ROCK CONCERTS

[Todd W. White]

Yes, I remember that comment by said scientist on the other forum (who, by the way, pushed the AES to have me lecture at their 92nd Annual in NYC on "Vintage Professional Loudspeakers, Their Care, Renovation, and Use", which was subsequently published in their Journal).

Quite a hilarious comment, indeed!

...My, my - the things we blurt out in the heat of the moment can be quite embarassing (remember, "I never, never had..."?).

With reference to the idea that we're really FAR AHEAD of those who preceded us, my friend Dr. John K. Hilliard used to make it painfully clear that we do NOT know everything when he said, "We're just beginning to nibble around the edges of the moderately difficult."

How true.

[tomdae]

Hi bob

You are right, I'was ingnorant of the fact, that the listening is still a criteria in the designing process of speakers. What I believe, but I can't proof it, that there are some properties in the sound that is usually not measured. This could probably be measured with Sine bursts of different lenght (1ms to 1 sec, different frequencies and different levels, adding many of them in an ever changing manner). Played on altec speakers, probably these bursts would not be very undistorted, but very precise in timing and level, what makes the sound realistic. And distortion would be mostly on 2nd and 3rd order harmonics, wich would not affect the sound image too much. Other speakers with lower (sine wave)distortion probably would get confused with all that bursts and would not be able to maintain the timing and levels all together. Thats why they do not sound so rich and realistic as altec do. That's just a fantasy, I try do technically understand why my ears tell me altecs sound nice and realistic. Has enyone heard of measurements of the above mentioned kind?

[Todd W. White]

I like to point out in these discussions (the type of which I have been involved in many times) that technical specifications are not the criteria for final pronouncement as the ability of a loudspeaker to perform in a way that is both accurate soncially and pleasing to listen to. Specifications are not the end - they are contribute to the final outcome, but they are NOT the only factor in the equation, if you please.

Not long ago, I spent a day at a high-end audio show where all kinds of speakers (and other things) were being demonstrated. Now, while the listening environments there were by no means pristine, they were usable, and the attendees were able to get a good idea of what the different speakers were able to do in less-than-ideal acoustic environments (most home listening rooms are less-than-ideal).

Of all the exhibitors of loudspeakers, almost all had published specifications that were outstanding: ruler flat response from 0 Hz to infinity, high power handling capabilities, impressive SPL output, esoteric approaches to sound reproduction covering everything from the nominally plausible to the downright ridiculous...well, you get the idea.

The consensus of the people - as heard and reported by those mingling amongst the crowds - was that two speaker systems present had the "best" sound at the show (whatever "best sound" is).

One system was designed by a scientist who is hearing impaired, and, truthfully, it works quite well. He did the math correctly, and came up with a system that does a very good job. I was not slightly impressed by them - they were well behaved, performed nicely at lower SPL's, and had good realism (no real coloration). My only real negative to listening to them, other than the fact that their space-age look might not go well in some home environments, was that, after a while, they were fatiguing to listen to. I don't know why, other than the fact that I could hear the acoustical center of the sound source moving up-and-down and up-and-down all of the time between his LF and HF devices. At any rate, they were quite nice, and quite expensive.

The other speaker system that, according to those who listened to the attendees' comments as they made their way from room to room, was actually preferred over all of the others present. That system used a coaxial full-range driver, a passive crossover, and utilized a more traditional approach to enclosure design and manufacture.

According to the attendees, this speaker, though it had less-than-perfect polar plots, less-than-perfect frequency response curves, a limited power handling capacity (75 watts, AES) - though it was extremely efficient (4.87%), sounded more natural, more realistic, and was more impressive to listen to than all of the others being deomonstrated that day, even more than the one designed by the highly regarded scientist who has an earned degree in physics and acoustics!

The most telling differences between that system and all the others came during the playing of a recording that had been made in a "nearly live" environment - it was a symphony orchestra recording live in a huge studio designed for the purpose, and used minimal mixrophones. The mix was very simple. It was as close to a recording of a live performance as you could get.

During one particular passage, the composer used a large 52" bass drum to accentuate the piece. It wasn't hit hard by the percussionist - only firmly - and it was not over-eq'd to sound bigger than it was. It was just a recording of a 52" bass drum.

Operating at a moderate listening level, this passage was played for the attendees. Every time the music got to the part where the bass drummer did his thing, the listeners, who had been favorably responding to the performance of this system up to that point, upon hearing the bass drum accentuate the musical passage, all broke out with a wide grin, their eyes widening with astonished pleasure.

Now, I've said all of that to say this:

YES it is important that a loudspeaker be capable of reproducing at sound levels that are true-to-life,

and

YES it is important that a loudspeaker be free from frequency response characteristics that alter the true nature of the sound they are being asked to reproduce,

and

YES it is important that we reckognize deficiencies in design and manufacture and improve speakers as best we can,

BUT -

It really all comes down to WHAT DOES IT SOUND LIKE?

If, in the final analysis, the sound the listener hears is NATURAL and exhibits true REALISM, then, the measured performance parameters notwithstanding, the speaker is successfully doing what it is designed to do.

To me, THAT is the final measurement criteria: DOES IT SOUND REAL?

If it takes perfect spec's and huge SPL abilities to do so, then that's fine.

If it can be accomplished on other ways using new methods, or using older ones, then so be it.

[amh]

tomdee:

I think what you refer to is called impulse response. The ability of a driver to not substantially overshoot or undershoot an electrical impulse of much shorter duration than it's acoustic output. Yes it can be tested for, and graphed.

Quoting from an article in Dec. 2004 AudioXpress by Mark McKenzie.

"In impulse testing the drive signal is a short duration, positive-going electrical pulse lasting a much shorter time than the acoustic output from the driver. The driver begins accelerating during the applied signal, continues to deflect in the same direction after the signal stops, and then once the spring force of the spider and surround equals the inertial energy of the cone structure, snaps back.

Electromechanical drivers always overshoot their rest position and produce a negative polarity acoustic output. For want of a better term these large magnitude acoustic outputs are called spikes. A perfect loudspeaker in a sealed cabinet or infinite baffle will show three spikes: a first positive, first negative, and second positive. These three spikes are considered the onset response. Each successive spike ought to be much less in magnitude than the preceding one.

After the first series of spikes, the driver continues to vibrate until finally coming to rest at the beginning position. This overshoot and later decay is caused by the combination of the motor system resonance and a complex variety of cone and surround resonant structures.

It is the magnitude ratio of first positive spike to first negative overshoot that makes a driver desirable."

Mark goes on to explain how the first positive to negative ratio is audible, and gives the example of how a tweeter with a larger second negative spike than it's first positive going spike can sound edgy and hard when reproducing cymbals.

Back to the original point. This attribute of a driver is seldom, if ever, listed in it's spec sheet, yet can clearly affect performance.

I hope I have not stepped on Marks toes by quoting him here, or broken any site protocall. He wrote a very informative article.

[joyspring]

Back again folks, sorry about the delay!

Todd:

While I agree that tech specs are not the final arbiter of how a loudspeaker system will perform, psychoacoustic research has codified a few criteria that is the minimum requirement for high performance.

For loudspeakers, flat on-axis frequency response is a minimum requirement.

Flat power response is now considered a minimum requirement now that a lot more research has been conducted on the loudspeaker system / room interface.

Now for the listening test example that you mentioned in your post:

Great, the 604-8K-derived made an impressive showing against a modern, well-engineered system.

I can agree that the Iconic did sound `more impressive' than the other system but assuming that the other system had flat power response (which the 604-8k has only above its crossover frequency), I doubt that the Iconic did sound `more natural, more realistic'; physics dictates that it can only sound uncoloured on-axis.

I have several UREI 811 / 811A monitor loudspeakers, a pair of 604-8K and 605As and I can assure you that ALL of these systems sound extremely coloured (and therefore un-natural and un-realistic off-axis. The `horn-in-the-centre' coaxial arrangement cannot assure a smooth transition between LF and HF drivers in the small physical space.

As an aside, the non-technically-savvy producer I now work with has complained about my UREI 811As because they only `sound right where you (me, the engineer) are sitting'. That's another reason why I had to purchase my Mackie HR824s.

Now to address your ultimate question: `WHAT DOES IT SOUND LIKE?'

For recording and sound design professionals, we want - no, DEMAND - flat frequency and power response; producers/directors need the assurance that the end-product translates to so many end-user systems as possible. Meeting those minimum requirements is the only means of ensuring this.

Now for the end-users, well, this has many engineers perplexed. Who can explain the resurgence of tube (valve) electronics and the insane preference for vinyl phonograph LPs? The latter in particular is much further from `natural' and `realistic' than 44.1 kHz / 16-bit digital media.

Perhaps end-users do not really want reality or naturalness... Perhaps they like noise and harmonics, frequency response aberrations, poor off-axis performance.

Perhaps they want a 3-6 dB rise in the one-octave band between 1.5 - 3 kHz (that adds apparent `impact')?

Remember that flat frequency response is dull, unassuming, reserved. Something that deviates from this (which most reputable high-end loudspeaker systems strive to attain) will definitely catch attention -- maybe even put a smile on one's face.

Lastly, I'm convinced that many end-users - especially audiophiles - listen with their eyes and make decisions based on faith. They cannot accept that small, inexpensive, modern technology can (and does) sound more natural and more realistic than grand-dad's old `Big Iron'. Harleys and old-American hot-rods are a perfect analogy.

One antidote is to convince these Luddites to go see live shows and purchase more music media than wasting money tweaking audio anachronisms.

Another is to significantly improve math, science, music and arts education in the US, all of which are severely deficient compared to the rest of the industrialised world. More important is to change the typical American `faith-based' disdain towards these subjects, which is baffling in such a technology-driven society. American's poor quantitative and analytical abilities are not only apparent in its loony audiophile culture but more seriously in its utterly debt-ridden consumer culture and its jingoistic and terribly militaristic foreign policy.

BobR

more on listening evaluations in a later post...

[joyspring]

Scotty J:

How come some gear, particularly tube amps, are noticeably rolled off in the high frequencies when measured on a scope, yet, sound so extended through speakers when playing music, despite what the measurements say?

<font size="2" face="Verdana, Helvetica, sans-serif">Possibly the addition of harmonics, which I've measured to be so high as 2-4% THD in some of the more ridiculous single-ended Class A designs?

Conversely, I have heard certain solid state amps, that have measured frequency response to 50K, and make some speakers sound gated and band limited in the high frequency!

<font size="2" face="Verdana, Helvetica, sans-serif">As I mentioned in a previous post, flat frequency response sounds unassuming and reserved, unlike a SET amp producing high amounts of harmonic components.

Also, an amplifier's ability to competently handle reactive components of a loudspeaker system load will also affect frequency response as well. The majority of quality, modern solid-state amplifiers do not suffer from this though it appears to afflict many tube amps (and many single-ended solid-state amps!), of which many - if not most - are incompetently designed.

BobR

[joyspring]

homme_de_terre and Todd:

Here's what he said about "sound": "I try and make accurate loudspeakers not 'great sounding ones' ".

<font size="2" face="Verdana, Helvetica, sans-serif">That scientist is entirely correct; isn't that the noble goal of `high-fidelity' or `faithfulness to the original'?

The creative process and the endeavour to make `great' sound ends at the final mix and mastering; consumer playback of the end-product should accurately reproduce the efforts of the artists, producer and engineer.

We certainly have not achieved that goal quite yet but we ought not corrupt the process...

I just doubt that psycho-acoustics with regard to sound reproduction is fully understood yet.

<font size="2" face="Verdana, Helvetica, sans-serif">Perhaps, but we know that flat frequency response and flat power response are essential, minimum requirements.

I always start feeling uneasy when a vendor attempts telling me what is best for me and why the competition's offers are inferior.

<font size="2" face="Verdana, Helvetica, sans-serif">When a vendor uses sound scientific methodology to cut down a competitor (i.e. - non-flat frequency response), that is entirely justified. The vendor is doing consumers a favour at that point.

As to scientists/engineers, my experience is that some of them tend to behave rather hubristically when their methodologies or theories are being questioned by "profane" people.

<font size="2" face="Verdana, Helvetica, sans-serif">No, we defend our assertions using facts gleaned from research.

Designing audio equipment is not art, it is science, or as I quoted Floyd Toole in a previous post, `Science in the Service of Art'.

With reference to the idea that we're really FAR AHEAD of those who preceded us, my friend Dr. John K. Hilliard used to make it painfully clear that we do NOT know everything when he said, "We're just beginning to nibble around the edges of the moderately difficult."

<font size="2" face="Verdana, Helvetica, sans-serif">On the contrary, we ARE incredibly far ahead in many aspects of audio electronics and media than we were even twenty years ago: witness 96 kHz / 24-bit digital audio which transcends the theoretical limits of human hearing, line-level analogue components with -110 dB noise floors, etc.

On the other hand, Hilliard's assertion is correct in other areas, particularly DSP and - more relevant to this forum - loudspeaker interface to acoustic environments.

So the big question: How to tackle these issues?

Improving and standardising testing methodology constitutes the foundation.

Floyd Toole's recommendation to codify loudspeaker testing procedures is an excellent starting point.

First off, sighted undisciplined comparison/evaluations that appear to be the norm (as in the example Todd provided) are out. Only double-blind testing can provide statistically meaningful results.

For example, JBL/Harmon uses a computer-controlled, pneumatically-operated `shuffler' to switch loudspeaker systems in the same position behind an acoustically-transparent screen. Different systems can be switched by the listener within three seconds! This should immediately become an industry-standard.

And preliminary results of Toole's research have only served to reinforce the results of previous research; listeners prefer flat frequency response / flat power response loudspeaker systems by a large, statistically significant margin. While controlling a user's acoustic environment is out of a vendor's control, assuring a flat power response is still the best attempt at achieving a good loudspeaker system / room interface.

BobR

[joyspring]

tomdee / AMH:

What I believe, but I can't proof it, that there are some properties in the sound that is usually not measured.

<font size="2" face="Verdana, Helvetica, sans-serif">If so, then it ought to be obvious that no vendor could then have mastery of such properties.

I think what you refer to is called impulse response. The ability of a driver to not substantially overshoot or undershoot an electrical impulse of much shorter duration than it's acoustic output. Yes it can be tested for, and graphed....

This attribute of a driver is seldom, if ever, listed in it's spec sheet, yet can clearly affect performance.

<font size="2" face="Verdana, Helvetica, sans-serif">Actually, quality manufacturers DO provide waterfall plots (Vifa is one, if I remember correctly).

Which also brings up yet another issue: can a 15" lightweight paper cone be both resonant and ring-free up to 1500 hz?

There was a reason for Tannoy's Girdacoustic cones (before they switched to plastic) and the fact that most loudspeaker manufacturers veered towards three-way systems with sub-300 hz crossover frequencies...

BobR

[bfish]

Guys, this has been a very thought-provoking thread! As it seems to have turned towards testing and standards, I'd like to throw out an idea to see how it flies;

Since various standards and test methods used by manufacturers lead to difficulty of accurate comparison of similar products, what if?:

An independant firm, with no ties to any manufacturer, provided fee-based testing facilities and services, using standards set by the AES, and agreed upon by participating manufacturers. I believe a well-equipped and staffed dedicated testing facility could save manufacturers time and money over having to provide their own facilities and staff, and provide the end-user with more valid and easy-to-use data for comparison purposes than is currently available, by being standardized. The makers of junk would be the only ones to find fault with such a system, though if voluntary and fee-based, they wouldn't have to participate. Such a system could lead to competitive research, which is a factor behind innovation, and the end result should be better, (and potentially less-expensive) products. Sort of like UL testing, only with different criterion.

The availabilty of standardized, reliable, and extensive test data might leave many of us that like to debate the various merits of our tools (or toys) with more time on our hands. Perhaps we could find a more rewarding pastime then,... like,...oh,say.....listening to music.

[Todd W. White]

The AES established measurment standards some time back, and Altec used them, but most others did not, probably due to the fact that their marketing departments wouldn't stand for it.

The TEF machine is THE way to do this, but few will use it...