Why Synergy horns?

In another thread I was asked, if I would provide more details about my speakers, so I thought why not?

I have played on active 4 way horn systems since 2016. First iteration was front loaded bass horn, midbass horn, tractrix midrange horn and tractrix tweeter horn. I worked nicely, with all the attributes associated with well implemented horns. Clarity, dynamics, realistic live sound etc.

However some problems will arise, with such horns. First of all, the center to center distance between the different horns is big, compared to the crossover frequencies. We need to be within 1/4 wave in distance at x-over for a seamless transition. For instance if you x-over from the midrange horn to the tweeter horn at 3 KHz the c-to-c distance would have to be 340/3000/4= 2.83 cm (1.11 inch). This is virtually impossible with "normal" horn configurations. This problem rears its ugly head, at every x-over throughout the audio frequency range. As frequency decreases, the wavelengths gets bigger, but so does the horns in the specific bandpass and then c-t-c also increases. It is a linear problem, that can't be solved with the regular approach, aka stacking horns on top of each other. This creates interference problems and lobing in the vertical response curves, that will color the reflection from floor and ceiling. Secondly a large column of vertically stacked horns, will push the sweet spot (SS) further back, for the horns to be perceived as more coherent and integrated, with one another.

But the biggest problem is that almost all horns beam with increasing frequency, it's their way of nature so to speak. What that means, is that the off-axis FR will not be similar to the on-axis FR. This translate into a poor power response, which is not considered a good thing, in terms of best sound quality.

Luckily we can circumvent all these problems with clever engineering and have our cake and eat it too, so to speak. Enter the Synergy horn.

 

How to EQ constant/controlled directivity horn

For the CD horn to work correctly, we need to EQ it to counteract the -6 dB drop off in the high frequencies. A normal horn loaded compression tweeter driver has a very high efficiency. 110 dB/1w at 1 meter is not unusual. That means there is usually power to spare, allowing the frequency compensation to be added. There are 2 ways of doing that. In the analog domain using passive circuits to create a low cut, a capacitor in parallel with a resistor, in the combination series with the driver. But in this modern day and age we don’t use passive components for signal shaping, we use digital EQ. The power requirements above ~2 KHz fall at around 6dB/ octave (20 dB/decade). Therefore, we can easily boost the top end above 3 KHz, without running into trouble, power wise. As long as the DSP do not distort the signal, which means that the DSP output reader does not go above 0 dB, all is fine.

Hello

YMMV Using digital EQ to boost the response is not the best approach. You are much better off using attenuation as used with a passive network. All you are doing using boost EQ is increasing the power into a driver where the distortion is the highest you are also reducing the available headroom. Compression drivers have a secondary resonance where most of the last octave response is from. The THD rises where this resonance occurs. You are also potentially adding more HF noise "hiss" to the overall system.

Rob

 

Hi Rob, as I wrote , you can either attenuate or boost. Situation will depend. Boosting has been done in both pro and high end circles for decades. What you normally do with the digital EQ, in an active system, is after the boost, you attenuate the output amplifier. Either by a volume potentiometer on the amp or you could place a resistor in series with the line input signal to the amp. In that way you bring up the signal-to-noise ratio.

I have a 200 watt amp on my CD tweeter, then I put 10K ohm resistor in series on the input to the amp. I can now boost the DSP until just below clipping the signal. In that way there is no hiss, unless you put your ear right next up to the tweeter.

Regarding power to the tweeter I wrote "The power requirements above ~2 KHz fall at around 6dB/ octave (20 dB/decade). Therefore, we can easily boost the top end above 3 KHz, without running into trouble, power wise."

 

CD horn continued...

Not all CD horns are created equal. In the 1980’s where the CD horn was developed, because it was recognized, that an even off-axis response was important to the overall sound quality. These early attempts on controlling the directivity, where based on a technique called diffraction. Altec Mantaray, JBL bi-radial (bum horn), JBL 2360 etc. all used this diffraction technique. The diffraction aperture itself is a vertical parallel-sided slot section near the throat that allows the horn to achieve a wide coverage pattern at high frequencies. It works to some extent, but there is a downside to such diffraction horns. The diffraction slot represents a severe discontinuity to the propagating wave front and creates something called high order modes (HOM), which are reflections, that happens within the horn, bouncing off the horn walls. This HOM phenomenon colors the sound, also known as “horn honk”. HOM’s in the higher frequencies has a sizzling character, similar to the sound like frying bacon on a pan.

In a perfect world, we only want one soundwave to travel down the horn as a plane wave, but the diffraction slot creates a number of different waves reflections propagating down the horn. These reflection waves arrives later in time than the original wave and these are the unwanted higher order modes.

Optimal modern CD horn uses a oblate spheroid throat, going into straight sided walls. At the mouth, the horn needs a termination roll off, to prevent waistbanding at the horns cut off frequency. The mouth termination also prevents soundwaves from reflecting off from the mouth (diffraction) and back into the horn.

All horns suffer to some extent from this HOM phenomenon. CD horns has the least of HOM’s compared to other horn types. Luckily, we can counter act these HOM’s by applying FIR filter to the incoming signal.

Below is the spectrogram of resonances in a typical high frequency horn.



With a correct FIR filter is applied it looks like this.



Impulse response without FIR filter




IR with FIR filter



With the correct FIR filter in place the improvement in the transient decay is very apparent and IR get's much better.

 

Why is CD horn important?

CD horns was developed for the pro audio industry, to control the coverage of a large space and provide a uniform sound field. Hifi consumers want controlled directivity, because it makes the power response more uniform in a small space, as in living rooms. Both user groups need the same thing, but for different reasons.

Two important things are happening when CD horns is used in home audio. First, the off axis response will be similar to the on axis response, so reflections will blend in with the direct sound with the same tonality and thus not skew up and alter the perceived sound. Secondly, what is generally accepted as the ideal in a hifi system is one that has good imaging, as in locations of instruments, as well as good spaciousness, the feeling of being in an acoustic space. The trouble is that these two criteria are counter opposed, in a small room, unless they are dealt with in a very particular way.

Image perception are dominated by the direct sound (first arrival), from the speaker and the sound that arrives in the first 5-10 milliseconds, called early reflections (ER). The ear simply integrates all this into one lumbed sound. This includes the speaker’s direct sound along the listening axis, cabinet diffractions and diffractions and reflections from nearby objects like equipment cabinets or televisions. Optimally, what we wants for good imaging is a point source response, a single direct sound, free from any diffractions or reflections for at least 5, but even better, a full 10 milliseconds.

It is difficult to get this kind of low ER in most listening rooms. If high amounts of absorption are used then the ER are decreased, but then spaciousness is lost as well. To get spaciousness you need a lot of reflections (preferably lateral) from the room and delayed more than 10 ms. In most typical home listening rooms using traditional loudspeakers you can have imaging or you can have spaciousness, but it’s difficult to have both. You have to give up on one to get the other. This compromise plays a big role in selecting a loudspeaker that suite ones taste and room.

This is where CD horns comes in. CD horns lets us have the two highly desirable features of imaging and spaciousness, without having to trade one for the other.

If the CD horn is directional, 90 degrees or less in the horizontal plane, then it can be placed and aimed such that the ER are minimized, simply because the CD horn does not illuminate the nearby room surfaces with sound. We are now free to design the room to be fairly lively or reverberant because we have controlled the ER with directivity, hence we will have good imaging. If the room is reverberant, the sound will radiate off, of the room surface but with a significant delayed reverberation sound, which is then heard as spaciousness.

Traditional loudspeakers do not behave in this way. They tend to spray room surfaces with sound, which creates many early reflections, thus degrading the image.

For CD horns the term “bigger is better” certainly apply. The lower in frequency we are able to control directivity, the better the potential sound quality will be. The circumference of the horn mouth will dictate its cut off frequency, of course providing that the driver(s) mounted on the horn, can support it. With that in mind, it makes sense placing them near walls or in corners, to decrease low frequency directivity and at the same time the cut off frequency.

 

@schlager Thank you for the very interesting topic and detailed posts.

Are you publishing a 'recipe' for diy or producing your own horn system for sale? Just curious Regardless a fascinating topology...

 

@schlager Thank you for the very interesting topic and detailed posts.

Are you publishing a 'recipe' for diy or producing your own horn system for sale? Just curious Regardless a fascinating topology...

Thanks, no this is just a run through the basics of the synergy horn and what I have done. There are different ways to do synergy horns, depending on situation and needs, by using different driver topologies, coverage angle etc. but they are all build using the same principals and function in the same manner.

I have a nice job, that I like and audio is just a hobby. Actually music is my hobby and the hifi part is just to make ends meet, though I find it funny and interesting to investigate

 

If you have come this far, you should by now have a grasp on how the Synergy horn functions. How we with smart engineering can get a true point source from multiple horn loaded drivers, radiating from the same place in space and time. Using the ¼ wavelength rule in combination with the horn load coupling, we can filtrate the bass and midrange harmonic distortion, above its specific bandpass.

You should also by now, recognize why controlled/constant directivity horns are important, in achieving the best possible sound in a home listening room, by avoiding early reflection and still allow later reflection, for an acoustic spaciousness.

The Synergy horn achieves both objectives, combining a broad banding point source with controlled directivity. Back in the days, before the DSP age, full range single driver speaker, was often referred to, as a coherent sound source with good imaging. That comes from two specific objectives. A point source without a cross over. What it could not do, was to reach low with power in the base and beaming in the high frequencies and lacking extension in the very top end of audio band. The synergy horn, don’t have these shortcomings and being horn loaded, it is in any home situation, practically with no limits SPL wise. Rest assured that your ears will give up long before the speakers do.

The Synergy horn is an invention of Tom Danley, from Danley Sound Inc. They precedes the former Unity design, with some small design changes, but they function in a similar way. Myself being situated in Europe, the chance to hear a Synergy horn, let alone buy one are relatively small. And being a DIY kind of guy, I like the idea to build my speakers to my specific needs. So I did.

Starting with a 3d printed horn stub, the transition from the horn tweeter to the horn itself, can be made very accurate. I use a 1” BMS 4550 ring radiator driver for the tweeter section, operating from 1200 Hz. The horn apex starts with a spherical oblate throat that transition into the straight-sided (conical) horn walls. The 3d print also incorporates imprinted beddings for the 2 particularly midrange drivers I use. They are closed back 4” drivers Celestion TF0410MR and are proven to have the right specifications, for working optimal in this kind of horn. For bass I use a 15” EVM 15L driver.

3d printed horn stub.



Notice imprinted bed for one of the midrange drivers. Two small tap holes is all there is needed, to horn load the driver. The trapped air between the driver and the horn is creating a natural low pass filter for the driver.

The two small holes at the end of the horn stub are for the extension plates to prolong the horn, for ones particularly needs. In theory, the horn can be expanded indefinitely.

Building process.







Attaching drivers to the horn.



Inclosing the drivers.



Finishing the horns.



The synergy horn is a 3-way design, measuring 135 cm (54”) by 90 cm (36”) and is covering 90 x 60 degrees. It has controlled directivity down to about 100 Hz. For bass below 100 Hz a have huge front loaded bass horns, in total 6 x 15” drivers in a total of 4000 liter cabinets. They are incorporated in a multiple subwoofer configuration.

Installment.



Using REW and a calibrated microphone, all measurements are done in the sweet spot and with impulse response correction in place.

Frequency response left.


FR right



Impulse response left.


ran out of max. 10 images continue to next post....

 

Impulse response right.


Step response left and right.


What can be seen is that left and right speaker measurements follows each other pretty closely in all 3 measurements, FR, IR and step for providing a very accurate uniform sound with a very precise imaging.

 

Very interesting. But aren't there drawbacks to DSP and FIR filters as well as the benefits that you have described?

 

Very interesting. But aren't there drawbacks to DSP and FIR filters as well as the benefits that you have described?

The discussion around DSP, room correction FIR filters are a whole topic by itself. In short they are merely tools and wrong use of tools, results in bad out come. As a general rule we only want to correct for minimum phase amplitude deviations, the IIR EQ part and correct for the excess phase, that comes from phase rotation created in the cross-over filters and that would be the FIR part. That is the short version.

 

They certainly measure well

What DSP do you use? XTA?

 

For x-over I use MiniDSP and for impulse response correction I use DRC Designer, equivalent to Acourate, Audiolense, Trinnov etc.

 

What you get with less precise horn speakers made with a more regular approach, will not be of similar sound quality, that a well implemented synergy horn can produce.

Here are measurements from two well-known and well regarded, similar priced horn speakers.





In the FR, we see strong resonances between 100-200 Hz, which is also seen on the impedance plot.

The step response is all over the place, partly because the horn mouths are lined up in the vertical plane. Therefore, the driver’s voice coils will be out of sync. The midrange arrive almost 2 ms later that the tweeter, that is a lot (about 65 cm) in relative distance. That will never translate into a stable imaging, between the speakers.


Next speaker is not much better.




In the FR, we can see that there is a mismatch between the midrange and tweeter horn. The tweeter is struggling to reach low enough to meet the midrange. In the time domain, again we see that the midrange is arriving later to the party.

Bear in mind that these measurements are done in the nearfield, about 1 meter from the speaker. Place the measuring microphone in the listening position and what you get is a 20 or even 30 dB alteration in response, because of the room impact on the FR. However, that doesn’t stop people from listening to those systems, because it is the only reference most audiophiles have.

Most audiophiles are facing these problems and on and on it goes. Lesser speaker designs combined with room interaction, unaccounted for, will just not reproduce high fidelity sound quality.

This is not for stepping on someone's toes and looking for at fight, its just to show what real world problems we, as audiophiles, are facing.

 

The discussion around DSP, room correction FIR filters are a whole topic by itself. In short they are merely tools and wrong use of tools, results in bad out come. As a general rule we only want to correct for minimum phase amplitude deviations, the IIR EQ part and correct for the excess phase, that comes from phase rotation created in the cross-over filters and that would be the FIR part. That is the short version.

Yes, it is a complex topic. There are on the one hand the effect of various DSP settings, but you also also have the effect of introducing additional conversion (ADC and DAC), with varying quality (even the quality of converters are not always easy to assess).

One thing is sure, is that I would love to hear a system like yours. Perhaps some day a similar system will be demoed in an audio show ?

 

Hi hopkins, regarding ADC and DAC, it is another explosive discussion. I use the MiniDSP OpenDRC-DA8 and it will only take a digital input. So my signal chain is kept in the digital domain and then converted to analog at the MiniDSP output to the amplifiers. Its as neutral as neutral gets, with a signal/noise ratio at 113 dB and distortion well below the human audible range.

If you recognize that the ADC and DAC process is altering the sound, it most likely comes from an attenuation of the high frequencies and we can correct for that very easily using EQ.

Sometime people refer to a DAC, that it sounds "musical" or "sterile" or even "boring". This is relatively uninteresting in the objectivist camp. It just tells us how that sounded to that particular individual. Objectivist rather work with the hifi system, as a holistic integrated unity. We know that with transparent electronics, we can get the sound to be light, dark, fat, rich in presence etc. just by altering the FR with EQ.

As a curiosum, users on a norwegian hifi forum, did a very unscientific test. They ran a signal, different small music pieces, into a 30 year old Yamaha AD-DA converter. When looped 10 times they heard small differences. That was until they corrected for the small frequency deviation the Yamaha introduced in the process. Once that was corrected for with EQ, they had a hard time hearing differences, even if the signal was looped 100 times. Food for thought.

 

One thing is sure, is that I would love to hear a system like yours. Perhaps some day a similar system will be demoed in an audio show ?

You are welcome if you ever come to Denmark But these Synergy horns are getting more attention within the DIY society, so maybe you can find someone near your location. Or look up Danley's new speaker for the home market The Hyperion.

 

I would like to point out some very basic pointers, I´m using when I analyze a particular hifi system. If voices and acoustical instruments sounds natural and nothing in the frequency response stands out and attract attention, then later in the listening session, I will try to determine image precision, soundstage, depth, dynamic capabilities, distortion issues, using a variety of music. If these objectives are met, to a satisfactory degree, then I would then describe that initial impression as high fidelity sound.

If I on other hand hear bass resonances, ringing or a skewed uneven FR, unstable image, imbalanced soundstage etc. then those shortcomings somehow would have to be fixed, to reach the level of high fidelity.

For me, measurements are important, not to judge the sound quality from a graph on a screen, but to see, what is happening with this particularly system in this particularly room? I would look at FR, IR, group delay (phase), decay and reflections, RT60, step response etc.

Without a common and replicable measurable standard, we end up comparing apples to oranges. After all, I prefer the hifi system, not to be an instrument that would be the job for the music.

 

Resorting to the objectivist camp does of course not mean that we don’t have preferences, toward some certain type of sound. Of course we do, we are after all humans with feelings, experiences, biases etc., just like everybody else.

20 to 30 years back, before measurement equipment, software and sophisticated DSP engines was available to the home market audiophile, we all ran our systems “by ear” and we were perfectly happy with that, not knowing all the shortcomings we were facing. You could get analog equalizers, but it sure as hell wasn’t looked upon having anything to do with high fidelity sound, within the audio community. In fact, for a long time audio equipment was resorting to the “no nonsense” philosophy, with no tone controls and what not.

Today we have all these tools at our disposal and we can in fact be our own sound engineer, with our particular system in our particular room. So why not use these tools? Maybe some find it to be too complicated, they have to get all kinds of information to make it work and then know how to apply the measurements to the audio system. Some might be feeling alienated toward this technology, some find it boring and not needed. And then some find it degrades the sound quality, to insert a DSP in the audio chain.

I often find that audio measurements are taken as an all or nothing proposition – some say “they aren’t perfect or completely reliable so why take them? I know what I hear so why not just listen and evaluate?”

One could argue that we all become “acclimated” to the sound signature of our system, loudspeakers and room. That gives us a bias and expectations, when we listen to unfamiliar systems. We kind of expect and try to find correlation in the sound, to what we have been accustomed to. What measurements and DSP can help us with is that we can objectively show, that the sound signature will be free from significant sonic aberrations. I think it’s a fair point to make that if you are going to get “acclimated” to a particular sound, then it makes sense to get acclimated to the most accurate one.

 

Sometime people refer to a DAC, that it sounds "musical" or "sterile" or even "boring". This is relatively uninteresting in the objectivist camp. It just tells us how that sounded to that particular individual. Objectivist rather work with the hifi system, as a holistic integrated unity. We know that with transparent electronics, we can get the sound to be light, dark, fat, rich in presence etc. just by altering the FR with EQ.

As a curiosum, users on a norwegian hifi forum, did a very unscientific test. They ran a signal, different small music pieces, into a 30 year old Yamaha AD-DA converter. When looped 10 times they heard small differences. That was until they corrected for the small frequency deviation the Yamaha introduced in the process. Once that was corrected for with EQ, they had a hard time hearing differences, even if the signal was looped 100 times. Food for thought.

I would like to believe that electronics are transparent. The way I see things, improvements in the accuracy of the speakers and accuracy of the "electronics" work together - you have to strive for both. We do have a set of tools to measure speaker accuracy (at least, to get some sens of it). We can also tweak speakers (design, crossovers & filters, placement) and corroborate measurements with our listening impressions. Changes to speaker designs are usually fairly obvious to hear. With electronics, it is impossible to do that (too complex, too many components, too many variables) so we have to rely on the assumption that the electronics we listen to are "good enough", but there is nothing really that tells us it actually is... This is why I personally tend to be more interested in others' experience with speakers (for example, what you explain in your thread), then with discussions about electronics (though the two interact).

 

I would like to believe that electronics are transparent. The way I see things, improvements in the accuracy of the speakers and accuracy of the "electronics" work together - you have to strive for both. We do have a set of tools to measure speaker accuracy (at least, to get some sens of it), and we can tweak speakers (design, crossovers & filters, placement) to corroborate measurements with our listening impressions. Changes to speaker designs are usually fairly obvious to hear. With electronics, it is impossible to do that (too complex, too many components, too many variables) so we have to rely on the assumption that the electronics we listen to are "good enough", but there is nothing really that tells us it actually is... This is why I personally tend to be more interested in others' experience with speakers (for example, what you explain in your thread), then with discussions about electronics (though the two interact).

I follow you to a certain extend. First, let me point out that well constructed audio electronics have at least 10 times or less distortion than even the best constructed speakers. And I do believe that noise, distortion and headroom can describe everything we hear in electronics. With that notion, I do agree that for example amps can sound a bit different on different speakers. Mainly due to different output impedance and distortion components. Higher order harmonic distortion and intermodulation distortion, leave a sonic imprint, if high enough. But in well constructed electronics, these distortion components are below the audible threshold.

In an active setup, with possibilities to change the tonality with EQ, I find it much easier to produce a lifelike sound that sounds natural to my ears and I can back it up with measurements.