Absolutely! It makes perfect sense.
So, it's a bit of a confusing balance. We like smaller drivers because they're fast. And so, if we take the double bass as an example, it's a standard bit of orchestration, that, the low E string (the diameter of a pencil, 6 feet long, supported by an approximately 5-6 cubic foot body) isn't going to "speak" quickly in its lowest register. So, if we use a larger driver, do we augment the physics of that instrument (slow rise time in the lowest register) by using a larger driver? But, if we use a small driver in a small cabinet, how can we expect it to accurately represent the tone of the bass?
So, it's a bit of a confusing balance. We like smaller drivers because they're fast. And so, if we take the double bass as an example, it's a standard bit of orchestration, that, the low E string (the diameter of a pencil, 6 feet long, supported by an approximately 5-6 cubic foot body) isn't going to "speak" quickly in its lowest register. So, if we use a larger driver, do we augment the physics of that instrument (slow rise time in the lowest register) by using a larger driver? But, if we use a small driver in a small cabinet, how can we expect it to accurately represent the tone of the bass?
Last edited:
In simple theory, if the small speaker can reproduce the tones that the bass can and in a linear manner, it will sound like the bass.
Of course it rarely works out that way in reality. But the dream lives on.
I think that some of the trouble you are hearing in that midrange resister could be due to harmonics. A complex subject, but I believe it makes more of a difference than we tend to realize.
I believe there is something going on in the typical case. Sound enters the room in a unique way with a freestanding box speaker in this range. Remember the goal here is reproduction, and even-handed treatment of each frequency under anechoic conditions should be able to achieve this. The size of the drivers may have a bearing on their competence, and the mechanics of stereo doesn't call for speakers the size of the source being reproduced.
Bring it into a room and this will dominate at the low end. A well designed speaker will have control at the top end and the question remains of how the transition is made.
A free-standing speaker will be necesarily smaller than the room so there is going to be a transition. The sound will physically overwhelm the size of the unit and disperse widely into the room down until the region where the room is taking control, which at mid frequencies exposes the discrepancy between the position of the speaker and the location within the room. Rather than full room control, the local walls can play a greater role, with a certain character of response, if not timing or apparent angle of the source.
In simple theory, if the small speaker can reproduce the tones that the bass can and in a linear manner, it will sound like the bass.
Of course it rarely works out that way in reality. But the dream lives on.
I think that some of the trouble you are hearing in that midrange resister could be due to harmonics. A complex subject, but I believe it makes more of a difference than we tend to realize.
Are we talking about phase behaviour again?
Get the harmonics there in time...
Er no. Audio magazines like to claim smaller drivers are fast. Reality is a lot more complex than that. Look at the Gedlee speakers, or any of the econowave designs and they run 12-15" pro drivers up to 1000Hz or higher. Shock horror noooooo, pro units? Yet they work, in some cases exceedingly well. The JBL M2, the current peak of the Toole skoole (sorry) of controlled directivity breaks every 'high end' rule, yet many report its rather good.
What you need to do is focus on what your speakers need to do well then chose how to approach that. But keep a very open mind and ignore the fashion du jour
This is a cogent description (simply broken into separate thoughts, just above) of one of the most difficult issues to control gracefully with real loudspeakers in small rooms, and the engineering involved to push that transition as low as possible to mask the low frequency directionality detection abilities of the human hearing system (but probably not an elephant's hearing system). For horn-loaded loudspeakers, it is possible, but the mouth size of the loudspeakers will be of the size of a SH-96, which will begin to lose its polar control below about 200 Hz.
If the loudspeakers are located in the corners of a room, that transition will sound seamless, I've found.
Note that the above discussion doesn't require the use of purple prose to describe, and it doesn't preclude the verification and evaluation of the implemented design using human hearing trials to identify issues and to follow up on their closure with design or implementation changes. This is also the technique that all the "objectivists" that I'm aware of use. This includes most major engineering departments of notable loudspeaker manufacturers. The "purple prose" folks hit the scene later when the loudspeakers are "beta tested" in controlled listening environments carefully set up by the manufacturers. It takes some effort by these engineering teams to decipher the language of these largely unpaid participants.
This is the development technique used by Toole, Olive, Paul Klipsch, and his successor (Roy D.). I simply subscribe to this approach because that method's track record over time beats trial and error using subjective listening only by an order of magnitude or two in terms of productivity and product performance results.
Others here seem to want to do it the old way, however. I don't recommend it.
Chris
Last edited:
One question remains, for me, at least. We've entered the room and placement in to the discussion and rightly so. While this certainly is an important factor, I think I probably drove my discussion in to a different register. By that I mean, specifically, in to the bottom end. This is a whole other musical ball game.
I started out by talking about midrange and mentioned a specific register on two instruments: the piano (the hardest to reproduce) and female voice (a very close second). I am trying to pinpoint in my mind's ear where that issue lays but of course, the "mind's ear" has a memory about as long as a puppy.
To be quite honest, I am leaning in the direction of the dominant 7th "chord" (it's not a chord because this is the overtone series) but the 4th through 8th partials do include a major third and the 7th of the "chord".
I want to put this out there: The fundamental pitch (the low C) through to the 4th partial contain only octaves, perfect 4ths and 5ths. In musical composition, movement(s) of octaves, fourth and fifths are *always* avoided because they stick out like a sore thumb. I am going to separate this sentence from the paragraph to highlight it:
When writing for 4 parts (SATB, or, soprano, alto, tenor and bass) movements of an octave or fifth (in parallel) are *always* avoided because one of the voices drops out. It "appears to disappear".
Listen to the first audio example on this page, then listen carefully to the second. The first example is an octave. Can you hear how "thin" that interval sounds? The second example is a major seventh. Can we hear how much more warm, interesting and complex the seventh is (ex 2). Now move on to example three. Listen carefully to how complex each interval is as we descend through each of them (sorry about the awful synthesizer vibrato).
Perfect Intervals
Now, let me get to my point: the reason I have gone to these lengths to describe musical pitches and their overtones is to ask a question about the limitations of our science. If we separate the midrange and bass driver via a crossover in (I would say) the most important register, are we making things more difficult for ourselves? Look at the overtone series again! We have put the crossover point (from the bass driver to the midrange)-beginning somewhere around- the 4th through 16th partials where all of the "warm stuff" is! *BUT* it's also where all of the nasty complex stuff is too. Middle C is 262 hz (the 4th partial of the overtone series), right at the spot where we would begin to roll off a bass driver and roll on a midrange. This becomes very complicated, right at the spot where we really need to simplify things.
In a "perfect world", musically, it would be wonderful if we could get a bass driver to shut up at *precisely* 165 cycles (E3) and a midbass or midrange to begin to speak at *precisely* 175hz (F3), then the midrange would silence at C6 or 1047hz.
That's precisely the area where the piano never seems to sound "real" in audio.
This may not make "loudspeaker sense"... but it makes "musical sense".
I started out by talking about midrange and mentioned a specific register on two instruments: the piano (the hardest to reproduce) and female voice (a very close second). I am trying to pinpoint in my mind's ear where that issue lays but of course, the "mind's ear" has a memory about as long as a puppy.
To be quite honest, I am leaning in the direction of the dominant 7th "chord" (it's not a chord because this is the overtone series) but the 4th through 8th partials do include a major third and the 7th of the "chord".
I want to put this out there: The fundamental pitch (the low C) through to the 4th partial contain only octaves, perfect 4ths and 5ths. In musical composition, movement(s) of octaves, fourth and fifths are *always* avoided because they stick out like a sore thumb. I am going to separate this sentence from the paragraph to highlight it:
When writing for 4 parts (SATB, or, soprano, alto, tenor and bass) movements of an octave or fifth (in parallel) are *always* avoided because one of the voices drops out. It "appears to disappear".
Listen to the first audio example on this page, then listen carefully to the second. The first example is an octave. Can you hear how "thin" that interval sounds? The second example is a major seventh. Can we hear how much more warm, interesting and complex the seventh is (ex 2). Now move on to example three. Listen carefully to how complex each interval is as we descend through each of them (sorry about the awful synthesizer vibrato).
Perfect Intervals
Now, let me get to my point: the reason I have gone to these lengths to describe musical pitches and their overtones is to ask a question about the limitations of our science. If we separate the midrange and bass driver via a crossover in (I would say) the most important register, are we making things more difficult for ourselves? Look at the overtone series again! We have put the crossover point (from the bass driver to the midrange)-beginning somewhere around- the 4th through 16th partials where all of the "warm stuff" is! *BUT* it's also where all of the nasty complex stuff is too. Middle C is 262 hz (the 4th partial of the overtone series), right at the spot where we would begin to roll off a bass driver and roll on a midrange. This becomes very complicated, right at the spot where we really need to simplify things.
In a "perfect world", musically, it would be wonderful if we could get a bass driver to shut up at *precisely* 165 cycles (E3) and a midbass or midrange to begin to speak at *precisely* 175hz (F3), then the midrange would silence at C6 or 1047hz.
That's precisely the area where the piano never seems to sound "real" in audio.
This may not make "loudspeaker sense"... but it makes "musical sense".
Attachments
Last edited:
By the way, I didn't want to confuse things but the image of the overtone series is numbered incorrectly. The fundamental (the number 1) is not a partial. in other words, the number 1 is actually on the second pitch from the left That is the first partial or "overtone".
I am going to repost the overtone series in pitches and the series as waves upside down (so we can connect them visually).
All of the "good stuff" is in the shorter wave lengths... fourth from the bottom and up. (C E G B flat)
This is what the chord sounds like -numbers 4 to 7 on the overtone series example I have here... and yes, that is right in the octave we're talking about. https://www.youtube.com/watch?v=FxaEwoEhs_Y
That is the first partial or "overtone".I am going to repost the overtone series in pitches and the series as waves upside down (so we can connect them visually).
All of the "good stuff" is in the shorter wave lengths... fourth from the bottom and up. (C E G B flat)
This is what the chord sounds like -numbers 4 to 7 on the overtone series example I have here... and yes, that is right in the octave we're talking about. https://www.youtube.com/watch?v=FxaEwoEhs_Y
Attachments
Last edited:
well 180Hz to 1KHz is covered by almost every 2-way out there albeit with the tweeter not coming in until around 1700Hz as a generality.
With active filters you can put steep slopes in and isolate the drivers from each other, but what you are aiming for is a smooth transition so you can't tell. This is affected by a number of things and people will argue over which of them is most important!
I do wonder if you might not be served well by a 3-way with a wide baffle so you can use Baffle step roll off to your advantage, aka FAST on this forum.
My current project will end up with x-over points around 180Hz and 750Hz, but those are driven by the design rather than anything else.
With active filters you can put steep slopes in and isolate the drivers from each other, but what you are aiming for is a smooth transition so you can't tell. This is affected by a number of things and people will argue over which of them is most important!
I do wonder if you might not be served well by a 3-way with a wide baffle so you can use Baffle step roll off to your advantage, aka FAST on this forum.
My current project will end up with x-over points around 180Hz and 750Hz, but those are driven by the design rather than anything else.
The reason I say that your piano and alto voice problem could be related to harmonics is that harmonics matter in music. I rarely, if ever, hear speaker designers talk about harmonics. Frequency response, phase, power response, crossover slopes, lobing and things like that are widely discussed, but not harmonics. Maybe because they don't know, or because they don't want to give away secrets.
The best way I can explain this is with an anecdote. Enginners - roll your eyes now, if you want.
While I was working on a crossover for my Altec A5, there was a little problem. The crossover had gotten to a good point, but I was hearing the paper cone of the Altec 416 woofer. A distinct sound of paper. This can be very pronounced on some guitar speakers, you may have heard it there.
When I mentioned this to my friend John, I told him that maybe the crossover point should be moved down, or made steeper, to get rid of the paper sound from the woofer. "No" he said "move it UP." What? Up? More woofer to get rid of the woofer sound? John explained that what I was noticing was an interrupted harmonic series. Because the woofer harmonics where truncated, my ear noticed the missing harmonics and that called attention to the paper sound. He told me to move the crossover point up by about 75-100Hz. I was dubious, of course, but did it. Presto! Paper sound gone. Move the crossover point back down, paper sound returns. It worked like a charm. A little more woofer bandwidth resulted in less audible woofer coloration.
Ever since then I've paid close attention to harmonics and how well the crossover blends. It's not something I measure, it's something I hear. The harmonics can be measured, I know how to do measure them - I just don't know what to do with the information. I believe the harmonic structure plays a big part in how well 2 drivers "play together". You will hear about matching sonic signatures of drivers, but I don't ever hear it expressed as harmonic structure.
That is why I wonder if whatever is happening in the tessitura that you find a problem, is a problem of harmonics. Frequency and phase response could look right, but if the harmonic structure is off it can sound odd.
The best way I can explain this is with an anecdote. Enginners - roll your eyes now, if you want.
While I was working on a crossover for my Altec A5, there was a little problem. The crossover had gotten to a good point, but I was hearing the paper cone of the Altec 416 woofer. A distinct sound of paper. This can be very pronounced on some guitar speakers, you may have heard it there.
When I mentioned this to my friend John, I told him that maybe the crossover point should be moved down, or made steeper, to get rid of the paper sound from the woofer. "No" he said "move it UP." What? Up? More woofer to get rid of the woofer sound? John explained that what I was noticing was an interrupted harmonic series. Because the woofer harmonics where truncated, my ear noticed the missing harmonics and that called attention to the paper sound. He told me to move the crossover point up by about 75-100Hz. I was dubious, of course, but did it. Presto! Paper sound gone. Move the crossover point back down, paper sound returns. It worked like a charm. A little more woofer bandwidth resulted in less audible woofer coloration.
Ever since then I've paid close attention to harmonics and how well the crossover blends. It's not something I measure, it's something I hear. The harmonics can be measured, I know how to do measure them - I just don't know what to do with the information. I believe the harmonic structure plays a big part in how well 2 drivers "play together". You will hear about matching sonic signatures of drivers, but I don't ever hear it expressed as harmonic structure.
That is why I wonder if whatever is happening in the tessitura that you find a problem, is a problem of harmonics. Frequency and phase response could look right, but if the harmonic structure is off it can sound odd.
Are you looking for a notch in the response?
An ideal crossover has the drivers in the same location and there is no clue to it's existence. Therefore, nothing can be achieved with an ideal crossover that cannot be done with a single driver.
It's possible to optimise or exploit the practical realities or the flaws of drivers and crossovers.
I don't think there is anything wrong or invalid about looking at the problem from a music perspective. And I certainly don't see anything 'magic' about Pano's experiences of moving x-over frequency. Makes a lot of sense.
Speakers are in general a bucket o compromise. Approaching the compromises from a different direction is always a good way of perhaps spotting something you have missed. And bringing things almost back on track, having some training in how to listen should help you get from 'the piano sounds wrong around C4' to what the root of the problem is.
Speakers are in general a bucket o compromise. Approaching the compromises from a different direction is always a good way of perhaps spotting something you have missed. And bringing things almost back on track, having some training in how to listen should help you get from 'the piano sounds wrong around C4' to what the root of the problem is.
I honestly don't know, but I am interesting in finding out. I have a personal reason as the current project speakers may well end up a 3-way (or even 3.1 way). I can cross over to the ribbons almost anywhere above 500Hz, but whenever you have a crossover you have a risk of something being messed up (directivity, lobing etc).
Take a concert grand where the fundamentals run from around 30Hz to around 4KHz. You can see why a full range speaker with some help top and bottom would seem a good thing (tm).
It's an area I am woefully unknowledgeable in currently. But there are some clever people here who do know
Take a concert grand where the fundamentals run from around 30Hz to around 4KHz. You can see why a full range speaker with some help top and bottom would seem a good thing (tm).
It's an area I am woefully unknowledgeable in currently. But there are some clever people here who do know
@Pano, while we're on the hypothetical, what if the paper sound was the result of distortions produced after the crossover and increasing the bandwidth somehow 'covered' that range of distortion harmonics by allowing the same driver to produce the 'natural' version of those harmonics and making a quasi minimum phase deal of it?
This really depends on what the crossover is 'doing'. It should be doing nothing. In reality there will be spatial changes and differing behaviours from the dissimilar drivers.
Allen. Don't know about the phase thing, but my guess is that I had truncated the woofer cone harmonics at exactly the wrong spot, making my ear pick out the odd sound. Once the harmonics were allowed to develop more fully, they didn't call attention to themselves. I had not thought about the phase.
I was stupid and didn't measure it. It was a perfect chance to see what was going on and I missed it.
I was stupid and didn't measure it.
It was a perfect chance to see what was going on and I missed it.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.