Upcoming Tom Danley Hifi speakers

[Rick Sykora]

Exactly, that is the exciting question. At what point are decay problems audible.

For the Genelec 8351 with the woofers behind diffraction slots, the decay delay is about one oscillation period - more details here. This is unlikely to be audible, but I don't know of any studies that have seriously determined the audibility threshold experimentally.

On the ocean-way-hr5 studio monitor measured by Amir with diffraction slot for the woofer, some slow decaying resonances have shown up, mixed with BR port resonances - more details here. There, it took up to eight oscillation periods until the signal was attenuated by -17dB. Something like this should, IMO, be avoided in any case.


If we look at a very simple simulation of a cone driver with, attention, symmetrically arranged "diffraction hole", then the potential becomes immediately clear and also the reason for the asymmetrical arrangement within the synergy horn (simulation with infinite baffle, 0-90° frequency response in the right diagram):
View attachment 133894View attachment 133893
We see optimal omnidirectional behavior in the operating range of the bass-midrange driver (in the example a 4'' driver). In the "pressure chamber" an uneven sound pressure distribution can be seen, which would not occur with a driver in the baffle.
Around 1kHz there is an extreme sound pressure level peak (resonance?), which makes sensible use impossible - hence always the asymmetrical arrangement of the "diffraction hole".

Unfortunately I can't simulate the decay of a driver - something like that would have to be done by a professional like @René - Acculution.com with COMSOL.

Exciting yes, but only interesting if it measures well and scales down (size and price IMO).

If the main advantage is as a point source, then needs to compete well with a KEF r3!

 

[kimmosto]

This is unlikely to be audible, but I don't know of any studies that have seriously determined the audibility threshold experimentally.

It's audible. See the link below.

massive mid bass decay, but no one ever complained about its audibility

Please do not mix audibility/sensibility and complaining. Possibility that owners would hear or care about any negative audible feature is close to zero due to strong belief to manufacturer and subjectively selected objectivity. Manufacturers won't highlight that they can't or want to produce products with very low excess GD. Why would they if sales and profit are okay.

Audibility of Loudspeaker Group-Delay Characteristics

 

[q3cpma]

Please do not mix audibility/sensibility and complaining. Possibility that owners would hear or care about any negative audible feature is close to zero due to strong belief to manufacturer and subjectively selected objectivity. Manufacturers won't highlight that they can't or want to produce products with very low excess GD. Why would they if sales and profit are okay.

I guess you're not completely wrong, but I'd say that the opposite is also present in some people: when you're expecting perfection, the smallest problem can trigger an inquisitive listening state, a bit like searching for a mosquito in your bedroom.

Audibility of Loudspeaker Group-Delay Characteristics

But group delay and decay aren't exactly the same thing, no (unless it is because of minimal phase behaviour)? Especially since you can correct for GD in the mid and high frequencies easily when you have DSP available. Well, to be honest, I'm quite sure that the research is already there, just need to dig a little bit on temporal masking. Also, it would be strange for Genelec to make such a paper and completely disregard it in their actual products.

 

[ctrl]

It's audible. See the link below.

Now I realize why the Finns are considered so talkative
Do you think that a slow decay behavior of loudspeakers triggers temporal masking effects and is therefore perceptible?
.

 

[AudioJester]

Not too bad

I have listened to these many times. Properly setup they sound incredible. I would go as far as saying they sound better to me than my Lx521 and can go impossibly loud with no distortion. The one downside.... that horn is almost 1m wide!

 

[Zvu]

Whaddayamean downside ?

It needs to be as big as it needs to be, to be able to perform in a desired manner.

Only downside i see is if someone wanting these cheap out on room size while building a house.

 

[Newman]

... you can correct for GD in the mid and high frequencies easily when you have DSP available.

Can IIR do it?

 

[Ilkless]

I have listened to these many times. Properly setup they sound incredible. I would go as far as saying they sound better to me than my Lx521 and can go impossibly loud with no distortion. The one downside.... that horn is almost 1m wide!

Wasn't there some trouble with Red Spade over potentially infringing on Geddes' intellectual property? I have been following it for a long time on DIYAudio and I think it's very promising indeed

 

[jhaider]

Wasn't there some trouble with Red Spade over potentially infringing on Geddes' intellectual property? I have been following it for a long time on DIYAudio and I think it's very promising indeed

Geddes probably not - don’t see any foam in the waveguide, and I assume the tweeter is a standard part - but it looks like it has taps near the throat for midranges. So depending on when it came out it may have run afoul of Danley’s “Unity” horn patents.

 

[Tom Danley]

The synergy horn concept is really very interesting and innovative.

The only thing that irritates me is the possibly less good decay behavior of the frequency ranges on which the cone drivers work together with the "diffraction slots/holes".
Several years ago I have dealt a little with diffraction slots, for me the poor decay was the reason for the discontinuation of the project.

In the synergy horns, the "diffraction holes" are arranged asymmetrically to the cone driver, which certainly helps to attenuate the resonances and "smear" them over a wide frequency range.

I have already searched the internet for measurements of the decay behavior of the synergy horns, but have not been able to find any CSD or burst decay measurement. In the PA context, the decay behavior certainly does not play such a large role, hence the low interest.

If someone has measurements of the decay behavior of synergy horns, I would be happy about a CSD measurement
.

Hi
I think i can explain the operation and purpose of the holes and the operation but it takes a couple layers, these are a bit more involved..

The efficiency gain a horn may provide has a "high pass" function based on the rate of expansion. For example a 30Hz exponential horn cannot expand any faster than doubling it's area every 24 inches but a 300 Hz horn's expansion rate is doubling every 2.4 inches and so on according to Frequency

In a horn, there is also a point where making the mouth any bigger does not increase efficiency any further (when the mouth size is about K=1) however the horn does continue to confine the radiation angle and at some point when the mouth dimension, wavelength being produced and the horn wall angle are such that the radiation angle is confined by the local horn angle at that point in the horn. Don Keele one of the horn pioneer's came up with a thumb rule for the pattern control / frequency / horn wall angle and width relate.

The pattern loss frequency is 10^6 / horn angle and horn width (inches).
I wouldn't be surprised if this aligned with baffle step as well, flat baffle confines the radiation to 180 by 180, a really wide horn haha.

What one finds is the larger the horn mouth is, the more confined to a "beam" the high frequencies are and so front edge cabinet detail is more like getting out of the way. That frame is removable and is for those who need a grill cloth (I have never had a horn at home with a grill)

Anyway the result is with a curved horn like an exponential, the radiation pattern gets narrower and narrower as the frequency increases. This is fine if your listening area is small, but if you wanted a larger area to have the same musical spectrum, you need a different horn where even the highs are spread out equally. To the degree one can confine the angle to "the same" angle over a wide bandwidth, one has a "Constant Directivity Horn". The more constant directivity it is , the more the off axis room spectrum is like the on axis spectrum.
If one examines pyramidal horn like the SH-50 or The Hyperion, one see's that the rate of expansion changes from very rapid at the apex slowing progressively as one gets to the mouth. This is why a conical horn like that has poor acoustic low frequency loading compared to the exponential but does have better directivity behavior.
All that was a realization, triggered by a comment Don Davis made about 8 years earlier at a Synaudcon. It dawned on me the reason for poor loading was due to the expansion rate at the apex. I thought what if i tapped in a driver suitable for a midrange horn, at the point in the horn where the expansion rate was correct, could i drive the horn as a single passage from two places and then three or more bandwidths according to the horns expansion rate. That is the basis of the patent.
Take two identical subwoofers, place them close together and they couple coherently into one new source, a spin o rama or polar plot shows a circle.
The cool part is the output power of 2 close coupled is four times that one of one, not twice but if you separate them, then you only get twice.
As opposed to a greater spacing like 1/2 wl or more, coherent addition means if you reversed the drive to one of the two they cancel each other our nearly completely.
This "close coupling" isn't limited to bass frequencies but it is wavelength dependent and is the condition that is met inside the throat of the Synergy horn where multiple driver couple, they are always less than 1/4 wl apart.

The mid and low ports and drivers are at the right places and properties to be horn loaded.
The holes or ports and the air trapped under the fronts of the cones form an acoustic "Low pass" filter which is somewhat above the crossovers electrical low pass and unlike an electrical filter, attenuates harmonics the drivers produced above the pass band. There is also an additional acoustic "low pass" filter effect for side wall mounted drivers related to the location in the horn as well.
One problem with loudspeakers is they add "free sound" in the form of harmonics of the input signal, noise and delayed radiations of of the signal. The first two problems get louder faster than the sound from the input signal and so "headroom is your friend".

One problem i faced developing these full range horns, beginning with the Unity horns about 20 odd years ago was the crossover.

Normal crossovers intend to have the sources in alignment at time zero and the result is there is a phase shift from high to low that is 90 degrees per order starting with 2nd order. The Time view is the highs emerge from the crossover first, then the mids, then the lf.
The arrangement of the drivers in the horn was the opposite of that result and the target seemed like to have the time come out at zero instead.

This was possible with non-traditional crossover shapes and a great deal of effort and billions on billions of clock cycles inside the computer and many cycles of build / measure / theorize / design / return. There was no design to follow.
Attached is an ARTA measurement of an SH-50 I made in 2006, it's a 3way speaker, with a passive crossover. As you can see, it appears to be a crossover less speaker with a 1w sensitivity of about 100dB 1M if i remember right. There are no lobes and nulls at crossover etc

In the late 80's, I worked on really loud source sources for acoustic levitation and my boss found out I was making electrostatic and other speakers at home. He asked me if I could "fix" his ESL-63's. By fix I mean more like neutering an animal that repairing anything as he wanted the protection spark gaps bypassed.
Like me at home with mine, they didn't quite go loud enough. Anyway i snipped one end of each gap and sat down to listen.
They did something i hadn't heard with mine, I had one on to start and with a quiet voice it sounded like the source was actually behind the speaker's radiator. I could see why, it was a set of concentric elements with a delay line so while flat, it radiated more like a single point, like a point source radiating through a window.
They did a decent job reproducing a square wave which was unusual. All that stuck with me and although the Synergy horn does it a different way , it also radiates like a single point source through a window and can reproduce a square wave over some bandwidth. Oh, his ESL-63's did burn out several months later.
I would also say something about how these came about. I like to build speakers, i still build prototypes for some of our smaller speakers at a shop / lab at my home.
When the Virus hit, things slowed down and I thought about what would i like in my living room, something smaller than what I had and importantly had a seamless subwoofer crossover and went down low enough for me.

I had two thoughts, i went with one based on the SM-60 3 way horn module which is an excellent shape but the hf driver isn't in an ideal place in time and the horn needed a minor change to be coupling to a subwoofer
We use DSP in all of our active products at work, FIR filters can be powerful but must be based of good data.
In this case, since all the sources combine into a single source, dsp correction applies every where in front. Since the subwoofer transition had to be seam less, another reason to keep it very close and use dsp.

The design / build / measurement and programing and listening took a couple months but i was really happy. I brought them in to show Mike and he asked if he could keep them in the demo room for a day or two and i said sure. I went ahead and built a second pair with a few changes and then the wood shop built some and then changed the shape a little as shown in that picture. The feed back was positive so we made a batch and decided to make them a product and we got off to a good start it seemed like.

We had an amplifier supplier who unexpectedly had major parts shortages and internal problems and we sold all the speakers we had amplifiers for. We have another amplifier maker who has supplied some prototypes which we are evaluating now.

So far as what these sound like, i am unavoidably biased but I think Erin is going to come out some time in the next couple weeks and have a listen so he may have some impression
Hope that helps
Tom Danley

 

[kimmosto]

Do you think that a slow decay behavior of loudspeakers triggers temporal masking effects and is therefore perceptible?
.

Maybe. That could partly explain why speakers with excess GD sound too bright and thin.

Timing problem becomes a dynamic problem due to energy smoothing with real life music material containing transients such as percussion and piano. Dynamic problems are percievable, but not so clearly audible. I'm just guessing, but that could be one reason why there's no agreement about threshold of audibility. Method and test material should be capable to reveal also dynamic problems - not just tone.
Fortune is that dynamics is combination of many features, and timing doesn't have to be perfect in order to get dynamic, presence/alive and realistic reproduction. Products with poor performance in most...every factor of dynamic reproduction are in the most serious trouble assuming that listener doesn't accept lame sound.
Reviews on ASR forum have something to improve too. Nowadays measurements and scoring system could rise weak products.

 

[Newman]

...So far as what these sound like, i am unavoidably biased but I think Erin is going to come out some time in the next couple weeks and have a listen so he may have some impression.
Tom Danley

Can't wait. Here on ASR we live and breathe for the next sighted listening report.

 

[Frank Dernie]

Exactly, that is the exciting question. At what point are decay problems audible.

I am interested by this.
I heard a pair of horns which I found very impressive so I looked into them and found there was a published review (in French). with measurements which looked good except for the "waterfall" which showed a slow decay - something I have pretty well always seen with horn measurements. It could, of course be the scale.
It certainly didn't seem to be a big problem when I listened and I went and heard them again at the then UK importer and ended up buying and enjoying a pair.
I don't know why but comparing them with speakers with a very good waterfall plot doesn't reveal the latter to be hugely better.

 

[thewas]

A good horn doesn't necessarily have much slower decay than a classic loudspeaker implementation though, for example the

Avantgarde Acoustic Uno XD

(source: https://www.fidelity-online.de/avantgarde-acoustic-uno-xd-messungen/ )

compared to the B&W 800 D3

(source: https://www.fidelity-online.de/bowers-wilkins-800-d3-messungen/ )

 

[ctrl]

I thought what if i tapped in a driver suitable for a midrange horn, at the point in the horn where the expansion rate was correct, could i drive the horn as a single passage from two places and then three or more bandwidths according to the horns expansion rate. That is the basis of the patent.

Thank you for the summary of your patent. As I said before, a really original/inventive idea.

The holes or ports and the air trapped under the fronts of the cones form an acoustic "Low pass" filter which is somewhat above the crossovers electrical low pass and unlike an electrical filter, attenuates harmonics the drivers produced above the pass band. There is also an additional acoustic "low pass" filter effect for side wall mounted drivers related to the location in the horn as well.

I had not thought of that at all. With this "acoustic filter", the frequency range of a driver can be exploited much better.

At the same time, the air volume enclosed by the cone driver and the diffraction port form a Helmholtz resonator.

This can be a possible cause for delayed decay if the resonant frequency lies within the frequency range used by the driver.

Besides that, there may be other "resonance phenomena" in the undamped compression chamber, probably not so dissimilar to those of compression drivers.

In the last post you showed us the Arta frequency response and phase frequency response measurement of the SH-50 speaker.
Then it's no problem to show the gated cumulative spectral decay and burst decay of the speaker.
This would show, depending on the gate, with the CSD at least down to about 500-800Hz, whether the four 5'' midrange drivers show decay problems in the applied frequency range due to the diffraction ports (and "compression chamber").

As already said above, I cannot assess at all whether there are relevant decay problems. When using diffraction slots there are good and not so good examples for their use - as shown here.

 

[Tom Danley]

Thank you for the summary of your patent. As I said before, a really original/inventive idea.




I had not thought of that at all. With this "acoustic filter", the frequency range of a driver can be exploited much better.

At the same time, the air volume enclosed by the cone driver and the diffraction port form a Helmholtz resonator.

This can be a possible cause for delayed decay if the resonant frequency lies within the frequency range used by the driver.

Besides that, there may be other "resonance phenomena" in the undamped compression chamber, probably not so dissimilar to those of compression drivers.

In the last post you showed us the Arta frequency response and phase frequency response measurement of the SH-50 speaker.
Then it's no problem to show the gated cumulative spectral decay and burst decay of the speaker.
This would show, depending on the gate, with the CSD at least down to about 500-800Hz, whether the four 5'' midrange drivers show decay problems in the applied frequency range due to the diffraction ports (and "compression chamber").

As already said above, I cannot assess at all whether there are relevant decay problems. When using diffraction slots there are good and not so good examples for their use - as shown here.

"I had not thought of that at all. With this "acoustic filter", the frequency range of a driver can be exploited much better."
The acoustic "low pass" filter is typically somewhat above the electrical low pass crossover frequency. If one thinks of a vented box (also a Helmholtz resonator), that is a low pass filter BUT one is using it at and above the low pass or resonant frequency and it is the Q of that resonance that governs a damped or peaked or optimal roll off shape at the low corner.

In the use in a Synergy horn, it is the output through the port that is the broad band output driving the horn and like a bass port operating well below resonance, the sound passes straight through. The electrical crossover has also low passed the drive signal to the driver so the acoustic low pass filter attenuates the harmonic distortion the driver added.

It's probably not the same cabinet as in that .png picture but i found a couple other old SH-50 ARTA file and put them in a drop box folder in the link below.
This is 1M away, in my living room with the speaker about 3 feet off the ground.
See what it looks like so far as that resonance etc. I saved the impulse response and put it in REW to see the waterfall. What one can see is the effect of the directivity and decreasing directivity below about 500Hz.
Best,
Tom
https://www.dropbox.com/sh/sbmotuz91g8avfr/AACi_7VvbIlgo4BUtELA11HOa?dl=0

 

[fluid]

Around 1kHz there is an extreme sound pressure level peak (resonance?), which makes sensible use impossible - hence always the asymmetrical arrangement of the "diffraction hole".

The resonant peak results from the band pass response of having a port and air mass in front of the driver. The size and shape of the peak and resulting rolloff after the peak can be modified by the length of the port and the amount of air trapped under the cone (volume plug). It is possible to get a much less peaky response that can be quite easily equalized. The decay of that peak can be reduced by EQ in just the same way as a room mode.

The rolloff can be incorporated into the crossover or pushed outside the intended passband.

The asymmetric positioning of the driver over the port is usually done to make the physical drivers fit within the space available as the best position for the ports doesn't make it easy to physically mount the drivers in some configurations.

 

[Tom Danley]

The resonant peak results from the band pass response of having a port and air mass in front of the driver. The size and shape of the peak and resulting rolloff after the peak can be modified by the length of the port and the amount of air trapped under the cone (volume plug). It is possible to get a much less peaky response that can be quite easily equalized. The decay of that peak can be reduced by EQ in just the same way as a room mode.

The rolloff can be incorporated into the crossover or pushed outside the intended passband.

The asymmetric positioning of the driver over the port is usually done to make the physical drivers fit within the space available as the best position for the ports doesn't make it easy to physically mount the drivers in some configurations.

Remember resonances can be damped with acoustic resistance and a horn can do that. Here is the raw response for 4x 5 inch mids in the sh-50 with no crossovers (SPL not calibrated). The notch above the passband is the 1/4 wl cancalation notch when the ports are 1/4 wl from the closed end and the sournd returns 180 degrees behind. Obviously that is well above where the driver operates.
As a direct radiator, these drivers have a rising response up to about 500Hz and then are pretty flat on axis to about 10k. As a horn driver, the sloped velocity response is what one wants for a resistive load.

 

[fluid]

Remember resonances can be damped with acoustic resistance and a horn can do that.

That is a good point. What is your take on diffraction in your speakers for home use. The moulded horns look like they would perform better than the SH50 as far as diffraction goes.

 

[Newman]

@Tom Danley You mean the notch at 2 kHz? 1/4 wl is 42mm. Are you saying the holes for the mids are 42mm from the closed end of the SH50?