Identifying Legitimately High Fidelity Loudspeakers: Myths & Facts about Crossovers

[admin]

The loudspeaker crossover can be considered the brain of the loudspeaker. It directs the bandwidth of frequencies each driver is optimized to reproduce while it also level matches each driver and can help to stabilize the load impedance the amplifier will see. No matter how much science a loudspeaker company may tout leading them to certain design choices, without careful observation of their "science" one cannot be sure of its accuracy. It's easier and often more profitable to justify using cheaper parts or less elaborate design practices than to take the time and do it right. This article explores some of the myths and facts about crossover design. It also discusses some of the mistakes often made by loudspeaker manufacturers done either as cost savings or design incompetence. It is our hope that the reader will gain a better understanding of the mechanics of loudspeaker crossovers so they can make a more informed purchasing decision.

Discuss "Identifying Legitimately High Fidelity Loudspeakers: Myths & Facts about Crossovers" here. Read the article.

 

[templemaners]

Identifying Legitimately High Fidelity Loudspeakers Series

This has been a highly educational and informative series. Keep up the good work!

 

[GranteedEV]

the first-arrival crowd feels that the on-axis, first-arrival “anechoic” frequency response

Balderdash, says the far-field power response crowd.

In most rooms, it's probably really a combination of both. The axial response does dominate, but the power response still contributes. The first ~5ms will dominate, followed by the next ~10-15ms to stereo imaging , and then the rest afterwards (especially reflections after 30ms) will contribute to spaciousness.

The better crossovers definitely won't assume that the person is listening in a reflection-free environment or even with absorption at the first reflection points - IMO. But unless it's an omni speaker(where the power response will dominate in most setups), the response will be very flat

 

[gene]

In most rooms, it's probably really a combination of both. The axial response does dominate, but the power response still contributes. The first ~5ms will dominate, followed by the next ~10-15ms to stereo imaging , and then the rest afterwards (especially reflections after 30ms) will contribute to spaciousness.

The better crossovers definitely won't assume that the person is listening in a reflection-free environment or even with absorption at the first reflection points - IMO. But unless it's an omni speaker(where the power response will dominate in most setups), the response will be very flat

Early reflections dominate in small rooms as per Dr. Floyd Toole's extensive research. Of course a speaker exhibiting a smooth and even power response will also likely produce good early reflections since its off axis response is consistent.

 

[GranteedEV]

Early reflections dominate in small rooms as per Dr. Floyd Toole's extensive research. Of course a speaker exhibiting a smooth and even power response will also likely produce good early reflections since its off axis response is consistent.

The worst thing about a small room isn't even the side wall reflections (they they rarely help at short distances), but the reflection from directly behind you. To compound that, in small rooms people shove their couches flush against the wall.

In such a room with all that comb filtering ( ) then the power response should definitely be good in order to fill in for a lot of the response error.

 

[gene]

The worst thing about a small room isn't even the side wall reflections (they they rarely help at short distances), but the reflection from directly behind you. To compound that, in small rooms people shove their couches flush against the wall.

In such a room with all that comb filtering ( ) then the power response should definitely be good in order to fill in for a lot of the response error.

Sidewall reflections are beneficial NOT detrimental. I agree most people stick their couches against the backwall which is bad for smooth bass response and a realistic surround field.

 

[GranteedEV]

Sidewall reflections are beneficial NOT detrimental.

At everyday SPLs Toole's research shows that it seems to help improve intelligibility and give the impression of higher SPL. But are they unequivocally benificial? I think that depends on the distance involved. 2ft, 6ft, 10ft... i'd think each of these distances to the side wall would impact the sound differently.

Anyways I'm probably dragging this off-topic.

 

[AcuDefTechGuy]

Early reflections dominate in small rooms as per Dr. Floyd Toole's extensive research. Of course a speaker exhibiting a smooth and even power response will also likely produce good early reflections since its off axis response is consistent.

Does that mean even an omnipole speaker could have a good on-axis and smooth off-axis response?

 

[GranteedEV]

Does that mean even an omnipole speaker could have a good on-axis and smooth off-axis response?

I think so, but its placement would be VERY unique for a lot of non omni instruments to sound good in rooms.

Which is why I think the B&O omnis measure so awkwardly on axis but have a smooth power response and supposedly sound great.

I think omnis are an extreme, just like 45 degree horns in super treated home theaters . You need to find a balance that works!!

 

[agarwalro]

Balderdash, says the far-field power response crowd. The room reflections etc. are so close in time to the first arrival that they swamp the first arrival completely, leaving the overall far-field energy output of the speaker—the so-called “power response”—as what you hear to determine the tonal quality of that speaker.*

Gene, hypothetically every speaker will have as many power responses as there are rooms. Is there any reference criteria for good vs bad power response? Or, does it boil down to a subjective preference of imaging (meaning good first arrival behavior) vs. soundstage (meaning good room reflection)? (Assuming that said speaker falls into the "accurate" generalization for its near field behavior and listening position or room or media does not favor an overly dead or reverberant space.)

So all those near-field considerations like driver alignment, cabinet diffraction, phase relationships, etc. are far less important in real world listening conditions than their theoretical importance might suggest .

On the contrary, wouldn't these constitute a minimum criteria for loudspeaker quality? It would be a logic defying speaker or "holy smoke magic" conditions in which a poorly designed/implemented product had good in-room sound reproduction behavior .

 

[agarwalro]

Does that mean even an omnipole speaker could have a good on-axis and smooth off-axis response?

Based on power response being an in-room measurement, I'd say it would be a speaker optimized the direct and reflected energy. Behold! We have a well placed Phil or SS. Hmmm... I'm f*ed.

 

[GranteedEV]

every speaker will have as many power responses as there are rooms.

I think there's something to be noted, in that sound power response and "sound that we hear from speakers in rooms" - are still not the same thing even though the latter is greatly infleunced by the former.

Sound power is an anechoic, imaginary, and maybe measured but mostly "calculated" concept. Think of it as the sound being radiated into all 360 spherical directions... averaged.

What we hear in room is first the direct response, then the reflected response and it doesn't all arrive at the same time, minus whatever was absorbed or reflected by every object or wall in the room.

So rooms have their own effect but that's a bit different from the speaker power response in terms of the crossover. The crossover does a few things to what the speaker is radiating..

1) It affects the lobing behavior of the speaker. Even a speaker that's flat on axis, will likely have various peaks and dips off axis from interference between drivers, because their center to center spacing is rarely close together enough. Coaxials are of course less guilty of this as they are a true point source. The lower the order of the crossover, the broader the frequency region of the lobing. Even 4th order crossovers look pretty bad as you move away from the design axis. 110db/oct or higher look better IMO as they operate over a very narrow frequency range. For these very high order filters to work you probably want linear phase active crossovers, and you probably want VERY transparent and SIMILAR drivers so that the "difference" between the drivers isn't revealed by the lack of overlap (bear in mind, 4th order is 24db/oct so 110db/oct is bordering on ridiculous). Here is one speaker that used this kind of crossover:

http://www.audioholics.com/reviews/speakers/satellite/nht-xd-loudspeaker-system-review

2) It affects the total radiated sound in the crossover region - odd order crossovers tend to have a smooth power response, but can have a a peaky summing between drivers somewhere off axis (assuming the on axis is made flat) because (-3db) + (-3db) in phase sums to (+3db) somewhere (rarely on-axis). Even order crossovers tend to have a power response dip (which arguably we don't notice as much especially with a higher order crossover) which means that the on and off axis response will be pretty smooth, but the tweeter, and midrange for example, are actually not radiation as much combined sound near the crossover point, as they are at other frequencies. This is because they're both (-6db) + (-6db) which is what's needed to sum to (0) on-axis. Mathematically, even coaxials have this problem. It's pick your poison. Have two drivers reproduce the same sound at the same, but reduced level so they add up flat at one point, or have two drivers reproduce the same sound at different levels but such that they add up flat overall with the caveat that there will be some peaking somewhere??

3) A crossover means that you're using multiple drivers each with their own directivity index. So the crossover should attempt to match the directivity indeces. The drivers themselves have a power response and it should be reasonably matched. This is why many higher end speakers like Genelec, Revel, JBL LSR, Pioneer/TAD, KEF all restrict their tweeter's off axis response - because the woofer's natural dimensions are restricting this.

So when we refer to power response, we're refering to what the speaker is radiating. What the room absorbs or doesn't, is definitely REAL - it's just out of the control of the speaker designer. The speaker designer's goal should be to put the final buyer in a position to succeed as widely as possibly. Ideally there's custom designed NEUTRAL rooms as well. But for non neutral rooms, speakers shouldn't compensate(beyond tweeter level/baffle step controls), and for non neutral speakers, rooms shouldn't compensate (that is.. you shouldn't have to cut off half of what your speaker is radiating for it to sound good)!

Based on power response being an in-room measurement, I'd say it would be a speaker optimized the direct and reflected energy. Behold! We have a well placed Phil or SS.

The soundscapes and philharmonics, while they have very wide dispersion, i think are still well away from being considered "omni". The measurements you see are what, out to 50 degrees? They're fantastic. But omni speakers look the same at 0 degrees that they do at 180 degrees off axis, and on paper at least, vertically too (though not in execution as that would be a tiny 1/2" full range ball!)

On paper, an omni speaker radiates the same sound in every direction (thus in phase unlike a dipole or an open back mid) - a theoretically flat power response. So most omni speakers are actually voiced with a tapering axial (any axis really) response. This tapers the power response to more closely mimic box speakers.

Toole's research shows that in most rooms, not a flat, but tapering power response was shown to sound realistic. Damned if I know why (probably a mix of the speakers the audio is mixed on, plus the way flat response + tapering power response simply sums up and because perhaps some directivity is desired)

 

[kevon27]

Question.. Are the crossovers used in Emotiva speakers of good or bad quality?

 

[cpp]

It's a shame Omni speaker types are so expensive as noted by the RAAL Eternity which cost more than my car or the Wolcott Omnisphere at $7,000.

I do think crossovers mater a lot for efficient speaker design, as well as room dimensions and room layout that support the design of the speakers internals but in the end it's the ear that counts, the components of the system and your wallet unless you just want to be techy and constantly change speakers every time a new on comes out.

Good post and very informative.

 

[zhimbo]

Question.. Are the crossovers used in Emotiva speakers of good or bad quality?

Labeling a brand as "good or bad" in this thread is probably not a good idea - better to talk about the principles and design. But you can get details on the crossovers of mosts speakers actually reviewed (or sometimes even previewed) by Audioholics if you search.

 

[bootman]

Question.. Are the crossovers used in Emotiva speakers of good or bad quality?

You can judge for yourself.

Emotiva ER 5.0 Speaker System Review — Reviews and News from Audioholics

Is there a specific model you are concerned about?

 

[gene]

Question.. Are the crossovers used in Emotiva speakers of good or bad quality?

Emotiva takes its time to design quality crossovers that don't display any nasty low impedance dips and they employ good parts and layout. Vance Dickinson does their crossover work and he is considered a legend in the field.

 

[bootman]

Any plans to comment on crossover less designs like Gallos and orb audio?

Rare to find any graphs on these types of speakers. (especially Orb Audio)

 

[gene]

At everyday SPLs Toole's research shows that it seems to help improve intelligibility and give the impression of higher SPL. But are they unequivocally benificial? I think that depends on the distance involved. 2ft, 6ft, 10ft... i'd think each of these distances to the side wall would impact the sound differently.

Anyways I'm probably dragging this off-topic.

I agree. Also just how much sidewall reflection do you really want? It is definitely a balancing act. It's funny. Years ago when I was setting up my theater system in my family room, (long before Toole's book came out), I decided to put absorption on the sidewalls b/c classically that was done for large room acoustics. I found in doing so that the soundstage seemed to collapse and the overall sound was too sterile. I wound up leaving the sidewalls untreated but I didn't give it much thought until Dr. Toole published his fabulous book that really dove into the topic of early reflections and how they dominate the sound we hear in small room acoustics.

 

[gene]

Based on power response being an in-room measurement, I'd say it would be a speaker optimized the direct and reflected energy. Behold! We have a well placed Phil or SS. Hmmm... I'm f*ed.

Power response is typically done anechoic in 10-15deg increments spherically around the speaker. The averaged results are then weighted according to the proportional area of the sphere represented by each measurement. There is a really good article on this I’d love to attach but I haven’t asked the author permission to do so yet.