Of course, you can't control for poor design. But all the things you listed, will eventually show up in the frequency response. Poor choice of crossover frequencies, less phase overlap through crossover, standing waves in enclosure, driver breakup, everything is revealed in the frequency response.
I'm not suggesting you EQ the heck out of anything and it's going to sound good. I'm saying you can have any sound you want ("slam" versus "clarity") by knowing how to manipulate the frequency response. There is no magic here. Of course, IMO, there is no such thing as "slam" and "clarity" because I've found that a flatter, smoother response allows me to enjoy a larger array of music, from the slam in "Smells like teen spirit" to the clarity of the piano in Chopin's ballades.
It would take around 64 series/parallel wired 12" drivers (less if mounted in a corner) to equal a single horn loaded 12. I think I'd rather drive one in a horn.
Bass horns are a thing of beauty.
Replacing horns with direct radiators can be futile.
http://www.electrovoice.com/downloadfile.php?i=970039
Yup, look for the core-midrange quality with the horn and mid-bass - and then "add" from there on either end. I think that comes from my experience with full-range drivers (..something I pursued later after horn systems).
I sometimes think though that Hoffman's Iron Law could be "broken".. sort of.
NOT extending below 50 Hz at any substantial near-average level, but rather modest form-factor and efficient.My thinking is somewhat nutty here
- but it "revolves" around the idea of a DML (or more likely a grouped mode version) grouping of panels which functionally are quasi-bipoles. Bipoles - and therefore no dipole cancellation = no pressure loss. Also, bipoles without the weird combing polar behavior of traditional bipoles.A distributed or grouped mode loudspeaker isn't something I'd want to extend higher in freq.. but below 800 Hz it doesn't bother me from an objective perspective.
It's not something I've ever experimented with - so it's just a crazy musing.
Just 2 B&C 12 PE 32's (similar drivers to the EV) in a 4 foot enclosure (total-both drivers) with a vent tuning of 45 Hz can provide an anechoic value of 98 db at 50 Hz and only -3db from there at 40 Hz.. (..granted it's a 4 ohm load with an extra 3db of voltage "gain").
However, 10-13 db more is quite a bit more, but I wonder what the size and form-factor is like for that added efficiency from the horn?
In any event, the point I was trying to make was with regard to a similar *sound* - not about which design will present the most amount of gain.
I'm sorry, I cannot agree with this. Here are two specific examples that are routinely encountered by loudspeaker designers:
1) It's normal for the bass or midrange driver to behave badly above the nominal crossover frequency. In the first system I designed (in 1975), the KEF B139 woofer had a monstrous peak at 1.5 kHz, in the range of 6 to 10 dB, and rather narrowband. The B139 was used as a near-subwoofer with a low crossover of 150 Hz, so the peak was substantially attenuated, and did not show at all in the overall frequency response.
The peak was very obvious on pink-noise audition, despite being some 15 to 20 dB or more below the in-band sound from the other drivers. To remove the peak from perception required either a sharp-tuned notch filter, or a 3rd-order crossover at 150 Hz (neither was cheap to implement with passive filters). If I remember right, I chose the 3rd-order lowpass, since it wouldn't need notch-filter tuning in production. (Both approaches sounded the same on audition.)
This was a real-world example of D.E.L. Shorter's "buried resonance" that evades detection with FR graphs. There was nothing imaginary about the audibility; flip the switch, and the mid coloration goes away, flip it again, and it's back. And Shorter was right about audibility on different program material: it was most easily heard on pink-noise, followed by massed chorus and massed string sections, and only intermittently audible on rock or blues. The sparser spectrum of rock or blues would only occasionally trigger the resonance, while broadband stimulus would reveal the coloration for what it was.
I've always followed Shorter's dictum of having the lowest possible subjective coloration with pink-noise stimulus, using other material to assess dynamics, imaging, or possible holes in the response. (The ear can acclimate to dips making a return to "flat" sound like a peak, so listening sessions need to be kept short.)
2) Since I worked with matrix quadraphonics before being recruited into speaker design, I was aware of the bizarre sound of 90-degree phase shifts spread across a wall. I was surprised to discover that 90-degree, or greater, phase shifts were audible even between drivers that were closely spaced. Once you train yourself to hear "phasiness" ... you start to hear it in many commercial speakers. At the time, the Daghlquist DQ-10 was the worst offender, thanks to many drivers with rough responses and low-order crossovers.
With theoretically ideal 1st and 3rd-order crossovers, the drivers have a 90-degree phase relationship. In the real world, drivers have ripples in their response, particularly at the edge of their passband (part of the reason for a crossover in the first place). The drivers are minimum-phase (in the usable part of their passband), so the innocent-looking ripples translate into phase deviations. At some frequencies close to the crossover frequency, the inter-driver phase angle goes beyond 90 degrees, and can approach 120 degrees.
This is partially visible as a slight dip the overall FR curve, but you have to sweep the microphone to find the real nature ... a deep null is close by, if you move the mike enough.
If you simply boost the dip region using DEQX or similar, that will not change the phase angle between the drivers. This is a very common fault in high-end loudspeakers. I hear it in many rooms at the RMAF and other trade shows; rough drivers with inadequate phase and amplitude correction in the rolloff region, which results in phasey and unstable sound in the crossover region. An MTM layout will partially disguise the fault, but walk around and you'll hear it.
Again, most audible with pink-noise, followed by mass choral and massed strings, with intermittent audibility on rock and blues. (If your music selection exposes design faults, don't expect to stay in the room long. Make sure to collect your disc from the CD player before you leave.)
These are just two, but very common, examples of design faults that make it into production on highly-reviewed loudspeakers that cost as much as $100,000/pair. There are many design faults that are beyond the scope of system equalization.
I know it sounds like I'm hard on the high-end biz, but many of the speakers with glowing reviews have outright design defects that are obvious to anyone that's designed speakers for a little while. Sure, the defects are correctable, but why do it for free for a company that is 100% sure that their product is the most wonderful thing on the market? They like their speakers just the way they are. Besides, they can always correct the defect and charge for the upgrade a year or two later.
Last edited:
I have a 8" driver that runs up to 10kHz without breakup or ripple.
And above 10kHz it does not fall abruptly in SPL response as most drivers do due to they are breaking up.
I also have a AMT driver that runs down to 100Hz.
When I change the crossover frequency between those two drivers I can get perfect measurements over a huge frequency (crossover frequency) range, but when listening it is the frequency that best matches the 8" and the waveguide radiation patterns that sounds best. So I have a 100Hz plus and minus range that are optimal.
I have not seen any commercial turnkey solutions, but amongst the most dedicated Norwegian DIYers this is sort of a trend. The Altec/GPA 515 in multiples is a favourite.
http://www.diyaudio.com/forums/multi-way/195734-nelson-pass-slot-loaded-open-baffle-project-51.html#post3153175
This is a link to some pictures on the forum posted by a really adventurous gentleman who threads unknown water (backed up with decades of experience) with his all-out approach. With 10 GPA 515 alnico on each side you are approaching the sensivity of a midrange horn.
Last edited:
As I designed the concept for pengesluk´s SLOBS they do not need EQ to reach -3dB at 30Hz.
Could have written this myself. Now you are seeing why I am circumspect about horns ( there are other things) My portable FM radio produces surprising good sound.
I am not going to delve into the physics here. Perhaps you too have a theory.
By the way.I do not use headphones. Whether Stax or whatever. They sound great but of course you dont feel the music
By your own admission this is fixed with a steeper crossover. Once the peak is about 24 to 30 db down, it should not be a problem. This requirement is no different than the requirement for good phase overlap through the crossover region. You want the phase of the two drivers to overlap through the crossover until the output of both units is down 24 to 30 db. Agreed that this is not directly revealed in the FR, but is an important requirement of good design.
90 degree phase shifts are pretty nasty, especially if on-axis or near on-axis. Again, the outputs of the two units will just not add to flat if the phase is not overlapping. Off-axis, the units might go out of phase, but this usually causes dips in the power response that have been said to be fairly innocuous. With multi-driver systems, it is impossible to get perfect summing everywhere. Maybe with some topologies you can. Again, this is revealed in the smoothness of the off-axis FR.
Agreed. But again, all this is revealed in the on- and off-axis FR curves. Perhaps the only thing that is not revealed is how loud it will go without distortion. And that mystical thing we call dynamics.
Things like resonances 30db down or phase cancellation at extreme angles might be audible to some people in special conditions, but for the most part what you hear can be completely described the on- and off-axis FR.
Maybe. But right now we have headphone systems that can't be distinguished from loudspeakers in a room.
(Smyth, DTS). What's the logical next step?
Most of us horn speaker listeners have written this dynamic coherence thing into our bibles. If we want a coherent sense of scale we use horns down to the lowest frequency. A "fully" horn loaded system - It's what makes the system whole.
The same can be said for the electrostat listeners, ribbon listeners and the cones and dome crowd. Make it fullrange or except the compromise.
How many people make the mistake of using horns with direct radiators then complain it isn't right? Electrostats? Ribbons?
Use corners and wall/floor loading so the horns can be scaled down is about the only way unless you have a huge room. I'm in a concrete bunker that helps but still use a rear dual 18 sub (in a corner) with my front horn loaded 18 for below 50 hz.
Strong corners, floors, sidewalls, ceilings, all can be your friend with bass horns. Above 80-100 hz I don't like folds but have heard folded horns that sound great up to 500 Hz. They are rare but it can be done.
I haven't built a tapped horn but they do show some promise for tight spaces and bass.
It is quite true and I had to design and build my own drivers to get both low distortion, enough headroom and dynamic that appeared un-natural for some hours before I acclimatized. The un-natural part of it was that when a guitar or drums etc. "attacked" it was sounding like 10dB louder than clipping, but undistorted and clean.
My drivers have very high xmax, very little or no mechanical spring effects (pure silk surround and no spider), magnet system that are way beyond saturation of the steel - thus the ripple effect from the magnet system are much lower than normal etc..
I have to admit that I also needed Silver Mica decoupling of the power amps PSU and also Silver Mica in parallel with the capacitors in the crossover. From before there was Mundorf silver / gold in oil that alone was unable to do the job properly.
Due to that I use my AMT that have -3dB at 100Hz and can play realistic 110 - 120dB SPL from 500Hz with 6dB/oct filter and 300Hz with 12dB/oct filter.
I have tested most of the other AMTs, ribbons, planers etc. and also connected up in series / parallel up to 16 units only to discover that it was impossible to crossover lower than with one unit and that it did not help the fidelity to split the load either.
If I run the AMT without crossover at all as a broadband driver and uses a DC coupled amplifier I can play a good 80dB SPL and if a BHL cabinet are used the low frequency response are extended down to below 60Hz.
I have full dynamic and can play at the SPL level I prefer (80 - 90 average and peaks up to 100 - 110) with my 1 watt class A JFET SIT amplifiers.
Most of us horn speaker listeners have written this dynamic coherence thing into our bibles. If we want a coherent sense of scale we use horns down to the lowest frequency. A "fully" horn loaded system - It's what makes the system whole.
The same can be said for the electrostat listeners, ribbon listeners and the cones and dome crowd. Make it fullrange or except the compromise.
How many people make the mistake of using horns with direct radiators then complain it isn't right? Electrostats? Ribbons?
A BLH folded horn with my drivers have a -3dB at 15Hz and AMT in waveguide from 500Hz and up to 15kHz where a super tweeter AMT acts as a 360 degree radiator.
The same drivers in a OB sounds much better and you totally avoid the problems the bass born creates..
Like the neighbors two or three floors above or under get serious problems with glass rattling in cabinets and vibrations.
I once fired up the BLH´s with 400 watt amplifiers and played well above 120dB.
No distortions was observed, but on the floor above the windows was rattling in addition to all other items..
Sounds fab...!
Ray,
Everything you are saying is " music to my ears " ....!
Your 8 inch cone and AMT driver combo in an open baffle design sounds like the best possible free standing full range loudspeaker I can imagine....
I have built a line array with eight 8 inch bass mids ( PHL) and 6 Beyma TPL drivers...it was fab, but not commercially viable ( driver cost plus DSP crossover) and being open baffle, it needed lots of space behind ....
Sold em to an Indian guy who had Apogee ribbons before, he loved them!
But your sounds even better...!
Ray,
Everything you are saying is " music to my ears " ....!
Your 8 inch cone and AMT driver combo in an open baffle design sounds like the best possible free standing full range loudspeaker I can imagine....
I have built a line array with eight 8 inch bass mids ( PHL) and 6 Beyma TPL drivers...it was fab, but not commercially viable ( driver cost plus DSP crossover) and being open baffle, it needed lots of space behind ....
Sold em to an Indian guy who had Apogee ribbons before, he loved them!
But your sounds even better...!