Properties of Speakers

[WmAx]

Question:

Is the sound heard through earphones identical to what one would hear in an anechoic chamber (i.e. all direct sound + zero reflected sound)? Does this explain why music through headphones is incredibly detailed?

From the aspect of 100 percent of the sound being direct sound, this is true. But the speakers will still provide a degree of spatial effect not possible with the headphones, because the speakers are located in front, at an angle, in relation to your ears/head. The head shape and ear shape will apply a frequency response filtration on the high frequencies, and the left and right channels of speaker sound will cross-mix(you will pick up left channel audio with right ear and you will pick up right channel audio with left ear). Headphones eliminate this effect. Also, resonances will be an issue that have to be considered between headphones and speakers. On a speaker, assuming high quality drivers and crossover, the cabinet is the most difficult part to get right concerning audible resonances. The size/mass of the cabinet on a full size speaker generally makes it expensive to manufacture while preventing audible resonances on conventional designs. Exceptions exist such a open back speakers which make the issue easier/cheaper to deal with. Headphones, being very low mass and small size, are easy/cheap to construct without the risk of problematic resonances of the frame/rear enclosure(homologous to the cabinet of a speaker). The headphone and speaker can be equally free of resonances, but it's much easier/cheaper to accomplish with headphones. However, in an anechoic chamber, resonances are less audible than in a regular reflective room. Resonances are [1]easier to detect when the signal is repeated/delayed. So, along with the inherant simple/small size construction of a headphone that makes reduction of resonances easier in the first place, they are in essence equal to listening to speakers in an anechoic chamber, so any resonances that are present are even more difficult to detect as compared to reflective/live environment. This must also be weighted/considered.

-Chris

[1]The Modification of Timbre by Resonances: Perception and Measurement
Floyd Toole, Sean Olive
JAES, Vol. 36, No. 3, 1988, March, pages 122-141

 

[Buckle-meister]

...the speakers will still provide a degree of spatial effect not possible with the headphones, because the speakers are located in front, at an angle, in relation to your ears/head.

Yes, it's odd isn't it? When listening to music through headphones, the music is in the centre of your skull. 'Tis a strange concept; people singing inside my head!

 

[Buckle-meister]

I've been thinking a little more on x-db/octave slopes.

I understand that the crossover in a speaker directs frequencies from x to y to say the tweeter and from y to z to say the woofer (in this example consider the speaker a two-way design).

Is it the crossover type that dictates the rolloff of the drivers? If so, what governs the decision to use a 12dB/octave crossover as opposed to say a 24dB/octave crossover? What I'm getting at, is that I would have thought that the theoretically perfect multi-driver speaker would have a brick-wall crossover between the drivers so they didn't overlap a portion of the frequency response. But if this were so, speaker manufacturers would simply always use the x-dB/octave crossovers with the greatest slope that they could, no?

...the passive radiator/ported system is a 4th order alignment, it will average 24db/ octave attenuation rate. A sealed system is a 2nd order alignment, thus averaged attenuation rate is 12db/octave.

What does 2nd order/4th order mean? Is there a first/third/fifth etc?

 

[WmAx]

Is it the crossover type that dictates the rolloff of the drivers?

Yes, the crossover, but also the natural acoustic rolloff of the drivers contributes. The crossover is integrated with the acoustic rolloff to achieve a final target response. For example, if driver X has a natural 6dB/octave roloff after 4kHz, and you want to low pass this driver at 4kHz, the natural rolloff is going to coincide with part of the electrical crossover's bandwidth. This means that if the final target response is 12dB/octave, that the electrical crossover will be 6dB/octave, becuase the natural rolloff will cascade with the electrical rolloff to provide a final effective target attenuation of 12dB/octave. This is over-simplifying things, but it is a fair summary.

If so, what governs the decision to use a 12dB/octave crossover as opposed to say a 24dB/octave crossover?

This is an engineering matter. It's like asking why use aluminum in one application vs. magnesium in another. It's a complex matter that must be addressed within a specific scenario, weighting the target objectives.

What I'm getting at, is that I would have thought that the theoretically perfect multi-driver speaker would have a brick-wall crossover between the drivers so they didn't overlap a portion of the frequency response.

(1) The steeper the roll-off, the more complex/expensive the passive crossover becomes.

(2) An analog brick-wall filter would create audible phase distortion if used in the lower midband range of frequencies.

(3) A linear phase FIR digital brickwall filter could be used that has no phase distortion, but this is only true for the on-axis response. Off axis(a driver's amplitude/phase response is not the same as you move around it's axis), a very odd pre-ring phase distortion will occur on almost all speakers. The extent of this effect on reflected soundfield in the room and the perceptual weighted effect has not been investigated with credible preceptual research as of this point, at least not of which I am immediately aware. Therefor I can not comment on the effect this will have on sound quality. However, in a situation where the speaker is used where no reflected/off axis sound is contributing the final response, this is a viable method of crossover.

What does 2nd order/4th order mean? Is there a first/third/fifth etc?

For simplicity's sake, you can interpret each 'order' to mean '6dB/octave'. Therefor, 1st order is 6dB/octave, 2nd order is 12dB/octave, 3rd order is 18dB/octave, etc..

-Chris

 

[Buckle-meister]

Many thanks. I understand.

 

[bigpapa]

Great analogy

The Spring on a Weight analogy really hit it for me and helped me understand.

Awesome exchange guys (and skirts). Of course, this only creates more questions. I'll divert to another thread so to not derail this one.

 

[Buckle-meister]

This has become something of a pet thread of mine...

Two questions:

1. When speaker manufacturers build cabinets as massive as they can and/or stiffen the cabinet internally, I understand this is to try and minimise as much as possible the colouration a vibrating cabinet adds to the sound from the drivers alone. Am I right in saying that it's not that the vibrating cabinet is bad in its own right, i.e. if we could somehow mute the drivers yet allow the cabinet to continue vibrating we'd not actually hear sound from the cabinet, but that instead it's the interaction of cabinet vibration with driver motion that causes us not to hear what the driver on its own is attempting to reproduce?
2. Assuming a speaker is considered as the sum of three components; the drivers, cabinet and crossover, what proportion expressed as (approximate) percentages do each contribute to the final sound that leaves the speaker? In other words, are the proportions analogous to the room/speakers/hi-fi (respectively?) where the room has enormous influence and the speakers and hi-fi far less?

 

[zumbo]

This is great thread. I am not as educated in the field as the people in this thread, but I do have a few opinions.

A curved or odd shaped cabinet is the best design.

Each speaker in it's on sealed section is the best design.

The crossover design/quality/match is extremely important. Great drivers can sound like crap with a poor crossover.

I am not a fan of passive radiators or ports. But, my speakers and sub are ported.

Driver arrangement is also important.

I like this idea:

My Quarts:

While I did skim some of this thread, I didn't see the properties of a 2 & 1/2 way speaker such as mine mentioned. The two mids share most of the midrange, while the top mid gets a higher rolloff, and the lower mid makes it to the lowest freq spec of the speaker. While not the best technical explination, you can get the idea.

 

[Buckle-meister]

Zumbo, you've just caused me think of something else:

Once its 'latitude', diameter and length are calculated, does it matter whether a port is on the front or rear face (assuming these are of identical width) of a speaker?

 

[Sheep]

This has become something of a pet thread of mine...

Two questions:

1. When speaker manufacturers build cabinets as massive as they can and/or stiffen the cabinet internally, I understand this is to try and minimise as much as possible the colouration a vibrating cabinet adds to the sound from the drivers alone. Am I right in saying that it's not that the vibrating cabinet is bad in its own right, i.e. if we could somehow mute the drivers yet allow the cabinet to continue vibrating we'd not actually hear sound from the cabinet, but that instead it's the interaction of cabinet vibration with driver motion that causes us not to hear what the driver on its own is attempting to reproduce?
2. Assuming a speaker is considered as the sum of three components; the drivers, cabinet and crossover, what proportion expressed as (approximate) percentages do each contribute to the final sound that leaves the speaker? In other words, are the proportions analogous to the room/speakers/hi-fi (respectively?) where the room has enormous influence and the speakers and hi-fi far less?

1.) That cabinet would have to compensate for EVERYTHING the driver did. If it was a subwoofer, it would have to act like a passive radiator. I was at FS the other day looking at the car subs, and one dolt wired the drivers in the same box out of phase. The subwoofer made no noise (barely) because one driver was compensating for the others movement.

2.) I would say drivers, crossover, then cabinet.

SheepStar

 

[Sheep]

While I did skim some of this thread, I didn't see the properties of a 2 & 1/2 way speaker such as mine mentioned. The two mids share most of the midrange, while the top mid gets a higher rolloff, and the lower mid makes it to the lowest freq spec of the speaker. While not the best technical explination, you can get the idea.

I think WmAx explained this at one point. It has something to do with dispersion and cancellation, and having a tapered crossover solves this problem. The Energy Veritas line uses these style crossovers.

SheepStar

 

[Buckle-meister]

A couple of things Sheep;

That cabinet would have to compensate for EVERYTHING the driver did. If it was a subwoofer, it would have to act like a passive radiator.

It sounds as if you're assigning the greatest importance to the cabinet here but...

I would say drivers, crossover, then cabinet.

Then you say that you reckon the cabinet's the least important.

Also, what about those percentages (even if rough)? For all I know, they could as easily be 33%, 33% and 34% as 90%, 6% and 4%...

 

[Sheep]

A couple of things Sheep;

It sounds as if you're assigning the greatest importance to the cabinet here but...

Then you say that you reckon the cabinet's the least important.

Also, what about those percentages (even if rough)? For all I know, they could as easily be 33%, 33% and 34% as 90%, 6% and 4%...

How many wood cabinets act like a passive radiator? As far as sound goes, the drivers are the first in the list, maybe 55%. The response of the speaker is determined by the drivers.

Next, the crossover has to split the 2 drivers at the right spot. Also very important for the sound. 25%

While I put the box last I didn't mean it was useless. That said, any 3/4inch MDF enclosure with some polly fill and bracing will suffice for a speaker. Subwoofers might need more, but in general, you can get away with the earlier. 20%

Edit: For your first remark, that whole question was based on the cabinet affecting the sound. Obviously I would stress the importance of the cabinet in that role.

SheepStar

 

[WmAx]

T
[*]When speaker manufacturers build cabinets as massive as they can and/or stiffen the cabinet internally, I understand this is to try and minimise as much as possible the colouration a vibrating cabinet adds to the sound from the drivers alone.

Am I right in saying that it's not that the vibrating cabinet is bad in its own right, i.e. if we could somehow mute the drivers yet allow the cabinet to continue vibrating we'd not actually hear sound from the cabinet, but that instead it's the interaction of cabinet vibration with driver motion that causes us not to hear what the driver on its own is attempting to reproduce?

Incorrect. The cabinet baffles, of an average speaker, will have substantial acoustic output. The majority of output is limited to the specific resonances of the panels. The contribution of this will alter the percieved timbre of the speaker, the degree of which, is determined by the amplitude, frequency and bandwidth(Q) spread of the effect(s). There are numerous ways to prevent the cabinet from producing appreciable acoustic output: high level of mass loaded dampening on walls, exotice low resonance construction materials, massive structure that is very heavily braced to spread resonant frequencies to a spectrum that is not very important, efficient lower mass structural design to the same, mechanical decoupling system(s), or a combination of these.

[*]Assuming a speaker is considered as the sum of three components; the drivers, cabinet and crossover, what proportion expressed as (approximate) percentages do each contribute to the final sound that leaves the speaker?

It is an equal distribution: the failure of one will compromise the entire system.

-Chris

 

[WmAx]

Once its 'latitude', diameter and length are calculated, does it matter whether a port is on the front or rear face (assuming these are of identical width) of a speaker?

On a midbass application, the port may output mid frequency resonances/noises. There for, it is preferable to have the port fire from the rear on such application(s).

-Chris

 

[WmAx]

While I put the box last I didn't mean it was useless. That said, any 3/4inch MDF enclosure with some polly fill and bracing will suffice for a speaker. Subwoofers might need more, but in general, you can get away with the earlier. 20%

The cabinet resonances fall in the spectrum of the midrange. There for, it is rarely a consideration of resonances, that one has to account for a subwoofer. Not just 'any' 3/4" MDF enclosure will be suitable. The bracing must be intelligently designed to shift/distribute the resonances to a minimal contribution. Another option is to use large amount(s) of mass loaded dampening materials on the walls such as Dynamat. However, it should be noted that a single layer is not sufficient. You will have to use such material at 1/4-1/3 the thickness of the MDF walls in order to achieve optimal effect. This can contribute significantly to cost, depending on the size of the cabinet. However, it is possible to use some alternatives, such as roofing materials, in the same capacity at a somewhat reduced cost.

-Chris

 

[Buckle-meister]

Incorrect. The cabinet baffles, of an average speaker, will have substantial acoustic output.

Interesting. I never thought it'd be so bad as to be audible, even on a less than 'average' speaker. Oh well. Live and learn.

It is an equal distribution: the failure of one will compromise the entire system.

I understand that, but I could just as easily say that failure of the speakers will compromise the entire system (in other words no sound at all!) in my room/speakers/hi-fi analogy. What I'm trying to get an appreciation of is what, assuming all parts are working, the relative importance of each is. With this in mind, do you still say they're all of equal importance?

Another option is to use large amount(s) of mass loaded dampening materials on the walls...

Like the material 'grania'?

 

[WmAx]

Interesting. I never thought it'd be so bad as to be audible, even on a less than 'average' speaker. Oh well. Live and learn.

In fact, the output is so high from lower quality floor standing speakers, that you can play music, lean your ear very close to the enclosure panels(almost touching), and hear the narrow midband contribution of the cabinet from the side panels. Now realize the large surface area of the total panel area. This contributes significant radiation. Think about this related issue, for a similar phenomenon: If someone has a stereo playing in the room next to you, and the door is closed, you can still hear it THROUGH THE WALL. That is usually 2 layers of 1/2" drywall(a fairly well dampened material) between you and the sound source. The majority of the sound is caused by the resonant modes of the structure(s) and wall(s). If you notice, the majority sound that breaches is not full range, but is distributed in narrow bands. Most of the full range breach is caused by air breaches between the the room you are in and the source room.

I understand that, but I could just as easily say that failure of the speakers will compromise the entire system (in other words no sound at all!) in my room/speakers/hi-fi analogy. What I'm trying to get an appreciation of is what, assuming all parts are working, the relative importance of each is. With this in mind, do you still say they're all of equal importance?

It will be up to you to define your weighting. My response was meant to apply to hi-fidelity reproduction, not basic operation. The connection wire is critical for basic operation....

Like the material 'grania'?

I am not familar with a material by that name.

-Chris

 

[zumbo]

On a midbass application, the port may output mid frequency resonances/noises. There for, it is preferable to have the port fire from the rear on such application(s).

-Chris

I agree. Even on a sub. If there is to be a port, I prefer it in the rear. Mainly, because I just don't like the nasty sounds that can come from them.

 

[Sheep]

I agree. Even on a sub. If there is to be a port, I prefer it in the rear. Mainly, because I just don't like the nasty sounds that can come from them.

It doesn't matter where they fire, you will hear port noise.

SheepStar