Are measurements the last word?

[ttruman]

I have read a lot lately about what makes a speaker sound good and the best ways to achieve these goals and a lot of the dependence seemed to focus around the outcome of measurement data (the discussions could get very intense depending on the views of the individual!). After reading several articles here and around I began to question my own seemingly inherit dependence on technical measurements and the importance of them. A couple of questions suddenly popped up in mind. Do I depend on them too much? Can they really tell us how good the loudspeaker will ultimately sound? Do perfect technical measurements mean my loud speaker will sound better than a non-technically perfect measured speaker given the same environmental and signal conditions? What is a technically perfect speaker? Well I thought about this for some and I am still thinking about it while I type this, my intent is not too necessarily to answer those question with any definitive answers (in some cases I don't think there are any) but offer my opinions and maybe spark some discussion.

To set the premise here let’s assume that you are attending a live classical event; The Mozart tribute quartet, a highly popular ensemble. You have managed the purchase price of a front and center seat. While you are enjoying the live experience two microphones, positioned very close to the quartet are recording the performance on state of the art equipment.

The next day you return to hear the Mozart tribute quartet encore. You have managed to purchase a ticket for the same seat except this time you see something peculiar. Two apparently very high end loudspeakers positioned where the quartet would be. Confused you open your brochure which explains that the encore is to be played on the most accurate sounding speakers in the world. The brochure goes on to explain the technical specifications. Every conceivable detail is expressed in terms of numbers and graphs. You’re into to high end audio and the specs are impressive indeed. In anticipation of the performance you settle in to your seat, close your eyes and it begins.

What is the outcome here?

Did you hear a live performance or the interpretation of one? You heard the interpretation of a live experience and you could tell it was not live. Why? Even if the recording did not contain the acoustic signature of the hall which would compound it’s self on playback there is one simple fact that remains; a loudspeaker is not an exact representation of the musical instruments that were playing. It is a physically different thing. A well designed loudspeaker can accurately reproduce sound but it does not produce sound mechanically the same way or have the same physical shape and acoustical properties of a given musical instrument and our ears can tell the difference. In other words to exactly duplicate the sound physics of a violin you must play a violin. Additionally there were two point sources per instrument where there was once one. (Carver new this and created the sonic hologram generator c-9 to restore a single source point to recordings. I own one and it does create some interesting results.)



Was the experience enjoyable? Yes it was. Although a speaker is not capable of fooling our ears into believing that there was a live performance happening it can be a captivating, enjoyable and sometimes even a strongly emotional experience. Why?
Loudspeakers are artistic representations musical instruments that allow us to experience music in a completely different and unique way that we find enjoyable.

(The exception , in my opinion, is that of a live concert recording, which was originally heard through loudspeakers, played through similar speakers especially if the performance was lip synced and used recorded material would probably sound much like that of the original because the original live experience was played through loudspeakers.)

Ok so back to the original questions the prompted me to write this. What is a technically perfect speaker? It is a loudspeaker that does exhibit any acoustical signature upon the original musical signal. And it just occurred to me that this speaker is not possible (yet). But in my opinion a loudspeaker is an artistic representation that allows us to enjoy music in different way and this is not a bad thing. Can measurements really tell us how good the loudspeaker will ultimately sound? I don’t think so. But they do tell us what we can expect to hear but they are no means a definitive representation of how enjoyable a loudspeaker will be. They help us to quantify our listening experience. Over the years standards have been accepted to which we follow and I am sure they will continue to advance as we learn more about how we perceive sound and the mechanical physics we will use to create it.



Tim

[Cal Weldon]

Originally posted by ttruman
Can they really tell us how good the loudspeaker will ultimately sound?

No, they can show you how it looks. They can tell you that with the source, mics and other measuring equipment you are using that you have achieved but one stage of a goal.

[i]Do perfect technical measurements mean my loud speaker will sound better than a non-technically perfect measured speaker [/B]

Absolutely not. What can be measured by the equipment versus what is being heard by the ear is different and so subjective that a ruler flat response IMHO is a good point from which you then start voicing your speakers.

[i]What is a technically perfect speaker? [/B]

I don't now. Subjectively I think it's a speaker that transports you to another place when you close your eyes.

[Svante]

Well, measurements can definitely point in the wrong directions. There is not one single measurement that can tell the whole story, several has to be made to get a reasonably complete picture of how the loudspeaker sounds.

On the other hand, only listening is also troublesome. IMO the "ears have to be calibrated" frequently in order to make them reasonably reliable. With this I mean listenling to live acoustic music and trying to remember the timbre of the sound. This should be the target for the loudspeaker design; to reproduce that acoustic performance at home.

I think the best method is compbining the two, and when they don't match, find the error. The error could for example be the wrong choice of measurement method or having the ears calibration twisted by listening too much to other boomy loudspeakers. Usually there is an error like these, and it is very educating finding it.

[Pano]

I agree with Svante 100%.

No one measurement sums it all up. You needs lots of different types. And you need to listen to real, acoustic music to get an "ear" for things.

That is what led me to my favorite speakers in the world, many years ago. Because I worked in live music and lived with a cellist, my ears were "highly calibrated." Live acoustic music for many hours every day. When I heard the speakers it was like a flash. "Yes! That's it! That what real music sounds like." It's rare to hear a speaker that does not sound like a speaker. Even very, very good speakers.

The ear is the final judge, but measurements are ohhhhh so useful to find the troubles. A real time saver.

[Conrad Hoffman]

As much as I believe in measurements, loudspeakers are a tough nut to crack that way. Collect a list of every possible speaker measurement that can be made and you'll be overwhelmed. If you can't come up with 50, you're not even trying, and there are probably hundreds. OTOH, I do believe measurements are essential if you're to avoid innumerable pitfalls that the ear is slow to pinpoint. The ear is very sensitive, but so easily misled as to the root causes of what it hears. Once you have a design that measures well, whatever limits you choose to apply there, you're ready for several weeks or longer of voicing it to sound the way you want, in the room it will live in, and for the program material it's intended for. I voice speakers differently for CDs vs. vinyl, and probably differently than somebody who listens exclusively to classical or exclusively to rock. I'm also sure I prefer a different sound than somebody younger or older- I certainly don't like what I liked 30 years ago in terms of music or speakers.

[BHTX]

Originally posted by panomaniac
That is what led me to my favorite speakers in the world, many years ago.

Mind saying what those loudspeakers were? Just curious.

[MadMutt]

My thoughts.

Measurements are the starting point.
A 'guide' to what they could be capable off.

After that the ears are what matters.

"What is a technically perfect speaker?"

I would think one that can accurately reproduce the signal that is fed into it.

Whether this actually means a 'pleseant' sound or not is another matter.

Example;
You design a speaker system using software,
the design has a ruler flat response from 20Hz-20KHz.
You build it as per design, and under a clinical Fq sweep it is indeed ruler flat.
Then you actually listen to it.
It sounds terrible, harsh treble, 'flat' mid and no bass.

Would you say the 'bbc dip' is incorrect because it isn't a 'technically' correct response ?

If so why do so many use it as a base line for their speakers ?

This is not a fantasy, it happens, every day, here for one place.

That's why when 'we' (DIY ETC) build a speaker, we expect to have to adjust the crossover 'by ear'.

If there was a technically perfect speaker that sounded 'perfect' I'm confident the chinese would have it copied by now anyway.

The biggest influence on any purchase be it a speaker/car/house/girlfreind/whatever, is the person doing the buying.

If it wasn't we wouldn't have Quad, Revox, Krell, ML, Nakamichi, JVC, Ford, Holden, Mitsubishi, blonde, brunette, etc.

What sounds right to YOU, might sound like rubbish to the designer of the Quad ESL range.

Who is right ?
Personal choice thats who.

Time and time again, people ask me "What is a better xyz ?"
I give the answer as a list of options.
The buyer will decide what is best for them.

Sometimes they are wrong, (to me anyway).....

[planet10]

Measurements are a useful tool.

Speakers are not very good -- even the best -- when it comes to doing what they ideally should (my estimate is that we are ~10% of the way). With even the best speakers doing so much "wrong" there is a lot of room for the art of loudspeaker design.

Further a loudspeaker should never be considered separate from the amp that will drive it. That adds even more degrees of freedom.

In the end the best speaker is the one that communicates the emotion of the music to YOU. And that may not be the one that does it for me.

That is why there is such a range of expression when it comes to loudspeakers, And with speakers as bad as they are on an absolute scale, why they are on some senses the least important part of your system & in other senses the most important.

dave

[Svante]

There is another crux in this as well...

The majority of the CDs available to us have passed a recording studio and a mastering studio. Here, the engineer applies various effects and eq so that the sound is "right" in his ears, using the loudspeaker and acoustic environment of the studio listening room.

Now if these loudspeakers have a "BBC dip" or whatever quirk in the response, the engineer will adjust the eq so that the sound is right under those conditions.

Seen in this light, loudspeakers at home should not reproduce the simple two-microphone-recording-without-eq as good as possible, but mimic the studio setting as far as possible.

It is a pity that studio monitoring is not standardised, this results in recordings that are essentially incompatible with one another. IMO the best would be if studios could standardise their monitoring to behave as close to good reproduction of "straight" recordings as possible. That would solve the entire problem.

[francis varkey]

My thoughts:

A speaker that's well designed should faithfully reproduce the signal fed into it. This will be where measurements help is getting close to that goal.
The fact that a recording which is miked close to the source(violin or whatever) will never provide you the sound which you would have heard sitting in say the 3rd seat from the centre in the 4th row. For reproducing the sound heard whilst sitting in this seat, it would require the mike(s) to be placed in that particular seat. Sadly, this will also reproduce the chronic cougher who spoilt that evening for you with his considerable sonic offerings.
All in all, I'd like to say that measurements will help you achieve a satisfactory response from the speaker. Then the ears take over.
All said and done, it seems that we are yet to find the perfect speaker.
We need to agree what perfection is; nice discussion coming around here.
Regards to all(haven't posted for quite a long time)