From Hi-Fi World - October 2008 issue
BUY THE MAGAZINE (back issues subject to availability)
Part 20 - Open Baffle Room Responses
Last month I was lucky enough to have the opportunity to investigate one of the great commercial open baffle designs, namely the Wharfedale SFB/3. As I pointed out, the designer Gilbert Briggs found that positioning the SFB/3 alongside the side wall, angled towards the listener, and a metre from the room corner gave the best performance. Last month’s listening seemed to indicate that this was, indeed, the best position, so let’s analyse why.
First, we can take a look at the response in a position that most hi-fi users would automatically choose – free standing well away from walls and corners. I’ve seen countless photographs of free standing open baffles and it is easy to be misled into thinking this is the best thing to do with baffled speakers.
In room measurement of Wharfedale SFB/3 free standing spaced away from walls.
Whether looking at open baffles, electrostatic or box loudspeakers I’ve never really been happy with speakers that need to be stuck in free space in order to deliver a balanced result, although I can understand the thinking behind it. Because speakers are traditionally measured in anechoic conditions the resulting designs naturally tend to gravitate towards a speaker which works independently of the room acoustics.
However, against that is the fact that room acoustics can never be dialled out of the equation where loudspeakers and rooms are concerned. In order to achieve a performance close to that measured in an anechoic chamber you’d have to play your speakers in a large open space like a field. Indeed, in the absence of an anechoic chamber, there were many designers, Gibbs included, who ended up measuring speakers hoisted clear of the ground in a quiet piece of countryside.
One of the problems with designing speakers for free standing is how do you take account of the inevitable interaction with the room in which they are eventually placed? There’s no real answer to this. In fact most of the problems that users experience with free standing speakers are invariably concentrated in the bass.
Traditional speaker measurement of an open baffle speaker on a hoist in a field.
This is hardly surprising when you look at what really happens as wavelengths get longer and stretch down to bass regions. Below 300Hz, reflections from the floor, ceiling and nearby walls start to add to the speaker’s output, unless the room is the size of a concert hall and the speaker hoisted clear of the floor.
As few listeners are prepared to put their speakers well away from rear and side walls, reflections from these boundaries add considerably to the perceived output of the speaker below 150 – 200 Hz (the exact frequency of transition depending on distance from the walls). In addition the room is, of course, a closed box in itself and low frequency energy builds up with reflections from ALL the room boundaries. We call this ‘room gain’ and a speaker’s output is boosted by progressively greater amounts as the frequency moves below 100Hz. Up to 9dB of room gain can be measured in smaller rooms.
Drive unit arrangement of Wharfedale SFB/3.
If room gain and the effect of room boundary reflections was entirely predictable then, of course, it could be extremely useful in the design of all types of speakers. Using the increased bass output due to the room we could design either speakers with high sensitivity in the midrange and treble or small loudspeakers with no real bass output of their own.
In reality, rooms differ so greatly in their low frequency performance that trying to predict how a free standing speaker will behave in a room is like trying to choose the outright winner of a horse race. Yes you can read the ‘form’ but there are always too many variables to be able to know for certain what will and won’t work.
As a result, speakers that are designed for free standing use and that have, or attempt to have, an extended bass response often end up delivering a lumpy, ill defined and, often, over-emphasised bass performance in typical living rooms. Listeners typically complain of bass ‘boom’ or ‘weak’ bass depending on where the reinforcements and cancellations of certain bass frequencies occur.
So, is there a method whereby we can predict how a speaker will behave in a room at low frequencies with a fair degree of predictability? Well we can have a good stab at it, if we assume close coupling of the speaker to two or three room boundaries.
The floor is an obvious one. It’s always going to be there anyway so we can couple the speaker to it by placing a bass unit close to it and making sure the baffle stretches all the way to the ground i.e. a floorstander rather than a standmount.
In room measurement of SFB/3 in Briggs' recommended position.
Then, if we ask the listener always to place the speaker close to one wall, we have two boundaries to which we can couple the speaker. Each boundary theoretically gives us 3dB of low frequency gain, so it is well worth attempting this feat as, suddenly, we no longer need to use other methods of boosting bass output, for example by using a bass reflex enclosure!
But what are the pitfalls of using nearby boundary reinforcement? There must be some, otherwise all speakers would be designed this way. We may have solved the problems at low frequencies but, by positioning boundaries close to the drive unit, we’ve introduced reinforcement and cancellation at higher frequencies.
SFB/3 placed across a corner to couple strongly to all room modes.
So this technique is really only successful when used in three-way speakers. The bass unit can be put close to the floor and, possibly, wall and crossed over to midrange and treble units which are kept clear of the boundaries. This, perhaps, is why you see the enthusiast’s open baffle speakers out in the middle of the room. As the typical open baffle design is often based around a single, full-range drive unit it does not lend itself to the close boundary design approach.
Now, this is where the SFB/3 comes in. Here’s an open baffle with a floor coupled bass unit, plus a bass/mid unit mounted close by and a treble unit on a separate baffle. Specifically designed for boundary reinforcement? That’s what I was interested in finding out.
Effect of moving the SFB/3 0.3m away from the corner position.
Analysis of the free standing position shows, obviously, some bass reinforcement from the floor but, also obvious, is the bass light balance compared to the midrange output. By the way these measurements are taken using an average of three microphone positions 2m from the speaker with absorbent (a large sofa) under the microphone to absorb the first floor reflection.
Compare this to Briggs' recommended position, with the side of the speaker abutting the side wall, angled towards the listener and a metre out from the corner.
What is remarkable about this measured performance is how the bass power in the room is brought better into balance with the midrange right down to 80Hz, resulting in a -6dB point of 65Hz. Now this might not seem like a very extended bass response when we think about the lowest audible frequency of 20Hz but it is all a question of balance.
If your aim with your speaker design is to achieve as extended a bass response as possible then you wouldn’t select an open baffle with which to achieve it! Surely the reason for choosing an open baffle is to achieve that open and uncoloured sound that comes from not having a box enclosing the drive units. All we can expect from a good open baffle design is to bring the bass power into balance with the midrange so that the speaker covers the full range of musical instruments.
Let’s put that into perspective. Many hi-fi enthusiasts live quite happily with miniature stand mount speakers that struggle to emit any significant bass power below 60 – 80 Hz. And yet the owners of these speakers speak delightedly of the fantastic bass response achievable from their favourites, how come?
The answer lies in the way our ears perceive the reproduction of bass instruments. Although the fundamental frequencies developed by bass guitar and percussion do, indeed, lie in the 40 – 60Hz area, these are not the strongest output of these instruments. Like many musical instruments it is the harmonics that make up the majority of the character of sound that we, as listeners, appreciate.
Experimental position for bass analysis only by pointing the baffle side nulled output towards the corner.
On that basis, providing that a speaker can generate a strong output in a room through the 80 – 200Hz region, this will make that loudspeaker sound as though it has a tremendously powerful and apparently extended response. An 80Hz pure tone is deeper than you would think by just looking at the numbers. By comparison, a 40Hz pure tone, whilst sounding incredibly impressive when air is really moved strongly by, say, a large pipe organ, is felt as much as heard.
The reason I say ‘pure tone’ is because most speakers have a strong second harmonic output when fed with a 40Hz sinewave anyway. This makes most people think that 40Hz is a deep and powerful bass note when, actually, they are probably hearing more output at the 80Hz second harmonic. Older speaker designers, well aware of this effect, used to call this ‘frequency doubling’.
The upshot of all this is that, if you really want to hear the fundamentals of musical instruments in perfect balance with their harmonics, then you should be looking either to a large loudspeaker with a big diaphragm area or a large subwoofer. Otherwise delivering a satisfying bass performance is just a matter of delivering bass power where it really matters – in the 80 – 200Hz region with, perhaps, some extension to a -6dB point of 60Hz or thereabouts.
That is exactly what the SFB/3 does in its recommended position in the room, coupling to both side wall and floor and angled to avoid upper frequency standing wave reflections and cancellations.
Now the strongest bass output in any room is delivered by putting the speaker in a corner. Obviously this couples the speaker to all room modes extremely well, but is a little disconcerting unless the speaker has deliberately been designed to fit with this format. Due to the width of the SFB/3 we can’t fit it snugly into the corner, instead we have to place the baffle across the corner.
Let’s look at what happens to the measured performance when we do this. Yes, there is a significant increase in bass output, as expected, but it is centred around a boosted peak at 110Hz. This not only unbalances the bass to midrange performance but also, note, provides a dip in energy at 200Hz. Not surprisingly the sound of the speaker at this position is ‘boomy’ and overblown in the bass region.
I wondered whether it was possible to achieve the best of both corner reinforcement and side wall coupling by moving the speaker marginally away from the corner by 200 – 300mm. A typical result is shown in the next graph taken at 0.3m from the corner. As you can see there is still considerable boost centred on 120Hz, the advantage here being that the dip at 200Hz has now levelled out. This position still exhibits the bass boominess when listening, however.
A choice remark by Noel Keywood led me to investigate one more possibility to use the extra gain available from corner reinforcement. All we have to remember is that an open baffle is a bipole. As such it will have a null in output either side of the baffle. Could we point this null into the corner?
The answer is shown in the last graph. Here I’ve pointed the edge of the baffle into the corner. Amazing, for a corner position, we’ve lost the peak at 120Hz and, instead, the bass output looks perfectly balanced with midrange! Could this be our perfect open baffle position? Obviously, as far as the midrange is concerned, this is something of a disaster as the output of the mid unit is pointing towards the side wall, not towards the listener. But leave that out for the moment as I’m just interested in the bass for this analysis.
For the final graph I’ve overlaid this angled into the corner response with the best sounding side wall coupled response. You can see that, although the bass performance down to 80Hz looks great in the corner response there is very little reinforcement from the room below this frequency.
Comparison of experimental position (blue trace) compared to Briggs' recommended position on side wall (red trace).
In comparison the side wall position looks a little bass light, when compared to the midrange level, in the 100 – 200Hz range but has good extension down to 65Hz.
I think this technique shows some promise, however, and it’s given me an idea with which to experiment in the coming weeks. I’ll report back on how this idea works out next month.