Rerurn to Romy the Cat's Site

This there is the last in the “THE PROBLEMS WITH HORNS” cycle. The previous threads might be found at:

Problems with horns: tweeters.
Problems with horns: mid-range horns.
Problems with horns: mid-range drivers.

This would be probably the most controversial and the most complicated article form “The problems with horns” cycle. Really, the middbas horns is a very complex subject where is EXTREMELY DIFFICULT TO MAKE ANY GENERALIZATIONS. First of all let define the middbas. For the sake of this article I would call middbas the frequencies from 45-50H and up. There are reasons why I would not define the upper threshold of middbas and you will understand further why. Also, it is important to understand that I do not really believe into a practical lower-bass reproduction by horns and there are many reasons why.

OK, the upper bass channel of your playback. I always say: “Show to me your middbas solution and I will tell you how your playback sounds”. I do not kid about it. The middbas is the most difficult, complicated but the MOST IMPORTANT RANGE of any playback and if you have a dedicated middbas channel then you better really take care about it because it will be a foundation of your entire system. If you do not have a dedicated middbas channel then it is most little that you have no good middbas, at list it was my experience with quite a number of playback systems to which I was exposed. A middbas channel sets up a framework, the skeleton of entire sound of your system. 99% of all bad sounding systems sound bad primary because thier major problems within middbas region. Would be too much to say that the middbas is the very first thing that I listen when I face any new system?

Well, the middbas DOES NOT REALLY EXIST in live music, however within reproduces music there is deficiency of middbas and there is a wrongly-made middbas. This is the key: to make the middbas “to sound” but at the same time to make it completely “unsoundable”. Does it sound like foolishness to you? Not really. The foolishness is how middbas sound within a typical audiophile system – wherever I heard it has a sounded of retroaction efforts instead of the “unsoundable” middbas. Also, there is such a concept as ”amplitude of the unsoundable middbas” – if it is not enough then it is not enough. Ironically there is no reveres situation and the “unsoundable middbas” can’t have too high amplitude.

 “ Ok, let stop to sing the “Ode to the Middbas” and let jump to some more or less practical discussion.

There are some fundamentally faulty sound reproduction methods that even theoretically can’t produce middbas. First of them would be so-called full range drivers. They idea of so-called full-range driver operation is so ridicules, so idiotic that it is no surprise the it has a lot of followers within audiophile community. They use the upper midrange drivers like Lowther, Fostex, Supravox, Jordan, LES, AER, PHY-HP and many other.

They force those drivers to go down to the range where they have no business to operate and then back-load those drivers with some curved horns horn. All of those full-range/back-loaded speakers sound incredibly horrible. Let me to street it again: all of them!!! Those loudspeakers violate the first too rules of the “Commandments of Mandatory Middbas Compliance”:

1) A middbas driver shell handle at least one octave below it lower cut off frequency
2) A dipoleness shell not be used anywhere near a middbas channel

The next candidate for middbas reproduction would be all versions of the box-reflex design.  Everything in ported sound is wonderful accept the fact that the ported enclosures kill all sound under 700Hz -800Hz. The third rule of “Commandments of Mandatory Middbas Compliance” said:

3) You shell run away from a speaker if you see a port

Certainly the Third Commandment is a derivation form the Second Commandment not to mention the countless problems that attached specifically to the ported enclosures. It would be worth to mention that the systems with passive radiators are essentially the same ported systems and they are full under a jurisdiction of the Third Commandment.

Evan the very best of the ported designs with the price tag close $100K (the largest, Wilsons, JMLabs, Avalons and many others) does not do correct middbas. I have to admit that in some very-very-very-vary rare occasion a very good ported speaker when it was painstakingly and supportably anal retentively set up in a room then it might demonstrate some almost acceptable result. However, they do it only for a very-very small “middbas” zone (geometrically it might be juts a few inches in your room) and second: they are literally a few installations around the world (I mean: a few) where such a level precision of setting the speaker up was taken. Still, in the end of the day, those installations’ middbas does not sound as good as middbas might be.

The next candidate for middbas would be an opened baffle. I know that there are many-many people out there who would swear about the fanatic quality and countless (mostly simulated) benefits of opened baffle. Unfortunately I have to disappoint you (and disappoint myself –because I spent quite time to learn it). The opened baffle is a FUNDAMENTALLY FAULTY WAY TO HANDLE MIDDBAS and not only because it violate the Second Commandment. The opened baffle (or any dipoleness) removes seriousness form Music and converts Sound into a stressful agitation of air in a room. Yes, it sounds “attractive” and instantly gratificative but in a long run it is a dead end of middbas reproductive efforts.

The next group would be the direct radiators in sealed enclosures. This probably is the most advantageous group that is capable to do better then the mentioned before groups, however in the real practical world they do not do better. There are many reasons and many variables involved into a proper sealed middbas:

Selection of a driver, learning how to use the driver, learn how to integrate the driver with mid-range, creating a back-chamber that would be suitable for this driver and many-many others, pay attention that all the I said is too obvious but it is very complex. For instance any person can look at T/S data and to design correct enclose to the driver but it will have a little relation to Sound. It will certainly assure that the driver would operate in it’s not-completely faulty mode but this is juts a beginning. There is no know ways to project the “minuteness” of a driver performance and therefore the are no ways to get the driver into it’s operation point via extrapolating it. A driver should be tested and evaluated sonically and using some proper evaluating techniques with are practically not at the disposal of the speakers builders. Considering that: there are virtually no good mid-bass drivers out there in existence, people who know what T/S is are completely clawless about proper listening techniques, there are no objective Sound Assessing Methodologies (at least know widely) and many other factors i… t is understandable why there are practically no good sounding direct radiators middbas solutions. Also, for whatever reasons, the lower-middbas in a sealed, even severely under-dumped incisures, with the best drivers sounds different (quality-wise) then the upper-middbas. What I mean is that I nave not seen any full-range middbas channels the would do identically-well 60Hz and 400Hz. Probably this all caused by the fact that no one produces any good middbas drivers that mend to be use ONLY for middbas and as the result people are forced with use whatever inappropriate crap they were able to dig out.

Well, would it bee so difficult to foresee that we are approaching in our observation to the horn-loaded middbas?  Yes, the horn-middbas is the only middbas know to me that could be OK. There are countless ways to frond-load the drivers. Let talk about them and what result it might lead.

We all understand that when you go for middbas then the size of horn grow faster then the President Bush’s budget deficit and therefore folding of horns looks like the reasonable solution. Unfortunately the horn folding has a huge number of negative moments the defeat the purpose and in very many cases the good underhanged direct radiators sound better then folded horn. A folded horn IS NOT A HORN in but just a consecutive chain of resonant chambers where each following section impose to sound a set of own rules.



It is similar to a person who does not hear his own speech but rather hear it as a translation form his own language to 5-6 different successive languages and then translated back to his own language. The each fold of a banged horn is a resonator that affects (screws) sound in own way and the all this mess get mixed and going out of the folded horn’s mouth. Also, each fold of the folded horn creates a HF roll-off and the harmonic content of sound after all those folds is kind of strange: it is always too dry. The more folds the stranger the HF knee and as the result, it is imposable to properly integrate the folded horn with an upper frequency channels. I think the problem in here that the type of the HF roll-off the takes place inside of the folded horn juts do not exist in a really world and we are not tuned to register it with out hearing. Therefore the folded horns always sound too much like an artificial sound reproductive effort not to mention numerous purely audio problems that folded horns introduce.

OK, we successful bitched about everything else and now we arrive to the straight, front-loaded upper bass horn. Fist thing first: from here and on I will be mention a chamber. Be advised that I will always be taking about the back-chamber of a horn. The front chambers, as it described in most of the incorrect horn literature, SHOULD NOT BE USED anywhere within horns. Most of literature form the mid of the last century that you probably read and that stressed the use of front chamber approach to horn design form a perspective of developing pressure but not from a prospective of development sound. Therefore some of thier observations are not applicable in case of high-fidelity home reproduction. Anyhow, if you driver too large and your throat too small (than might occur with middbas horn) and you have some kind of accidental front chamber in there then juts disregard it and move forward. Yes, keep it smaller but disregard it is have happened.

Eventually: The Problems with the midbass horns - the horns.

If with mid-frequency horns the profile is quite important, with the HF is fatefully important (up to the point of being mysterious) then with the middbas the horn the profile kind of less critical. Any exponential curve could work, even a straight conus but the entire horn should be deigned and used properly. The curve will dictate the geometry of the horn. The larger mouth = more LF, the smaller throat = longer horns, the longer horns the more sound get EQed down. The more Sound get EQ down the more “interesting” the horn loads the room. Each single horn (profile and size) will demonstrate own behavior regarding room coupling and there are NO WAYS to foretell the result, at least know to me ways. Some horn might work great in a corners, some would like to see one single boundaries, another would like to be placed in a middle of the room, or even lifted up form a floor…. and so on and so on… Do not forget that behavior of a middbas horn DO NOT exist  “as is” and without it being fully embraced by the sound of your room. Therefore to talk about a middbas horn within considering the room is a foolish thing to do – and as you understand to talk about a room we can infinitely.

Next is a rule: middbas horn should be massive, its walls (would it be a direct radiator horn or not-dipole corner-loaded horn) should be strong and as heavy and you can afford.  You can use the walls of the horn as a creative resonators but here you at your own as the result become completely unpredictable. 

Next: a horn might be spherical of rectangular. The spherical horns integrate better with high-range and image way better the thier rectangular brothers…. The middbas might run a full profile of its rate of to be truncated. It also might run a combinative profile and to start with one curve and continue with another. I have difficulty to propos that one way is better the other. I do not care about the theories but I never seen anyone built two huge horns with different curves and then able to properly evaluate the. Usually people build a horn (whatever the decided to do intellectually) and then use the result “as is” trying to get the best from whatever the horn does. Even if people do bult two identical horns with the different profiles then each of them would act different in the same room and therefore to say anything defiantly is tonally unnecessary. What I really hate is when people get an accidental result and than build artificial theories around those results. Unfortunately considering the complexity, size and the price of the manufacturing of the middbas horn and considering that each of them demonstrate its own behaves in each room I do not believe that there is an only “best way” to make an upper bass horn. Therefore this portions of “Problems with horns: is more liberated then all others.

The biggest problem that usually people face when they listen the middbas horn is that people tend to make the middbas horns too short. In the short horn the driver shot right through the horn and the LF go not “get loaded” with the horn mouth. For instance the JBL 4560 is a typical size middbas horn, the size that most of audio people use.




The 4560 would give you with a correct driver ~120Hz but it will be only 200Hz of horn loading and the rest below will run as a direct radiator. If you put in this type of enclose a driver with lover resonant frequency you get even lower response from it but it has no relation to horn loading. Most of the popular Altec horns are even shallower then the 4560. For instance audio people stick into the Altec 5-9 the 515B with Fs hear 20Hz and they happy to get “bass” out of this speakers. In reality the horn loading in there do not go lower then 200-250Hz and the rest is juts sound of mass-centric direct radiator driver boosted with a port.

My rule to detect that horn does is to analyze the EQ at lower frequency. For instance I use in my horn a driver that without baffle rolls-off somewhere at 650Hz or at 250 in a baffle and has Fs=86Hz (designed to be used at 400Hz-8Khz). This driver being placed into a horn get EQed down 120Hz anechoicly and ~100Hz in a room. This bass comes from the loading of the drive with horn not form the driver itself. However if you use a driver with Fs around 50s and that in open air has –3dB at 80Hz but in horn it has lower response at 120Hz then you did not bult a horn but just a God knows what. Also, the total gain of sensitively across the entire bandpass should be around (or not less then) 6dB (depends of the curve you selected). If you meet the criteria of EQ and sensitively gain then you have a horn-loaded middbas. If you do not then you have a direct radiator with some “wings” around the driver and that looks like a horn

Anyhow, to get better middbas horns we need to use sampler throat, no-stepped exponential curves (I personally use tractrix just because it is shorter. Longer horn = more sonic problems), larger mass, exploit to a driver with a higher resonant frequency and …the horn should be properly used in context of playback installation.

The Problems with midbass horns – the drivers.

OK, a paper driver for middbas should have as low as possible excursion, as high as possible sensitively, the Fs juts slightly less then the lower mouth cut off, somewhat lower as Q but it is not really necessary. All the rest rules do not work with the middbas drivers and that is no ways to find out how it will behave until you actually place it into your horn. Be advised the if you in a curved horn then you might use any crapy drivers you wish - it will be practically irrelevant for the result, you might use even 2-4 drivers in the compression chamber. However, if you use a single driver in a straight horn then the quality of sound of this drive become dominating. 

The dreamer of the drivers also might wary. Usually it is a good idea to use a larger diameter than the throat of horn as the larger drivers more likely to have more powerful magnetic structure. A driver should handle at least 2 octaves above your desirable HF cut-off. It might sound ridicules for you but the ability of you upper bass driver to handle MF is superbly important and it will huge responsible for how your midbass channel sounds and how it will be integrateable with your MF channel. Be advised that some drivers would “strangely” react to large indictors if you use a speaker-level crossover and the driver’s coil reactance would prevent you to roll-off the HF frequencies. The very best solution is to find a driver where the HF EQ of the smaller throat along with the longer horn would roll-off the driver’s HF without utilizing any low path crossover. Also, the long decal of the HF of middbas channel is not as bad thing as it might be with the direct radiators. In most of the cases (and practically when you do not drive your MF driver very hard into upper bass – that is always is desirable) you might do not use any low-pass filter with your midbass. 

It is highly desirable do not let the middbas driver to handle the LF that are beyond the mouth rate. By introducing a high path filter that would be one octave under the horn’s cut off would be substantially beneficial, and partiality if you use the Alnico magnets on your midbass. If you use the speaker-level crossovers then we are talking in here about really huge capacitor. Do not use plastic caps (Ansar, Solen and so on) but go for a tandem of electrolytics Nichicons with active biasing.  Be advised the unloading the LF from the horn will affect sound, even it is not miserable with your RTA. In some cases the combination between the increase the Fs via the back chamber and a certain LF unload might produce very-very interesting results: there is no universal rules in there, also each of your driver will react differently. There is not other ways to get it right without listening it. Also, the performance of your middbas would change hugely depend of what and how your LF channel used. Do not forget that you do not bult middbas channel but rather you shape sound in your room. As a starting point I would suggest getting RTA and ran a low knee slope from you listening position with a high resolution. Then introduce a cap (at the speaker level of at line level… where you use it) and find values of the cap where you detect that decrease of the cap for 5-10% would be visible as a measurable response (if you see anywhere +. 25dB then stop it as now the cap because stripper then the horn roll-off). Now, add 20% to the found value of the cup and it will be the value you need. It would not always be a calculated value as the 50Hz-100Hz your driver would most likely have a near Fs impedance fluctuation.

The Problems with midbass horns – the chamber.
 
A middbas horn should have sealed back chamber in order do not violate the Second Commandments, period. An open chamber, even a small leak, injects into room all negative attributes of dipoleness and in addition it would make the horn aware about all imaginary boundaries of room and it will be in the way that would be more aggressive then opened baffle.

Make the back chamber bid and strong. Big is necessary to be able to pace inside all imaginary drivers size and strong because it will have a huge pressure in there. Brace back chamber with metal from tanks if you need but it should be overbuilt. Eventually your chamber will be very-very small and you will need to fill the chamber with some filler (not damper but filler) but it always easy to fill the chamber with bricks then to realize that if you had 1/8” more then you would be able to try a new kinky driver that you juts picked form eBay. If you know that you have the driver that you will not change then you certainly might go for a small chamber initially.

The correct size of the chamber would be when the Fs of your driver would be near the mouth rate and it would cancel the throat reactance. However it is just a theory and in the real would it might not necessary work. The throat reactance would be different with the different curves because the different values of the air in the horn’s bell. By staffing the chamber and increasing the Fs of the driver you will VERY aggressively would change the way how your horn sound. The different profiles would behave different. Also all of it would greatly depend from the way in which the LF and MF channels joint the sound of your midbass. I do not know any other ways then to listen it. A properly sounding midbass horn should reproduces piano perfectly and should not expirees any problems to listen piano with juts the upper bass horn for days or weeks. If you have any musical problems with the sound of piano with your midbass then fix the horn. Only after everything is perfect with midbass, ONLY THEN, you might introduce the complimentary MF and LF channels. DO NOT FIX WITH MF AND LF CHANNELS THE IMPERFECTIONS OF YOU MIDBASS HORN – this is imperative.

There are some people out there who use middbas compression drivers with midbass. Do not go there under any circumstances. The existing park of compression drives and some of them can go down to 150H are not good candidates for your middbas channels. Remember that your driver MUST be able to handle at least an octave below the LF cut-off. The use of compression drives for middbas do not lead to anything besides the to sonic disasters because you will be using a driver that it is already lifeless. Would you ask a prison who juts ran a Marathon to make a pubic speech? I do not think so… So, let a person to breathe and let the compression drivers to operate where it meant to be.  If you want to know how the best compression drives sound in midbass horn then breathe out ALL air out of your lungs and count out-loud to 30-40 with frequency 2 count/sec. When you reach 15-20 pay attention to what happen with your voice – this is how the compression drives sound in midbass.

The only exception to ban of using the compression drives at midbass would use the compression driver were initially designed to operate at 30-50Hz. There are some drivers like this in Japan, quite expansive and exotic. I personally have no experience with them but I have some concern that they all use titanium diaphragms. I do not feel comfortable with at middbas. With middbas your cone must break up and the different segments of the cone should radiate the different shadows of the tones. I never had seen so far any metal-based cone operate properly at middbas but certainly to make a defiant collusion it should be properly evaluated – something that I did not do. Theoretically Evan a metal decagram might be impregnated into a soft suspension and I might work...  I’m sure that that the initially designed to be midbass compression drivers would do a wonderful pressure but I do not know if they would do a good Sound. I also, do not know anyone who would use them where I would be able to get reference points.

The Problems with midbass horns – the setting up.

The better bass horn loading comes with a reasonably smaller throat diameter and consequentially with the unfortunately-longer horn. The longer horn forces to introduce a time delay for HF channels because you most likely would not place the MF and HF horns near the location of the middbas horn’s throat (because of reflections). The delays at analog domain are big problem and particularly for HF. The digital delays are way simpler but I do not have a good experience to use them – they jut do not sound good to me (look for Horns and digital crossovers.) In some cases if to position the middbas horn and the rest of the channels as close as your real-estate allow and then would have some very minor deals left not taken care… and if you have a large room then you might live with it. The perfect time arriving is more critical in a nearfield and if you are 25-30 feet away then 2 mS might be not so auditable.. Certainly to build a shorter horn with large diameter of the throat is an easy solution that helps with the delays problem but you would lose in here some “horn-ness” of your horn. Anyhow, I always was able to hear the timing misalignment with any distance I personally reject any ability to have horns time not coherent.

If your middbas horn run over 500Hz then the mid-line of the horn should be perfectly aligned to the midranges mid-line. The way in which GOTO and some others alight the timing via angling the horn is a summit of horno-idiocy. If your horn is at sub-500H than you might play with aligning. But turning the forward radiating straight midbass horn off the soundstage, sometimes quite aggressively, you might get some extremely interesting result but be careful if your horns do into a lover midrange.

Generally the simple rectangular 70-100Hz horns with the throats of 6-9”, built around a regular 4-515G-like driver are very friendly to use, they are not large and the MF channel can be ingrate easily via positioning. They sound quite OK and very not-demanding on setup. The next level up is to go do for sampler throats, better drivers and not even-sided (5-7-9 sides) or a spherical horn.

Usually it is very unlikely that you get a probably made midbass horn to sound right the way in the way that it might. It passable but it requires s lot of experience with the subject. Generally if you get a midbass horn be prepare that it would take 1-2 moths init you figure out how to use your horn. However, when you set it up and bult your playback around it then you get sound that would be at very different level that anything else out there.

In the end: collect your upper bass sensations. Whenever you are search the opportunities to experience the sound of upper bass horns and if you find one that was properly designed, bult and used in context of it’s installation then you would experience an understanding that would be eyes-opening.

Rerurn to Romy the Cat's Site