[bigshot]

We talk a lot about the science behind digital audio and electronics, but not so much about acoustics. I've been working on my speaker system lately and through trial and error, I've stumbled across something that works really well, but I'd like to understand the science behind it a little better.

My system is a blend of wide dispersing speakers and more directional horn loaded speakers. I understand the stuff about on and off axis, but I'm interested in the difference in sound between a wide dispersion speaker and a directional one when hearing them on axis. My theory is that wide dispersing is best for rear channel speakers, which have to bridge a wide space to create a phantom center in the rear, and a combination of directional and wide dispersing is good for the mains which have more sound localized hard left or hard right, but still have to bridge in the middle. For center channel a highly directional speaker is good to add presence to vocals and dialogue in movies. Voices seem to cut through better that way. I'm not sure how dispersion affects room reflections. Would it be similar to the difference between focused and diffuse light? I'm also wondering if horn loaded speakers are better in specific frequency ranges (i.e.: treble) or if they work the same with lower frequencies. Would it be possible to create a highly directional subwoofer? Not sure what application that might have, but it's interesting to think about.

There's definitely a difference in sound between a directional and a widely dispersed sound. Does anyone have any good resources for web links that talk about the differences and how to apply them in sound systems?

 

[Strangelove424]

Acoustics can seem like voodoo sometimes, but to my quite limited understanding your metaphor in the visual lens is accurate. With a certain amount of energy, how does it get focused or diffused? Engineering can bridge the compromise, like a good lens that’s capable of taking a wide angle shot with good color, sharpness, and no barrel distortion. The extent a speaker can disperse sound energy without sounding dull or “out of focus”, losing its ability to image a convincing soundstage, the better it is considered to be. But all that is based on so many different factors, including size and price, which most people have no major choice over. The choice, as always, is where to put the money and how to use technology to our advantage. That’s what fascinates me about speaker technology.

There’s some basic issues for dispersion of sound in a speaker, and they’re often tweeter-related, or more so how the tweeter and woofer interact, the crossover having a lot to do with it too, as well as our ears' fluctuating sensitivity to different frequencies. The first problem is that the higher the frequencies, the shorter the wave, the more directional they become, hence the importance for tweeter to be aimed at your ears. I don't think a directional subwoofer would be possible due to the length of bass waves, but at the same time I think the non-directional nature of bass has been overemphasized, and there's good reason to use two in symmetry. The second problem is that of timing (and its relation to reflection) since tweeters are often mounted differently than the woofer, and can create arrival time differences. There’s some rather innovative ways to come up with solutions here though, your KEF R100s being a great example. They set the tweeter into the woofer with a concentric driver, so bingo timing problems solved. They also directed the tweeter, using a waveform guide of sorts to disperse the high frequencies more effectively. KEF is quite astute about acoustics, and invested in acoustic measurements early. Klipsch’s horn solution is great for dispersion, but the issue with the loaded horn has often been resonance at certain frequencies. For music that has a natural horn dynamic like a trumpet it’s hard to beat a horn’s imaging though. I almost bought a pair of Hsu loaded horn bookshelf speakers after hearing the Saints Go Marching In on them. They were only $250, and for the first time ever I felt like a real woodwind was in the room with me. But there’s other ways to solve dispersion issues too, like separating the tweeter completely from woofer cabinet like B&W does, or Magnepan’s solution, which is known to create an expansive soundstage with lots of detail, but lacks low end. (Are they really just giant tweeters?) However, you don’t have to worry about timing when the driver is one monolithic sheet.

The simplest thing you could do to improve dispersion is to align your speakers as best as possible so all the sound sources are equal distance to the listening position, and arrival times are as close as possible. Reflections are difficult to account for, and are best handled with mics and software these days (or good ole accoustic treatment) but your best bet to creating a cohesive soundstage would be to work on arrival time as best you can, particularly from the fronts/center. Your emphasis of where to place focus on your surround setup sounds spot on to me.

I hope these links are useful.

General info:

http://www.neumann-kh-line.com/neumann-kh/glossary.nsf/root/E739BF069CE5E3F7C125728C006784FC?Open&term=dispersion pattern

A dispersion calculator that shows the mathematical relationship between dispersion/direction (I’m sure advancing technologies can change the efficiency of this ratio):

https://geoffthegreygeek.com/calculator-speaker-dispersion/

A sales brochure for some speakers, but covers the problems with dispersion in detail before the sales pitch starts.

http://www.actem.de/upload/The_advantages_of_using_waveguides[2].pdf

This is an interview with the former engineer of KEF and TAD, who also designed my current speakers. It’s a discussion about the process of designing speakers, a lot of KEF history, and the goal to create 3d output. He also talks about concentric drivers, like your KEFs have. I really think you’ll dig this.



Not trying to get you to buy anything else after your speaker upgrade, but I really do think accoustic treatment might also be a good thing to start considering. It will help both dispersion and imaging. I couldn't believe the difference I heard here with some acoustic diffusion at the end:

 

[bigshot]

Thanks! That interview answered one of my questions. The tweeters in my 70s JBL mains are bullet tweeters with an exponential wave guide. He explains why wave guides on tweeters create directionality and improve sound. That's a bingo.

It seems that a lot of the theories on timing and reflection are based on a simple two channel setup. They assume the distance between the speakers and seating distance and move right on to working out the optimal layout of the drivers in the speaker cabinet. But a multichannel speaker system is MUCH more complicated than that.... especially in home theaters, where the scale of the sound needs to relate to screen sizes that might vary from 24 inches all the way up to 120 inches. That means that a standard 6 to 8 foot distance between speakers and equilateral triangle arrangement for seating isn't going to work all the time.

When they talk about room reflections, it's a lot simpler to speak about a standard 2 channel setup. Multichannel is going to have several times as many reflection points and varying arrival times. You also don't have just one direction that direct sound comes from, so direct sound is blending with reflected sound all around you. My theory on that is to just worry about primary reflections and let the rest be the natural quality of the room. He is absolutely correct that ears don't like anechoic chambers. Although heavily treated rooms can have more clarity of sound because muddling reflections have been eliminated, it can sound considerably less present and live than a room with a complementary natural acoustic. That's why symphonies perform in concert halls with natural reverberation instead of performing in soundproofed auditoriums with acoustic panels blanketing the walls. I think multichannel audio might relate more with concert hall acoustic theories than traditional home stereo ones. In fact, since you have sound coming from all directions, it might need to have its own set of theories.

If I understand correctly what I've arrived at through trial and error, I've got very time accurate and widely dispersed sound coming from my rear channels (concentric drivers). This creates a wide focused sound field that bridges the wide gap between the speakers (almost 20 feet). Up front on the far left and right, I have my 70s monitors with synchronized bass and treble because of the exponential wave guide on the tweeter. This provides highly focused directional sound at the extreme sides of the room... about 16 feet apart. Between those two directional mains, I have vertical towers which are less directional filling out to the right and left bridging the gap between the mains. These are about 10 feet apart, and are more midrange focused, so the high frequencies don't get muddled mixing with the studio monitors. In the center is a horn loaded speaker that has a very directional sound. Set between the diffuse towers, it's like the cheese between the slices of bread- it bridges the gap in the middle for both sets of mains, but it has a distinct horn loaded sound of its own. This perfectly sets off vocals in multichannel music and dialogue in movies so they have their own clear separation. Surprisingly, this distinct sound isn't apparent in orchestral music. The sound field feels unified all the way from left to right. That may have something to do with percussion, woodwinds and horns being seated in the center of an orchestra. Those are instruments that would benefit from having a clear separation and directional projection.

The other day I was listening to a jazz album that had a bass solo in the left hand speaker. It was clearly focused on the left, even though my subwoofer is on the right. It seems that the aspects of a plucked bass that indicate direction are all above the crossover at 80Hz. I may have been hearing 80 and above on the left and 80 and below on the right, but to my ears, it all seemed to be coming from the left. A lot of home theater people seem to think you need two subs. I really don't see why unless your crossover is significantly above 80Hz.

Directionality is the thing that sets multichannel apart from both headphones and 2 channel speaker setups. I think there might be some theories that go beyond normal 2 channel room acoustics. In fact, it may be that a whole different set of priorities apply when you have so many sound sources in a room.

 

[Strangelove424]

No problem, I just enjoy being able to discuss this stuff. You are correct about surround sound setup. I don't think its 100% understood yet, which is why so there are so many variations of surround speaker locations, toe angles, EQ/timing algorithms, and room layouts. As Andrew said, "If I knew what I was doing, it wouldn't be called research." That still appears to be the case for much of audio, and surround sound never even gets respect as an important standard. It certainly never replaced stereo to the same extent as stereo replaced mono. There are many albums that would make incredible 5.1 mixes, but they're so rare. Much of the 5.1 complexity is dependant on each person's particular room, and all the compromises or benefits that go with your own acoustic space. Reflection is well understood, but getting an acoustic map of each room is the hard part. Generally, speakers are tested individually, so surround placement and room interaction are a few steps of consideration down the road for a speaker engineer, but it's probably something that deserves more attention. Additional speakers, when harnessed correctly, can reduce IM distrotion and could probably increase the coherence of reflections too, with enough calibration. There's an audiophile myth that lots of speakers = cacophony, and stereo is the only purity (sound familiar?... it's mono->stereo all over again) and that's not true, it's just how the speakers are implemented.

I have one major rule I try to follow based on the idea that the back wall creates the most incoherent reflections - don't let sound go the opposite way it was originally travelling. So I try to control reflection of the front speakers off the back wall best I can, and angle in the rear speakers so their dispersion or wall reflection have the least effect on front sound stage. I'll admit, it's a little busy back there during some passages, but I'd rather it get busy back there than up front, and the transitions around the room are all smooth. Airplanes sound just like they're actually flying overhead. I realized at some point I didn't have to worry about filling the room's space with sound, I only had to worry about creating a convincing sweet spot, and from that spot the whole room will sound like it's filled with energy, even if it's somewhat of an illusion. So that airplane sounds exactly like it's flying overhead the entire house... but not from 5 feet over.

Your sandwiching idea is a good one, I think you are using the speakers you have as intelligently as you can. It wouldn't make sense to bunch the focused speakers all together and space out the mid-emphasis at the extremes. You are putting the speakers that are strong at imaging where they belong: at the dialogue center to anchor the picture to the sound, and at the extreme of the soundstage for increasing expansiveness of the effects. The mid-frequency woofers placed between will widen dispersion and 'fill in' the soundstage without becoming too directional. The only thing I might experiment with is maybe raising the height of the extreme L/R focused speakers slightly above the mid-placed ones. Just to play around.

Yes, the the room must be considered part of the speaker system, especially in the lower frequencies. In a certain way, I like to think of the actual 'cabinet' of my system as the room it resides in. Could you hear how dead absorption treatment sounded in Ethan Winer's video? I know he poo-poos book shelves, but they do sound better to me than straight up absorption, although not as good as his diffusers obviously. Just trapping all reflections anechoic style is not a great way to make something sound good.

Most subs are at or below 80hz crossover in a full-range system, which on its own shouldn't be directional. The justification for two subs I think has more to do with how troublesome sub placement can be. The length of the bass waves makes room interaction a royal pain. I have a weird listening room - it seems to keep the bass isolated to half the room no matter where I put the sub, so I keep it as mid-placed as possible, almost nearfield. Directionality is not so much a problem seated, but if I get up and move around doing stuff, I can hear the bass changing through the room. It would be sort of ridiculous to use two 200w subs at 10% power each though, and I don't have the space for them anyway, but I can't say I would mind the extra coverage.

 

[bigshot]

I have one major rule I try to follow based on the idea that the back wall creates the most incoherent reflections - don't let sound go the opposite way it was originally travelling.

The only thing I might experiment with is maybe raising the height of the extreme L/R focused speakers slightly above the mid-placed ones.

Thanks for the comments! I have bookcases on the front and back walls, and I have couches on the side walls at the primary reflection point. My roof goes up to a point, it isn't flat, so I think that helps divert some reflections into the rafters where they don't matter.

Raising the outside mains might be tricky, because that JBL tweeter has a pretty narrow on axis range. The center speaker and rears are raised above ear level though, which tends to pull the soundstage up higher without having anything really important off axis. The center sounds like it tends to throw out vertically rather than horizontally, so it finds its way down to ear level OK. The rears are tipped down a little, but they aren't pointed right at the listening position. They have that wide dispersion that makes that OK too.

By the way, I spoke to a friend of mine who does PA systems for concert venues and he says that bass can be made highly directional if you use horn loaded speakers with an exponential design. To get down to 20Hz, the horn has to be something like 17 feet tall, but he's designing a modular horn system to do just that. The advantage is that in outdoor amphitheaters, you can focus the bass like a laser beam at the audience, and not neighboring houses in the hills to the sides. That's a big problem at the Hollywood Bowl and the Greek Theater here in LA.

 

[Strangelove424]

Don't take the tweeter too far out of ear range. It was a suggestion that I thought might increase soundstage height and separate frequencies for imaging, but not worth taking tweeters out of ear range. If it sounds good, just enjoy it.

Whoa! 17 feet... I wonder if that's even safe to aim at a structure, or how close you could stand in front of it. How do they plan on going about unleashing this fury on a helpless audience? I cannot even fathom the feeling of that because my experience of subwoofers is essentially feeling the bass from all the solid structures around me. But sitting in a wide open space, and still feeling a low frequency wave from pure space must be intense.

 

[bigshot]

He is estimating that unobstructed the bass could carry for as far as a mile in just one direction, but the volume to the audience would be pretty much the same from the front row to the back row. It's focused directional sound, not necessarily loud. He gave me a demonstration in his living room with 1/4 of the speaker pointed out the front windows and doors. It had to be short because he has neighbors a block away in that direction and their house was probably full of Paul McCartney bass lines.

 

[Strangelove424]

It's amazing what focusing can do. I can't begin to imagine all the challenges associated with large scale open air audio reproduction. The sheer power requirements are mind blowing. And we complain about the complexity of structures, but it must be even worse not to have any at all! Or to deal with reflections that sound like the grand canyon.

 

[abm0]

The other day I was listening to a jazz album that had a bass solo in the left hand speaker. It was clearly focused on the left, even though my subwoofer is on the right. It seems that the aspects of a plucked bass that indicate direction are all above the crossover at 80Hz. I may have been hearing 80 and above on the left and 80 and below on the right, but to my ears, it all seemed to be coming from the left. A lot of home theater people seem to think you need two subs. I really don't see why unless your crossover is significantly above 80Hz.

That was in your room, to your ears, with your crossover filter and with that particular song. Professionals have seen many more examples and that's why they end up recommending something different as "good enough for most people". 80 Hz is a bit high, for example; a better crossover for most contexts is 60 Hz if you're going to use a single sub. Also the crossover slope can make a difference: 12 dB/octave is pretty bad, as it lets audible amounts of "wrong-frequency" material into both of the channels it's supposed to separate, and then the sub will give away its position by reproducing things it wasn't supposed to reproduce. Also the sub should be high-quality, i.e. with very low THD, otherwise, again, it might give its position away. So the choice seems to be roughly between an excellent sub + an excellent crossover on one hand, or whatever crossover you can get and whatever 2+ subs you can get, if the goal is to preserve correct soundstage down to the bottom of the spectrum.

 

[bigshot]

I tried crossing over at 60 once. It didn't sound as even of a hand off. I would have had to boost the EQ on the mains and I didn't want to do that since it was working so well at 80. Two octaves to the sub was just perfect.

 

[abm0]

I tried crossing over at 60 once. It didn't sound as even of a hand off. I would have had to boost the EQ on the mains

Well nobody said you'd be able to do this equally well with any speakers - ideally they should have flat response all the way to the crossover point and even a bit beyond.

 

[bigshot]

The thing is, most speakers have more controlled response in the middle. Even if they produce sound in the lower register, it's not going to be as flat. If you crossover below that point, EQ becomes more complicated. Subwoofers are designed for the lowest two octaves. It's what they do best.

 

[abm0]

Yet another reason to prefer the dual-sub setup: among the many constraints it relaxes for you - while still giving you reliably correct soundstage - is the crossover frequency.

 

[bigshot]

I really don't know why I need a second sub when the one I have fills the room and is able to work with my mains crossing over at 80Hz without affecting the directionality. I think the only reason one would need a second sub is if their room layout isn't optimal or the sub isn't powerful enough to fill the room. I have a 12 inch Sunfire True Sub and I have the thing dialed way back to make it balanced. I'm sure it could fill a small nightclub all by itself. I know that it can make all the walls in my listening room vibrate like mad. It totally fooled the auto EQ in my Yamaha AVR. It pumped out so much sub audible bass that the auto EQ freaked out and tried to dial it all the way to zero. I ended up using the Sunfire setup and then adjusted the crossover to the mains manually. My mains have 15 inch JBL woofers that go low, but the Sunfire goes down well below 20Hz and it's clean and flat all the way up to 120Hz. 80Hz and one sub work perfectly for me.

 

[abm0]

I really don't know why I need a second sub

You personally don't, it seems. I was discussing the general case.