About my little horn system

I was not a big fan of horn systems back in my audiophile days and early career years.

Most of the Hi-Fi horn speakers on the market just don't sound quite right to me. Perhaps due to how they were tuned, or maybe the listening distance is simply not far enough for some of the big multiple-horn designs in the demo rooms at shows. It was until I listened to Bjorn Kolbrek's horn system back in the time when we worked at Celestion - With an agreeable tonality, I was deeply drawn into some of the characteristics of such system, the uncompressed dynamics, the effortlessness, and the very unusual, extremely focused imaging putting things right in front of you (due to his super narrow +-20 degree directivity mid range horn). I joyfully backflipped myself into this rabbit hole and started messing around with a relatively small horn system design. To this day, Bjorn and I still visit each other every year, listening to different designs, which has been great fun.

Back in 2 years ago, I moved into a shared house with my colleagues (good old days!), and there is limited space for acoustic treatment in my room, unless I really compromise my living quality. The only thing I can do is to make use of the acoustic properties of the furniture as much as possible (e.g. a filled bookshelf for a mix of absorption and diffusion). To get a better sound quality in that environment, I decided to design a new system for myself with the following characteristics:

1. Passive system, for the sake of having the flexibility to mess around with other Hi-Fi gear.

2. A consistent and slightly narrow directivity (+- 45 degrees), with directivity control all the way down to the lower mid-range region.

3. A shallower-than-usual cabinet so I can push them quite close to the wall.

4. The bass reflex alignment should have a slow and gentle roll-off, which will start a bit early. Works better with close to boundary placement in my opinion.

5. As always, very low distortion. Ideally, keeping the mid-range THD below -50dB at 96dB/1m situation.

5. Good output capability so I can annoy the hell out of my housemates (no, it didn't happen).

The result is what you see in the photo above, which is a 2-way horn system I temporarily named Type 1425. The sensitivity of the system is 88dB/2.83V/1m,4Ohm, F6 at 40Hz and F10 at 30Hz. The Spinorama is shown in the picture below. You can’t see them in the photo, but the bass reflex ports are on the left & right sides of the cabinet. They work as a handle grip too, making it easier for me to move the speaker around when optimising their placement.

The interesting-looking driver layout is actually nothing new; you might have seen something similar in the past. But some of those classic designs are in fact 3-way where one of the woofers is a "sub" woofer, in some cases, one may be a passive radiator. For those actual 2-way designs with dual woofers, I haven't seen measurable evidence showing that they deliberately targeted a very smooth and consistent directivity like what I intended to.

I call this design a 3-stage directivity control. The wide baffle (42 cm) enforces a frontal-focused radiation starting from around 250 Hz. Then, the acoustic interference between the two horizontally arranged mid-woofers takes over and narrows down the directivity further. Finally, it hands over to a constant directivity horn, which takes care of everything from mid-range to above. As usual, I simulate the horn using Bjorn's MMM approach. The horns are SLS 3D printed, lightly sandblasted and then spray painted. The finish is lovely.

It is quite important to measure a speaker at a far enough distance when it involves a sizable waveguide/horn and multiple drivers that handle the same frequency range, so you get its far-field polar response, which describes the tonality of the system more accurately. When I first measured the speaker, it was at 1m in a mid-small size anechoic chamber with a slightly high cut-off, which I wasn’t totally sure if it's good enough. Recently I built another pair of Type 1425 for a friend, and I took the chance to scan it on the Klippel NFS. The software allows me to obtain the polar response at any given distance, which is very handy. The picture below shows the horizontal polar map of the Type 1425 from the scan (SPL normalised to 1m).

To hit my directivity target, I must crossover to the HF driver in the range before the beaming effect shrinks the beamwidth of the mid-woofer section much less than +-45 (and definitely before the crazy lobing occurs), which is around 850Hz. As you know, a simple constant directivity horn in this size with a +- 45 degree dispersion is not gonna provide much meaningful acoustic loading for the compression driver at its lower end. I also can’t really find any 1” exit driver that satisfies all my criteria, so that leaves me no option but to go with a 1.4” exit driver. The frequency response at the very top end (>10kHz) will probably be compromised, but in my opinion, the performance in the range below is way more important. I chose a Celestion driver for the project at the end, not because I worked there before, but I, as an audio enthusiast, do find their modal suppression approach sounding great. The treble is just nice and smooth. At 14k Hz there is a huge peak from the breakup of the large titanium diaphragm. Luckily, it fades away in off-axis quite significantly, so it's much less audible than what you might think from looking at the graph. I do want to try other 1.4” exit drivers in the future if there is one that behaves nicely all the way up to 20k Hz, just to make it perfect.

My Fidelia (dispersion +-70 degree) and the Type 1425 have a very similar estimated in-room response, so when you compare them side by side, they sound fairly similar to each other in terms of tonality. But the second you switch to the Type 1425, you can immediately notice a difference in the presentation of the soundstage in the room. It's obviously more focused and precise; everything just sounds noticeably closer to you. It's like moving your seat a few rows forward to the front in a live concert. However, I do have to say that under low listening volume, the compression driver does sound a touch less refined than a good dome tweeter. But the moment you really crank the volume up, the poor little tweeter will start to struggle and puke out intermodulation distortion, while the 1.4” compression driver is still singing effortlessly like a diva.

I didn't commercialise the Type 1425 like the Fidelia for a number of reasons. Apart from the manufacturing cost, people's reaction to it is also a mix of love and hate. I conducted many listening sessions comparing the Fidelia and the Type 1425 with groups of friends, and many of them just don't get used to the soundstage it presents. More people prefer the wider and more laid-back soundstage from a traditional bookshelf speaker like the Fidelia. Only a few are really drawn into the Type 1425, and one immediately decided to order a pair. However, now I do see the potential of this design style being used as super high-output professional home theatre speakers; the cabinet can be made shallower, which works well for on-wall / in-wall settings. I can see myself working on a new series based on the Type 1425 in the future... Well, quite a distant future, probably. For now, let me get back to my work on the Cadentia series. Customers are waiting!