Loudspeaker Measurements Explained

[Stereoeditor]

Well, this is what makes the whole thing controversial, and to paraphrase Mr. Karsten on a recent a'gon thread on Wilson speakers, some speakers are designed - he claims - to be used with tubes, therefore said interaction is part of the design...

Okay, what "tubes"? An amplifier with a relatively low output impedance, like a VTL, or one with an enormously high impedance, like a PrimaLuna or Mystere? The speaker will measure very differently with either of these tube designs. The point of using a solid-state amplifier with a very low output impedance for the measurements is to remove the amplifier from the equation, to allow people to see the fundamental performance of the loudspeaker so that the measurement will be maximally transportable. If someone then decides to change their speaker's response by using a tube amplifier, the impedance graph I publish in every speaker review will allow them to predict how the measured frequency response will be changed.

But in the end, Mr. Karsten does have a point in at least indirectly raising the question - isn't it all about system matching?

For listening, sure, but not for measurements for the reasons I have given you.

John Atkinson
Editor, Stereophile

 

[Robh3606]

I would just repeat a comment that I have made before that the better the system the less relevant the frequency response accuracy is to the listening experience.

Hello Frank

Could you elaborate I am not sure what you mean. As I see it the frequency response accuracy is extremely relevant. If I understand you correctly you are saying a "better system" can have poor accuracy compared to say a less expensive more accurate system and sound better??

Rob

 

[Ron Party]

It seems to be a position that is not borne out by science. I believe Dr. Olive's blind testing suggests the exact opposite conclusion.

 

[NorthStar]

A simple question...

You guys are touching the most important point in Audio: System Synergy!

* Why don't loudspeaker's manufacturers recommend the correct amplifiers matching with their speakers?
{And the same for amplifier's manufacturers towards several good speakers matching with their amps...}

Would be nice if the speaker manufacturer (designer) could tell us which amp(s) they used to build their speakers, and other good matching amps with them, no?
Few do, but they are indeed just too few and far in between.

Is that mean that most speaker manufacturers aren't designing serious speakers?
I'm asking this question right here to you guys, the experts.

 

[Mark Seaton]

Well, I cannot speak for JA but one reason for testing with an amp that is insensitive to speaker impedance is for consistency over time and for the results to be unbiased. OTOH, if the amp is "sensitive" to impedance variations, the amp itself is a variable as one would presume that the interaction would not be the same for all such amps.

...The point of using a solid-state amplifier with a very low output impedance for the measurements is to remove the amplifier from the equation, to allow people to see the fundamental performance of the loudspeaker so that the measurement will be maximally transportable. If someone then decides to change their speaker's response by using a tube amplifier, the impedance graph I publish in every speaker review will allow them to predict how the measured frequency response will be changed.

John Atkinson
Editor, Stereophile

Echoing the Kal and John here, small trends in response, even when audible, are only a small part of what loudspeaker measurements capture and reveal of a speaker. The integration/hand-off of the drive units and how the speaker radiates sound into a space, along with details like sensitivity and impedance tell us much more. A low output impedance is the only way to examine the speaker in isolation, just as we measure subwoofers outdoors. It provides a common reference where you can then consider the additional effects of the listening environment and connected electronics.

The primary addition I would love to see is a few more measurements covering further off axis horizontally and vertically to show some of the strong midrange lobes common on some of the physically larger speakers available. Admittedly these are very cumbersome measurements to create. I would love to see some publication work out something with Ron from NWAA Labs or Pat Brown of Syn-Aud-Con/ETC Inc. who can easily capture such macro measurements. I won't hold my breath, but it would be interesting to consider, especially for more home theater centric designs where larger seating areas are a more significant consideration and a greater deal of pre-construction design & planning goes into the room acoustics.

 

[fas42]

Hello Frank

Could you elaborate I am not sure what you mean. As I see it the frequency response accuracy is extremely relevant. If I understand you correctly you are saying a "better system" can have poor accuracy compared to say a less expensive more accurate system and sound better??

Rob

This is a matter I have referred to a number of times, and is as Gary would say, pyschoacoustic. To rephrase, one's perception of the sound quality changes as certain intrinsic qualities other than the frequency response change, which renders the frequency response of a far lower level of importance. This is something that I discovered for myself, when investigating the effects of extensive tweaking of a system. As a simple example, at one stage I was playing with a normal Sony integrated amp, which had the usual treble and bass controls. When this amp was in a normal state it was obvious the impact of severely boosting and attenuating these two areas of the spectrum. By dint of some devious tweaking, of very short term benefit, I could get this gear to perform well above its station, very transitory high end behaviour. While in this state I could rotate these tone controls from minimum to maximum, and to my amazement found it almost impossible to audibly pick that anything had happened!

How can this be so? As mentioned above, I believe it is psychoacoustic, the mind playing tricks on you. When the sound is at a very high quality level, very low distortion, then the musical message of the recording overwhelms everything. The ear/brain tunes into listening intently to the performance and ignores elements that it considers irrelevant, allowing you to gain maximum pleasure from the experience.

There is a live music analogy: imagine listening to a very fine violin solo; you will be totally drawn into the moment (assuming such things are your bag! ), oblivious to other stimulus. On the other hand, an acquaintance showing you how far they had got in learning to play this instrument: the last thing you tune into is the musical content, you are aware of every tiny imperfection in the violin's sound, the person's foot scraping on the floor, a window in the room slighty rattling. This is equivalent to being highly sensitive to the frequency response, which is what you need to transcend. In other words, your system should always perform at the level of a fine violinist, because then minor or less important imperfections become irrelevant to the listening exerience.

Frank

 

[fas42]

I will throw in another example. Tim at one stage posted the in ear measured frequency responses of two very highly regarded headphones. The differences in the response curves of these two, measured using identical procedures, was quite staggering, certainly sufficient to consign any respectable loudspeaker designer to the sin bin. Yet these were transducers considered to be close to state of the art ...

Frank

 

[caesar]

It is an updated version of the articles on speaker measurements I published in Stereophile at the end of the 1990s:
http://www.stereophile.com/features/99/index.html
http://www.stereophile.com/features/99/index.html
http://www.stereophile.com/features/103/index.html


Ralph Karsten tends to get a bit bitter about many things

Seriously, to get repeatable speaker measurements, you need to use an amplifier with a very low output impedance, which means solid-state as far as I am concerned. If you use a tube amplifier, you get arbitary changes in response due to the interaction between the output impedance of the amplifier and the manner in which the speaker's impedance changes with frequency. If you do wish to use a tube amp with a speaker that has been reviewed and measured in Stereophile, it is a simple matter to examine how the response will be changed by that interaction.

John Atkinson
Editor, Stereophile

Stereoeditor,

Thanks for the links. I will check them out when I get some time. Also, I do hope you guys make them more prominent on your website and in the magazine. Very few people do what you do, but worse, many people do not understand what you do. So you guys gotta teach them about how you and why you differentiate yourself from the pack and add value. It's kind of like those beer commercials that constantly remind people they are drinking beer from the best ingredients on earth.

 

[ack]

It provides a common reference where you can then consider the additional effects of the listening environment and connected electronics.

Well, that's what's at the heart of this issue. Do we care about this common reference more than we care about "optimal measurements" on a per speaker basis, at least according to the speaker designer's definition thereof? Not to mention that tube guys have absolutely no clue as to how their amps might handle all these speakers being tested.

I love measurements, but I am on Karsten's side on this one...

 

[microstrip]

You guys are touching the most important point in Audio: System Synergy!

* Why don't loudspeaker's manufacturers recommend the correct amplifiers matching with their speakers?
{And the same for amplifier's manufacturers towards several good speakers matching with their amps...}

Would be nice if the speaker manufacturer (designer) could tell us which amp(s) they used to build their speakers, and other good matching amps with them, no?
Few do, but they are indeed just too few and far in between.

Is that mean that most speaker manufacturers aren't designing serious speakers?
I'm asking this question right here to you guys, the experts.

As you say, system synergy is the keyword for good sound synergy. However, system synergy can have two distinct meanings - electrical synergy or subjective sound quality synergy. I will address this last one.

Many manufacturers try to get this synergy building complete lines and selling complete systems - then the buyer can get exactly the sound that the designer wanted. However, in market terms many times thinks are not so easy - distribution agreements can make the life difficult for shops and the choice or upgrading capabilities are very limited if the client wants to change the sound type. Krell was one of these cases - they manufactured an excellent line of great speakers to complement their electronics, but it entered in conflict with their dealers compromises and interests.

Other manufacturers discretely let us know what complementary gear they use for development - you have to look at notes about factory tours, or even at join presentations at audiophile gatherings.

It can happen that a few brands are owned by the some group, and as they are mixed during the development phase they show a high synergy.

But sometimes brand matching is due only to the distribution channels, and brands having no synergy are sold together with very poor results.

 

[NorthStar]

Thanks Micro for taking the time.

* I'd take from that that it is up to the customer to do some research, and also experiment on his own for new discoveries.

'Loudspeaker's Measurements' is quite a vast and complex field, that needs to be understood in the right circumstances or context (room and amplification as well preamplification, and the material from the source too). ...Speaker cables?

 

[garylkoh]

But sometimes brand matching is due only to the distribution channels, and brands having no synergy are sold together with very poor results.

Thanks, microstrip. As a manufacturer, that is my biggest problem. If other manufacturers were friendlier, or if dealers in the same town were friendlier towards each other and cooperate to grow the pie instead of stabbing each other in the back to jump at the crumbs, the industry would be a far friendlier, and dare I say much improved place. Instead, like the ancient Chinese kung fu masters, they keep their secrets so close to their chest that they never grow beyond their individual ability.

Even when my competitors ask me about my techniques, I tell them. I tell them what I do, but may be not how I do it. It's pretty safe, because first, there will be disbelief that I will actually tell them the truth. And then they will believe that I'm trying to send them down the wrong path, and do the complete opposite. Either way, I'm still safe.

 

[NorthStar]

^ Lol. ...In a very friendly way of course.

 

[Mark Seaton]

Well, that's what's at the heart of this issue. Do we care about this common reference more than we care about "optimal measurements" on a per speaker basis, at least according to the speaker designer's definition thereof? Not to mention that tube guys have absolutely no clue as to how their amps might handle all these speakers being tested.

I love measurements, but I am on Karsten's side on this one...

I see this as two issues, not one...

Quantifying and verifying the performance of a loudspeaker should be done with a common source, and most are designed around the concept of an amplifier which produces constant input/output Voltage vs. frequency. A tube amplifier will vary from this behavior with different connected loads (speakers) and different amplifiers will have different affects on the same speaker based on the design.

The amplifier does not affect the impedance of the loudspeaker, it doesn't change the on/off axis behavior nor the crossover integration of woofer to mid or mid to tweeter. There is no reason to examine the metrics of every loudspeaker with 10 different amplifiers, nor is it practical for ongoing testing.

What would probably be more useful, and likely get plenty of audiophile panties in a bunch, would be to coordinate a separate test of a few tube amplifiers of differing design and output impedance vs. frequency. Set a base line with a stout solid state amplifier which measures up in delivering constant Voltage vs. frequency to real loudspeakers. Pick 3-5 speakers with differing impedance, low extension and number of frequency divisions (ie 2, 3 or 4-way). Now test 3-6 tube amplifiers while driving each of these loudspeakers (matching Voltage output at some chosen frequency) and compare/plot the resulting responses showing the different amplifiers on a given speaker, and then compare plot a single amplifier on the different loudspeakers.

This would at least set and clarify some expectations of how much a tube amplifier selection can impact the response of a given loudspeaker and how that correlates with the impedance curve posted of each speaker.

 

[JasonI]

Stereoeditor,

Thanks for the links. I will check them out when I get some time. Also, I do hope you guys make them more prominent on your website and in the magazine. Very few people do what you do, but worse, many people do not understand what you do. So you guys gotta teach them about how you and why you differentiate yourself from the pack and add value. It's kind of like those beer commercials that constantly remind people they are drinking beer from the best ingredients on earth.

He misposted article two in the series. This is the correct link: http://www.stereophile.com/features/100/index.html

Informative series. A little light on actual explanation of measurement details, but I suppose a math book would be required to get into the nuts and bolts.

RE: Phase Coherence and Time Alignment - Part 2

Floyd Toole concluded thusly: "The advocates of accurate waveform reproduction, implying both accurate amplitude and phase responses, are in a particularly awkward situation. In spite of the considerable engineering appeal of this concept, practical tests have yielded little evidence of listener sensitivity to this factor...the limited results lend support for the popular view that the effects of phase are clearly subordinate to amplitude response."

I feel that if other factors have been optimized—on-axis response, off-axis dispersion, absence of resonance-related problems, and good linearity—like a little bit of chicken soup, time coherence cannot hurt.

RE: Driver Resonance in Cumulative Spectral-Decay Plots - Part 2

Floyd Toole and his associate Sean Olive did considerable work on the audibility of resonances [39, 40]. It is generally held that high-Q, high-but-narrow peak resonances are less objectionable than low-Q, low-but-broad peak resonances. It is also held that dips in the amplitude response that might also be associated with resonant behavior are less audible than peaks. In my experience, the cleaner-looking a loudspeaker's CSD plot in the upper midrange and treble—above 1kHz, say—the better the chance it will receive a positive review. Loudspeakers that are praised by listeners for "good clarity," "low grain," or "excellent transparency" tend to have clean-looking CSD plots.

In the above section, Mr. Atkinson goes on to discuss panel resonance and suggests that, in most cases, box resonance is a fruit hanging high in the tree.

RE: Extended FR of Ported Designs - Part 3

Up to the middle of 1997, the vast majority of the 360 speakers I measured were reflex designs—300 models, or 83%—the designer using the port to extend the design's anechoic low-frequency performance. Yet in an actual listening room, the increased rate of low-frequency rolloff of a reflex design leads to less low-bass output than with an equivalent sealed-box design, with its 12dB/octave rollout.

...it is my considered opinion that in almost every case, the same or better bass performance could be achieved with an equivalent sealed-box alignment.

 

[NorthStar]

^ Great post Jason!