Theoretical question - differences between speaker systems

[RayW]

If you pumped in a sine wave, and the speaker gave you back a square wave, the frequency response could be identical and they'd still sound very different.

Do you mean one speaker spits out sine waves and the other square waves of the same frequency? In that case you would have the harmonic structure to deal with. A look at FFT's would show that the sine wave is a single frequency but the square wave has a whole harmonic parade following it. The frequency response wouldn't be the same unless you redefined what you mean by frequency response.

 

[barredowl]

Gordon thanks for adding transient response which I think is often forgotten, I think that accounts for most of the difference in sound say between small bookshelves and big efficient horn or monitor type speakers IMO.

 

[Negotiableterms]

Do you mean one speaker spits out sine waves and the other square waves of the same frequency? In that case you would have the harmonic structure to deal with. A look at FFT's would show that the sine wave is a single frequency but the square wave has a whole harmonic parade following it. The frequency response wouldn't be the same unless you redefined what you mean by frequency response.

I was thinking only of a distortion in which the waveforms of the two speakers are not the same, but the volume level at each frequency remains the same. With the FFT's, etc., you went above my pay grade.

 

[TerryO]

O.K., if we then take it that "frequency response" does not mean a single on-axis measurement, rather, a definitive sample of perhaps hundreds of measurements with infinite resolution taken in spherical planes at representative radii, all being identical, I submit that it would still be slam-dunk simple to tell whether a silk vs. titanium dome tweeter was playing the VHF.

Somehow, however, it seems doubtful OP was postulating such a scenario.... :tongue:

Zilch,
I don't know the person posting the question, nor to what degree of detail he had in mind. He didn't specify only on-axis measurements, which I've noticed is only the tip of the iceberg as to total performance and gives a somewhat tenuous grasp of what a speaker is doing.

As far as your example as to a silk vs. titanium tweeter comparison, are you saying that careful measurements would be unable to detect a difference between them even at VHF?
Dan Wiggins has often stated that if you can hear a difference, he can measure it. This may be an overstatement on his part, but so far his batting average has been pretty high.

Best Regards,
Terry

 

[Nikko75]

If you can hear it, it can be measured. We can measure far beyond the threshold of human hearing but the equipment must be equally high resolution and precision, we won't get that with Clio or other budget hardware/software.

 

[Zilch]

While I am clearly a proponent of objectivism in these forums, and frequently cite and argue extreme positions on such matters, in the more practical context of OP's question:

When one speaker is "better" than another, or more detailed, or has "more air", or "more presence", or even if an instrument section appears to be in a slightly different space than when heard with another speaker...are these observations solely a function of the speaker's frequency response?

The answer is "No, there is more, some of which we have only just begun to explore and understand."

[Where is Ken Kantor when I need him...? :tongue: ]

 

[Nikko75]

When we see loudspeaker measurements on the internet we're looking at low resolution compositions and its easy to make assumptions on what is measurable based upon the aforementioned. In reality we can even measure conductors and know why they sound the way they do, and have been capable of such for over thirty years. The problem started when magazines bought the budget audio measuring equipment in the 1980's with limited accuracy and resolution. While it was good enough to get an idea of what they were working with, much of what the speaker or conductor did was not readily visible on that equipment. The printed conclusion was then drawn that they could not measure what they could hear. In regard to the equipment being used, they were correct but the ill-informed statement aimed at technology as a whole was obstinance.
A major contributor to a speaker's sound other than response is it's settling characteristics and each implementation of radiating plane technology and material used displays it's own characteristic ringing or delayed settling that changes with duration, amplitude and primary frequency being used. CSD is often represented as a compressed version of a speakers sinusoidal settling response measurement over a bandwidth. When a portion of a speakers CSD is broken down into a narrow frequency bandwidth and expanded considerably both in regard to frequency and time domain, software can then aid in highlighting the waveform characteristics that create the sonic "signature". With a live tone or series of tones being played this is quite a bit easier as I can then see a live waveform for the whole 360° duration. Metals, woods, plastics all have their own characteristics adhering to trends and that is why they sound different. Sinusoidal response can encompass every characteristic of a speaker with the exception of polar dispersion unless multiple measuring devices are in use.

 

[TerryO]

Several FAMILIES of factors that would ALSO have to be identical, for the speakers to even have a chance of sounding alike:

-Frequency response- on and off axis response curves, extension on top and bottom end.

-Transient response- group delay in the bass, resonant modes (and their die-down times) of the individual drivers, damping ratios in the crossovers, enclosure damping/resonance, driver time alignment (difference in path lengths to the listener between all drivers in the system) and so on.

-Dispersion characteristics- monopole vs. dipole, baffle width (baffle step effects), driver diameters vs. frequency range (lobing in individual drivers), driver center-to-center spacing vs. frequency of crossover (lobing from adjacent drivers), driver location on baffle (diffraction effects)

Without EVERY ONE of those parameters matching exactly, there's very little chance of the speakers sounding "identical", even if their on-axis response matched. And even if all these matched, as Zilch mentioned, there are still whole fields of research devoted to other aspects of performance and design, many of which we know very little about to this point.

Regards,
Gordon.

Gordon,
Thank you for listing the factors that impinge upon the reproduction of sound from a speaker. I'll try to explain my understanding of your listed concerns.

1. Frequency Response:

I agree, and in fact, have specifically stated the same previously.

2. Transient Response:

Group delay is controversial, but I believe that there is a general consensus that it is not usually audible.

Resonant modes (and their die-down times) of the individual drivers: I believe that these would show up on a frequency response if they're audible.

Damping ratios in the crossovers: At what point, that is, what is the ratio that this become audible?

Enclosure damping/resonance: I believe that this would also be a factor in the measured frequency response. Otherwise why would Harbeth, and others like Sonus Faber, expend effort to create certain conditions, that they feel affects the sound in a positive manner if they're unable to measure the results. I would think that any resonance of a cabinet would show or affect the frequency response.

Driver time alignment (difference in path lengths to the listener between all drivers in the system): Some designers feel that (to a certain degree) this can be corrected with the crossover. I'm not convinced that off axis measurements wouldn't be affected in this case.

3. Dispersion characteristics- monopole vs. dipole, baffle width (baffle step effects), driver diameters vs. frequency range (lobing in individual drivers), driver center-to-center spacing vs. frequency of crossover (lobing from adjacent drivers), driver location on baffle (diffraction effects):

I concur that every example you list has properties that affect frequency response. What's the point? I never said, nor do I believe that, these are factors that do not change or influence the sound, and I fully believe that careful measurements would back me up on this point. Again, when he said identical frequency response, I assumed he meant identical.

Perhaps it comes down to what we mean by identical. What is the essential nature of identical? Are there degrees of "identical?"

Finally, while I have participated in the design of several commercial speakers, I'm well aware that I don't know it all. As my undergraduate and graduate degrees are in History, I'd be the first to agree that this is hardly a qualification for speaker design. However, I do believe I am able to read the English language and have a fairly adequate vocabulary.

As it seems evident that I don't have a remote clue in these matters, I'll refrain from commenting in further discussions.

Best Regards,
TerryO

 

[valve_amp]

if two speakers are playing in the woods..are they putting out sound? jk...if a bear takes a chit in the woods and leave does his chit still stink..jk...

ok my answer about the speakers...if two identical speakers are playing in the same listening area at the some time..they
they are acting as one..so they do sound alike...otherwise hell no.jk

 

[trwham]

If you can hear it, it can be measured. We can measure far beyond the threshold of human hearing but the equipment must be equally high resolution and precision, we won't get that with Clio or other budget hardware/software.

How about: if you can hear it, it should be measurable if we understand how to measure it? I can hear the difference between Sonny Rollins and John Coltrane, but I am not sure we can measure and identify that difference. In the case of these hypothetical speakers, if they are not identical, then I believe they cannot sound the same, but I cannot be sure the difference would be measurable with known technology.

 

[Blast]

While I am clearly a proponent of objectivism in these forums, and frequently cite and argue extreme positions on such matters, in the more practical context of OP's question:

The answer is "No, there is more, some of which we have only just begun to explore and understand."

[Where is Ken Kantor when I need him...? :tongue: ]

I would agree.

 

[Nikko75]

How about: if you can hear it, it should be measurable if we understand how to measure it? I can hear the difference between Sonny Rollins and John Coltrane, but I am not sure we can measure and identify that difference. In the case of these hypothetical speakers, if they are not identical, then I believe they cannot sound the same, but I cannot be sure the difference would be measurable with known technology.

The difference will be found within the sinusoidal waveform from which the recording is formed.

The question in that situation would be: What has led someone to feel it is not measurable? Who has wrote or said that and with what equipment were they using that was such a low resolution that it could not relay a difference?

 

[barredowl]

Hmm I interpret good transient response as not compressing dynamics during a sudden short louder spike of music. Small bookshelf speakers simply aren't capable of the volume reached in a real sharp transient IMO.

 

[trwham]

The difference will be found within the sinusoidal waveform from which the recording is formed...

Except perhaps for a lone flute or organ playing a single note, music is not a sinusoidal waveform, but rather the sum of several waveforms. A symphony orchestra may be 90 pieces playing at once. This will not be a single, simple waveform. This is why I am skeptical that it can be measured so certainly.

 

[Nikko75]

I chose the term sinuous because it's easier for people to picture electrical current and sound waves in that regard and themselves imagine the alterations that make up different sounds. When music is played back through a sound system, the signal is an ever changing electrical waveform: One for each of the two channels forming stereo that carry the information regardless of the number of instruments or sonic complexity.

Sound in nature is the same way. From an easily measurable and predictable 3D perspective an orchestra contains many separate point sources that radiate sound if different ways and the sound and intensity will change depending on where we are listening, but no matter where the listener is, the sound entering the individual ear is received as a single wavefront; a single waveform. Through timing, pan and phase for example, these two channels work together to offer an infinite number of perceived sources and vertical& horizontal instrument location and distance between and outside these two channels.

 

[jwalker]

In other words, if two speakers have an identical frequency response, and are placed in the same spot in the same room, will they sound identical?

Frequency response only describes one aspect of a speaker system. There is also phase response and the effect from non-linearities (distortion).

I would say "no".

 

[jwalker]

Do you mean one speaker spits out sine waves and the other square waves of the same frequency? In that case you would have the harmonic structure to deal with. A look at FFT's would show that the sine wave is a single frequency but the square wave has a whole harmonic parade following it. The frequency response wouldn't be the same unless you redefined what you mean by frequency response.

A true square wave is chock full of odd order harmonics (Fourier series) and would appear in the frequency response.

 

[jwalker]

Gordon,
Thank you for listing the factors that impinge upon the reproduction of sound from a speaker. I'll try to explain my understanding of your listed concerns.

1. Frequency Response:

I agree, and in fact, have specifically stated the same previously.

2. Transient Response:

Group delay is controversial, but I believe that there is a general consensus that it is not usually audible.

Resonant modes (and their die-down times) of the individual drivers: I believe that these would show up on a frequency response if they're audible.

Damping ratios in the crossovers: At what point, that is, what is the ratio that this become audible?

Enclosure damping/resonance: I believe that this would also be a factor in the measured frequency response. Otherwise why would Harbeth, and others like Sonus Faber, expend effort to create certain conditions, that they feel affects the sound in a positive manner if they're unable to measure the results. I would think that any resonance of a cabinet would show or affect the frequency response.

Driver time alignment (difference in path lengths to the listener between all drivers in the system): Some designers feel that (to a certain degree) this can be corrected with the crossover. I'm not convinced that off axis measurements wouldn't be affected in this case.

3. Dispersion characteristics- monopole vs. dipole, baffle width (baffle step effects), driver diameters vs. frequency range (lobing in individual drivers), driver center-to-center spacing vs. frequency of crossover (lobing from adjacent drivers), driver location on baffle (diffraction effects):

I concur that every example you list has properties that affect frequency response. What's the point? I never said, nor do I believe that, these are factors that do not change or influence the sound, and I fully believe that careful measurements would back me up on this point. Again, when he said identical frequency response, I assumed he meant identical.

Perhaps it comes down to what we mean by identical. What is the essential nature of identical? Are there degrees of "identical?"

Finally, while I have participated in the design of several commercial speakers, I'm well aware that I don't know it all. As my undergraduate and graduate degrees are in History, I'd be the first to agree that this is hardly a qualification for speaker design. However, I do believe I am able to read the English language and have a fairly adequate vocabulary.

As it seems evident that I don't have a remote clue in these matters, I'll refrain from commenting in further discussions.

Best Regards,
TerryO

True transient response happens strictly in the time domain and can be very easily determined by the frequency response and vice-versa (Fourier transform).

But the complication happens because frequency and phase response are power dependent (i.e, distortion/compression, etc). So stuff like "dynamics" is simply describing an "end effect" from the fundamental characteristics of frequency, phase, and distortion performance.

 

[barredowl]

jwalker while I don't have the expertise you do isn't there an absolute db limitation going on in smaller less efficent speakers as well? For example when I had my Klipsch Hersies they would occasionally startle me with just how loud an occasional transient peak was, which I never hear in such a dramatic way on my NHT Super Ones. If you hypothetically equalized the frequency response of say a big horn loaded speaker and a small bookshelf they might sound very similar on softer material on axis but I think a difference in volume will be heard on transient peaks that will suddenly make them sound very different.

 

[cwall99]

Perhaps it comes down to what we mean by identical. What is the essential nature of identical? Are there degrees of "identical"

Or "is"?

Sorry... just a throwback to the good ol' late 90s.

When it comes to speakers sounding identical, my definition for that would be that, if I were blindfolded, and someone were doing an AB test, I would not be able to hear any difference at all between the two speakers' playback. I would not be able to distinguish between Speaker A and Speaker B.