After reading through the Loudspeaker Measurements Explained thread, I thought that maybe someone could answer a question that has been floating in my mind for sometime. If you have two speakers that measure the same, can they sound different?
Measurements Vs. Sound
- Thread starter Matt193
- Start date
garylkoh
WBF Technical Expert (Speakers & Audio Equipment)
It depends on what you mean by "measure the same". If you use a fully exhaustive suite of measurements, then they should sound the same. The problem is that I haven't figured out what "fully exhaustive" is.
There is a lot of science in loudspeaker measurements, and there are a lot of measurements you can do. I think that I can get only 80% of the way to the final result. The last 20% takes probably 80% of the design time - because there is then a lot of iteration that takes you away from some ideal, and then you have to work towards it again.
There is a lot of science in loudspeaker measurements, and there are a lot of measurements you can do. I think that I can get only 80% of the way to the final result. The last 20% takes probably 80% of the design time - because there is then a lot of iteration that takes you away from some ideal, and then you have to work towards it again.
"Measure the same", yes, if sufficient measurements are taken, but getting a complete set of measurements is very difficult. Frequency response and distortion vs. power level, dispersion, phase and time delay, etc. are difficult to measure and require lots of equipment (and a good anechoic chamber). In practice, making all the measurements required to say that two speakers measure identically is impractical for all but the loudspeaker companies, some reviewers/magazines, and some independent labs.
Here's a hypothetical situation where speakers might measure the same but sound different.
If the measurements were taken from only one point in space and the speakers measured the same, if they have different dispersion patterns, they would sound (and measure) differently from other points in space.
If the measurements were taken from only one point in space and the speakers measured the same, if they have different dispersion patterns, they would sound (and measure) differently from other points in space.
Just consider the crossover components. Most manufacturers will tune the speaker choosing component types (I am not referring to the values) that do not show in measurements.
Speaker measurements are much less accurate than electronics measurements - manufacturing tolerances and dependence on external parameters is much higher. IMHO, they are much more useful to the manufacturers during development and production than to the consumers.
Speaker measurements are much less accurate than electronics measurements - manufacturing tolerances and dependence on external parameters is much higher. IMHO, they are much more useful to the manufacturers during development and production than to the consumers.
Mark Seaton
WBF Technical Expert (Speaker & Acoustics)
Just as others have hinted at, the problem in answering this question comes down to the qualifier "measure the same." Unless they are nearly identical, this won't be the case in any absolute sense...
The problem is that many do not get far past on axis frequency response of a loudspeaker. We can take hundreds of additional, identifying measurements to quantify different aspects of a loudspeaker. Ironically the most telling metrics are the most complex to measure and least easy to depict, particularly for a single graphic comparison. If you are familiar with the grayscale of a video display we could liken these differences as all being deviations from a perfect grayscale. Some deviations are more towards red, some more toward green, with many possible points in between. Now consider that a real loudspeaker operating in a real room adds many more dimensions than can be represented in a simple gray pattern, and you might get an idea as to just how many variations there can be.
That said, yes, if you get enough metrics to be very similar (and that is a lot of metrics) two speakers will sound very similar... under certain conditions.
Hello Matt
Depends on how comprehensive the measurement set is and if the measurements are done under the same conditions. For example if you look at gated on axis frequency response the amount of smoothing used and the gate time used could give you similar curves between two different sounding systems. Lets say one with 1/3 octave smoothing and a short gate vs 1/12 octave smoothing and a longer gate say 2X longer. The 1/3 smoothing and short gate time could give you a very nice looking smooth curve. Lets say you get an identical curve on the other speaker using the longer gate and 1/12 smoothing. The chance are good that if you remeasure the other speaker using the same paramaters the curves will no longer look the same.
The problem you have is there is no standardization on how the measurements are taken or a standard set of measurements that are accepted industry wide. Take a look at Tooles book, it's a great book that will help you understand how measurements can be useful and what they tell you.
Rob
Other people will have more experience than me, but years ago I helped a friend replacing the capacitors of an old JBL speaker with much better quality ones of the same capacitance. I remember we measured them one by one to keep the values within 2 or 3%. As we did one speaker first, the second after we were able to compare them and they sounded really different.
Recently I compared my Soundlab A1 Px backplates to the so called hot rod level - resistors, capacitors and coils were exchanged by more powerful thick film ones, paper in oil and special coils respectively. The global difference between the sound pre and after is more than what I get changing the medium adjustments by 3dB.
BTW, I did not say the differences are immeasurable. I say that using current, commonly used measuring techniques within normal standard tolerances, these differences do not show significantly in the measurements.
Are you sure? If the differences were that audible - and I don't doubt for a second that they were - a very common on-axis FR measurement should reveal them.
Tim
Hello Tim
It depends. Ever hear of running capacitors Charge Coupled with a DC bias voltage?? You will never be able to see a difference in the on axis frequency measurements with two networks one CC and one not. That said you will hear a difference. Try it yourself and see what you think. The differences are not night and day they are subtle but audible.
Another example is Aquaplas, a dampening compound, treated compression driver diaphrams. You won't see a difference in the on axis in two systems set-up with and without. What you need to do is look at the difference in decay charateristics in the time domain using ETC or Waterfall. Again the differences are not night and day. When you put both the CC and the damped diaphrams together it's makes both more obvious.
Rob
Mark Seaton
WBF Technical Expert (Speaker & Acoustics)
That will depend on how you are measuring and what you are using to measure. Some measurement systems have much more variation sweep to sweep than others, some are more or less susceptible to noise. In this specific case of a crossover, you are just as likely to have the significant difference be further into the roll of one driver element in a multi-way speaker, where even significant changes in response make for very small differences in the summed response. Finally, you have to account for differences as they relate to drive level. At the lower end of quality and with some options for very large capacitor values you will see measurable differences. Some are superior, some are just different than another.
In a fairly comprehensive set of measurements, like the ones we do at NRC, I can say with confidence that you'll never see two different models of speakers measure the same. Heck, it is even hard to get a matched set of stereo speakers to measure the same (pair matchning is something not all companies do so well).
Thanks for the fantastic replies. I am more curious as to how some people say that they can tell how a speaker is going to sound by looking at the on-axis FR graph. To me it would seem like a speaker comprised of cheap drivers but with good frequency response is going to be at a disadvantage over another speaker using quality drivers with the same FR graph or am I wrong thinking that?
At the same level the two systems will have the same FR. Says nothing about how linearly either plays (at any volume), what their radiation pattern is, etc. FR is barely a starting point for measurements.
Definitely true. That's why we publish a "listening window" that is an average of on-axis, 15 degrees right, 15 degrees left, 15 degrees up, 15 degrees down.
Even still, we publish up to 75 degrees off axis as well to get a more complete picture.
And example of a recent Amphiom model:
http://www.soundstagenetwork.com/in...hion-argon3-l-loudspeakers&catid=77&Itemid=18
I did not say "insignificant", I said it's a starting point, barely. Especially one made at a single point (he did say "on-axis") and power level. A more accurate statement would have been that I weigh it roughly equally with a number of other parameters, but like anything else that assumes a certain base level of performance. To be truthful, I tend to weigh FR a little more heavily, but so many other things are so important that I would not assign it dominance. I probably should have not thrown in "barely"... I think too many people put too much significance on FR to the exclusion of everything else, and in the real world the room tends to dominate the actual FR anyway. However, I do tend to forget that not everyone has a pair of Maggies or B&W's (or whatever) and good subs to start with... Let alone what some of the real WBF members have! (Yeah Steve, I'm talking about you, buddy! And the rest, you know who you are... )
I have Toole's book, and a number of others -- the problem is finding time to read them all!
)I have Toole's book, and a number of others -- the problem is finding time to read them all!
garylkoh
WBF Technical Expert (Speakers & Audio Equipment)
There are lots of decisions that a loudspeaker designer can make that have the same measured result with respect to FR and phase. Say you have a mid/woofer that will go up to 3k and a tweeter that will go down to 1k. That's a 2kHz range where you can crossover the two drivers, still stay flat, and still be phase coherence. If the mid/woofer is not too large, the dispersion characteristics of the speaker will still be pretty good within the range.
Will they sound the same crossed over at 1kHz vs 3kHz. You betcha! Which will sound better? It depends.
Will they sound the same crossed over at 1kHz vs 3kHz. You betcha! Which will sound better? It depends.
Two speakers that measure 'quasi' the same are from the same spot.
When in real use they are positioned at different spots (for a pair). ...The same with women.
When in real use they are positioned at different spots (for a pair). ...The same with women.
Similar threads
- Sticky
