have series resonance to do the same trick.  Is that what's in the Infinity Kappa 9 amp killers?

 

 

Kappa 9 is not ported... 

Link to comment
Share on other sites

1 hour ago, jvandyke_texas said:

You might do a tube top and SS bottom.

Giving fashion advice? LOL :D

 

 

 

But as a comic in all seriousness, Tube highs and SS lows is my bucket list stereo dream. I'm trying to simulate that with M-1.0t highs and M-1.5t lows!

  • Haha 1
Link to comment
Share on other sites

Rereading Greg's posts, I understand them and now agree with much of them.  Yes, inside the SS amp, it's not a good voltage source, does not have low enough impedance, and has efficiency issues.  Improvement on these problems has been the subject of amp development for the past 50 years.  However, with negative feedback, outside the black box amp, the load has the illusion of being driven with a near zero impedance source.  That is, until the internals of the amp operate out of the linear region, and can no longer respond to feedback to keep distortion low.  Then it is not longer an amp.

Link to comment
Share on other sites

11 hours ago, jvandyke_texas said:

You might do a tube top and SS bottom.

This statement here leaves the thoughts wide open, although I understand it's technical meaning....grin.  On a lighter note, every time I see SS, I think Stainless Steel instead of Solid State........:D

  • Thank You 1
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...