Welcome back for part 2 of this article; ‘What can Hifi learn from PA?’. Part 1 focused on two sound systems I was particularly impressed with at this year’s Glastonbury festival; The first was a Funktion One system (who I’ve long been a fan of), and the other was from a pretty exciting company in the US called Danley Sound Labs.
Below is the RH stack of the Funktion One system at Glade, Glastonbury 2017:
If you haven’t read the first installment and are interested to have a look, you can find it here:
In this second segment we’re going to take a more detailed look at horns; exploring how they’re utilised in PA and asking whether the same principles can (or have) been used in Hifi. We’ll then have a look at drive unit design in part 3, and if people are still keen to find out more, we’ll look at quarter-wavelength designs such as the Danley tapped horns in part 4.
Just to quickly mention who I am; my name is Ceri Thomas and I run a small British Hifi company called Code Acoustics.
Below you can see our ‘SYSTEM-1’ and if you’d like to find out more you can visit our website here:
Just to be completely clear, I have no commercial connections with any of the companies I mention in this article (other than Code!).
One defining feature of PA systems is their high sensitivity i.e. they are highly (in relative terms) efficient at converting electrical energy into sound. The origins of this are fairly well known; In the early days of speaker development amplifier power was very low; often just a watt or two. So you needed to make that limited power go a long way. Two ways of doing this are the use of horns and highly efficient drive units i.e. those designs with a powerful magnet (motor) and light cone.
Here is a fairly large horn speaker intended for cinema use, from the 1930’s:
A second aspect of PA sound systems is that the venues they operate in are often temporary. This means to keep costs down you want to use the minimum number of speaker boxes to make the maximum amount of sound. i.e. sensitivity and max SPL (Sound Pressure Level) are the top requirements. Keeping distortion down has not traditionally been the no.1 priority.
History of the horn loudspeaker (very briefly)
The first practical speakers were actually horn loaded and started popping up around 1880. Edison filed a patent for a ‘speaking machine’ in 1978. Below is a working example next to the man himself:
If you’re interested to find out how it worked please have a look here:
It was essentially a (non powered) device that could make a recording of your speech, then play that same recording back to you.
Over time, sound recording technology improved and the LP record was developed. In combination with the gramophone, the public now had access to a device that could reproduce the performances of their favourite artists in their living room. The advent of domestic Hifi!
This was still non powered (note the handle), so max volume was pretty low. In 1909 the first amplifier (of the valve variety) was invented by Lee De Forest. By 1920 (ish) you were starting to see the first public address (PA) sound systems:
How do horn speakers work?
Horns effectively work by better coupling the drive unit diaphragm to the surrounding air i.e. for a smaller cone displacement, you can create higher air pressure than an equivalent drive unit that wasn’t horn loaded. Ball park figures would be 2% efficiency for a non horn loaded drive unit and 20% for the horn loaded version!
The diagram below shows a treble compression driver (A) coupled to an exponential horn flare (B):
Whilst horn flares are fairly compact for the treble region, they can become pretty unwieldy when dealing with bass frequencies.
The reason for this is the horn flare dimensions are related to the lowest frequency the horn system can support. Below this frequency, efficiency will drop (either fairly slowly, or very abruptly), depending on the profile of flare you are using.
The diagram below shows some of the most common horn profiles:
It’s beyond the scope of this article to discuss these profiles in detail and how their performance varies, but if you want to find out more, the link below is a good place to start:
I normally steer clear of using maths in my articles, but if there is just one equation you should learn in the field of acoustics it’s this one below:
Frequency = Speed of Sound* / Wavelength
*344 meters per second
A theoretically perfect horn flare should be the same length as the lowest frequency you want to produce and 1 wavelength mouth circumference in free space.
To use two examples:
- A 3kHz horn flare loading a tweeter would needs to be 11.4cm long and 3.65cm diameter.
- A 20Hz bass horn flare would need to be 17m long and 5.4m diameter.
So you can see that producing a theoretically exact treble horn flare is fairly easy, whilst creating a theoretically exact sub bass horn flare is pretty challenging (to put it mildly)!
Now I want to quickly come back to my comment about free space and use the diagram below to explain what I mean:
Treble and mid horn flares can largely be considered to be operating in free space, even if the speakers are in a living room. Distance in the field of acoustics is a relative term and related to the wavelength you are interested in. Please see below for a sample of frequencies vs. wavelengths:
- 10,000Hz = 3.44cm
- 1,000Hz = 34.4cm
- 100Hz = 3.44m
- 10Hz = 34.4m
So even if the horn is 1m from the floor, sidewall and rear wall (a little like the pic below), in relative terms it’s a very long way away when you’re producing a 10kHz signal, fairly far away at 1kHz & fairly close a 100Hz.
So now it makes sense that a treble and mid flare are acting in free space. However, a bass horn is a different beast as the wavelengths are so much larger. Even a PA bass stack in a field is not operating in free space, but rather 1/2 space (there is the floor). A bass horn in a living room could be considered to be operating in 1/8th or even 1/16th space (depending on the ceiling height).
What this means is the boundaries are effectively enlarging your bass horn flare and you can now get away with shortening the horn length and mouth area. Once you’re operating in 1/8 space, your surroundings are pretty much a horn and this will help greatly. In fact there is a specific type of horn called a ‘corner horn’ that uses exactly this principle:
Even in half space you can ‘get away with’ reducing the horn length to 1/2 your lowest wavelength or at a push 1/4. It’s more important to keep a fairly large mouth area.
To this end PA companies often stack multiple short bass horns to give a fairly decent mouth area. This will extend the bass frequency range, but not as effectively as also increasing the horn length.
Funktion One have a novel solution to this problem by creating a ‘bass horn’ that is perhaps not a true horn. Well it’s either a bass horn operating well out of it’s high efficiency band, or its a 4th order band pass box with a flared port. I mentioned in the first article that I think the argument is largely academic as it’s still a very efficient design (compared with sealed or port bass loading) and gives really tight punchy bass.
Below is the HD15 ‘kick’ horn designed by Rog Mogale, who runs Void Acoustics. Although not a Funktion One design, the concept is similar:
We can compare this with another one of Rog’s designs (the 1850), which is a slightly more traditional bass horn with a longer path length:
Rog discussed this in a little more detail in the following link:
But generally speaking the longer horn will be a little more efficient and go a little bit lower. In smaller stacks the shorter horn will sound a little more natural.
What does a horn speaker sound like?
Well if they are well designed, they sound incredibly dynamic. By that I mean they have huge dynamic range i.e. the ratio of smallest to largest sound pressure that can be accurately reproduced. Far greater than your standard low efficiency domestic speaker.
They can pick out the tiniest details in the music with pin-point accuracy i.e. they’re highly revealing. This can be good as well as bad. If you’ve got good electronics and the music is well recorded, it’ll sound incredible. However, if the electronics are poor and so is the sound recording, you won’t want to listen for very long! This is a trait that’s highly regarded in studio mastering.
Below is a pair of ATC SCM110ASL studio monitors, renowned for their flat frequency response and revealing nature (big fan!):
Due to their efficiency they can also play very loud with very low distortion, which conventional speakers find hard to do.
There is also just a sense of air, freedom, energy and excitement to the sound. You just want to tap along to the music or dance.
On that note, the kick drum for dance music is very important and horn loaded speakers can reproduce this frequency range in a way non horn loaded speakers just don’t seem to be able. The sense of attack produced by their excellent transient response is fantastic. It just sort of hits you in the chest and pulls you into the music.
Void Acoustics, like Funktion One use a dedicated speaker box just for the ‘kick’ frequency range; circa 70 – 200Hz:
Some people are adamant they don’t like horn speaker and I suspect that might be because the supporting electronics and sound recordings aren’t up to the job or the horns are badly designed e.g. the horns are too short and the mouth area too small, which can give an uneven frequency response, which sounds ‘honky’!
Others say they sound great, but is the sound ‘really’ Hifi, as the frequency response can be a little uneven. Or they’re so used to listening to inefficient speaker designs, they just can’t adjust to the sound.
Your typical inefficient domestic floor-standing using multiple small diameter woofers:
Or maybe they just don’t like the sound… but these differences of opinion are what make sound reproduction so interesting!
Rear loaded horns
There is a particular type of horn that I’m not so keen on:
I should note the above is yet another design from Rog Mogale and he’s also not a big advocate for the same reasons I’ll outline below. Here is a link to his ‘Speaker Plans’ page:
The Hifi equivalent is very similar:
This type of speaker is part of the ‘quarter-wave’ family, which I’ll cover in detail in part 4.
The issue with these designs is that at certain frequencies (related to the length of the rear horn), the sound coming out of the horn will be exactly 180 degrees out of phase with that coming out from the front of the driver. At these frequency you’ll get a cancellation and a big drop in SPL:
This first big null is quite often right in the crucially important kick region of 70 – 90Hz. If it is, the sound just seems to be heavy and has no impact. It just completely loses it’s excitement.
I should note that the big PA scoops are often used for reggae and dub, which have very little output in the 70 – 90Hz region, so won’t necessarily present a problem.
The Danley Sound Labs tapped horn are very well designed to minimise these nulls, but even then they are crossed over at circa 70Hz:
So do horns work in Hifi?
Well I guess you already know the answer to this… they are! Well not really in the mainstream, but there are a whole host of primarily niche companies making them. There is also a small, but highly enthusiastic following for them.
One of the best exponents is Avantgarde Acoustics. Here is their top of the range Trio’s with bass horns:
In fact it’s very reminiscent of Void Acoustics Air Motion PA speakers:
Both have a tweeter, upper mid and lower mid horn sections. Then stacked bass horns for everything below.
Avantgarde have a slightly more practical solution for the majority of living rooms:
To keep the speaker size down, they’ve ditched the lower mid and bass horns, and instead opted for high efficiency non horn loaded drivers.
There is another horn loaded speaker that has almost legendary status in Hifi and this is the Tannoy Westminster’s:
If you’re playing speaker ‘Top Trumps’ this is the one you want! It’s vital stats are:
- 530 liters
- 99dB 2.83V @ 1m
- 1395cm x 980cm x 560cm
- 18Hz to 27kHz (-6dB)
- 2″ compression loaded tweeter and 15″ mid/bass
- acoustic crossover at 200Hz, electrical at 1kHz
This design is interesting for many reasons. Firstly, in Hifi I prefer to combine a tweeter with a 3 or 4″ mid-driver. To combine one with a 15″ mid/bass driver really is PA territory. I suspect the reason it works here is the speaker is so efficient, the 15″ driver is barely moving, even at high volume. As such it is able to reproduce fairly high frequency sound without suffering from much intermodular distortion*.
*discussed further here:
I mentioned above that I don’t like rear loaded horns in general, but in the Westminster they do something quite clever:
The front loaded horn becomes rapidly less efficient below 200Hz. Then the rear horn has a chamber just behind the driver with a small slot. This slot acts as a low pass filter, limiting the amount of high frequency sound that can enter the horn. As such you end up with a situation where almost all the sound above 200Hz comes from the front side of the driver. Then almost all the sound below 200Hz comes from the rear side of the driver. This way they integrate pretty well, without too much of the interference you can get with badly implemented quarter-wavelength designs.
If you do ever have a chance to hear these speakers, which is probably easiest at a Hifi show, I would highly recommend it.
That’s about all we’ve got time for here. I hope you found this installment interesting and if you have any comments or suggestions please get in touch. The plan is to discuss drive unit design (PA vs. Hifi) in part 3, then quarter-wavelength speaker design in part 4. After all that, perhaps it’ll be a good time to finally answer the question… what can Hifi learn from PA?!
Thanks for reading.