December 20, 2010

What really matters in audio - part 2

Part 1 covered the aspects of a sound system that I consider most critical >

But how can all these goals be met? It turns out many of them work together quite well.

1. Controlled directivity

A compression driver mated to a waveguide is needed. Horns fail to adequately control directivity generally, and the exceptions have lesser sound quality. The waveguide should match the mid in dispersion at the crossover point. Typically this means a 15" waveguide mated to a 15" driver.

2. High efficiency

The speaker described above will have high efficiency. 97 db can easily be achieved. A conventional large hifi speaker will have 89 db at best. Even where large drivers are used, the efficiency is low due to the extension. In keeping with our goal, our speaker will have to either be very large, or accept less extension. Typically it will be better to trade off extension and use a 3 or 4 way system.

3. Low compression

Again as a bi-product of the controlled directivity and high efficiency, we will have also achieved low compression.

4. Absence of resonances

Here we need to choose good quality drivers. Breakup in the bandwidth should be avoided. Drivers generally have well controlled breakup with a smooth native response, or a very sharp peak at the top end. Both can be used effectively, but where the peak is very sharp, it should be sufficiently removed from the crossover point.

5. Distortion profile

This can be determined with measurements and generally indicates good driver design. Where these measurements don't exist, one can choose drivers that are known to be generally of high quality. Typical examples include Acoustic Elegance, B&C, PHL, 18sound, Precision Devices among others.

6. Wide bandwidth

In a 3 or 4 way system this is easy to achieve. Consider the extension that is necessary. For a music only system, consider a 3 way system. Essentially this can be a passive two way with one or a number of subwoofers. For a music only system, consider using more efficient drivers that are better described as woofers. Consider 15 or 18" pro woofers in a vented alignment.

For home cinema, efficiency must be traded to achieve deeper extension. It's certainly worth it. Start with one LFE sub that meets output and extension targets. It may be a few subs clustered together. Greater efficiency is gained in this way.

7. Minimised very early reflections

Controlled directivity helps in this area a great deal as well. Placement is also important. Toe in the speakers to avoid side wall reflections and where placement close to boundaries can't be avoided, use treatment to tame the ill effects.

8. Live vs dead acoustics

Controlled directivity speakers require less treatment, so the problem of a dead room is avoided. Consider diffusers on the ceiling and rear wall, although be careful with nearfield placement.

9. Smooth bass response & rapid decay time

Start with a room that has some damping. Plasterboard (drywall) lining is preferred, with two layers joined with a flexible adhesive such as Green Goo. The cavity should be insulated. Then add bass traps - as many as you can fit. A rectangular room has 20 corners. 4 of them are vertical, 8 are horizontal and the remaining are tri corner where both walls and ceiling or floor meet. Add broadband bass traps to as many of these as you can. This will smooth out mode issues and also reduce modal ringing.

Next, find the best placement of speakers and bass sources.

For a small room with a small seating area, one sub may be enough. For a bigger area where you want consistent bass in every seat, you will need multiple subwoofers. The first should meet output and extension targets, the others fill in dips which persist even after bass trapping.

Then after all else is done, you may apply some EQ to reduce peaks. It's important to ensure that the correction does not create problems in other parts of the room. EQ works more for a music system in just one small area and may not be useful at all in a large room with many seats. One may be improved at the expense of another.

None of this can be properly optimised without measurements.

10. Time alignment

Time alignment can be achieved by various methods. The best for DIY enthusiasts is with a digital active crossover.

11. Response shaping

In the system described so far, this is quite easy. Passively you can do it with an L pad on the tweeter. Actively you simply adjust the gain and in either case, the subwoofer will have it's own gain control. I don't recommend magic EQ boxes. If used, it's important to understand what they are doing. Sensible correction involves a room measurement for bass and a gated measurement for everything else. Even this can be done incorrectly. They may be an improvement for uninformed users, but there's a good chance that a capable DIYer can achieve a better result with a manual system that they can control and tweak over time.

12. User adjustment

The system described here will be easily adjusted with greater precision than tone controls and the result will be a satisfying balance.

Overall the synergy that is achieved here will count for far more than lesser important issues like choices of cables, DACs, preamps and sources. Of course, valve based systems do have different requirements and so that would tend to skew priorities and make matching more critical. Amplifiers also have more critical compatibility issues than most components. Some amplifiers are speakers are just not suited, in particular due to the impedance load. Class D amplifiers also have not yet reached the level where they are ideal for treble. However, once basic compatibility issues have been addressed, excellent performance can be achieved with low cost components.

What does all this mean? It isn't cheap to set up an accurate sound system. However, it's possible to put together such a system mostly with components that are low cost. Insert these low cost compatible electronics into a well set up and well rounded system that meets as described here and they will far exceed the performance of much more expensive "high end" systems that neglect these issues. You could choose all the drool worthy components - turntables costing more than a luxury car, snobby cables costing more than many complete systems, amplifiers built with unobtainium and speaker cables made from cantaffordium. Now put them in a room with no treatment and no consideration of speaker-room synergy. Now all that you have gained will be masked by big problems. The bass will have big swings of +/- 20 db and never mind the fast transient response of those little sealed woofers. It will be completely swamped by modal decay. Those woofers might stop fast, but the room won't.

Don't get me wrong, the high end system should still sound excellent. It will simply fail to yield all of it's potential. It can easily be beaten by a well rounded system with more attention paid where it counts the most. In many cases it would not be difficult to spend 20 - 50% and get a better result.

For those with high end systems

If you already have a system put together with high cost high end components, are you getting the performance you paid for? If you are like most, probably not. I have some suggestions for you. Firstly, focus on learning some new tricks. Learn how to measure the bass in your room. Use those measurements to assist in placement. Read up as much as you can. Secondly, invest seriously in room treatments. Get as many bass traps as you can. Thirdly, make your speakers and the room the most serious investments in your system. And if you have SAF issues with treatment, do whatever you have to do to get a treated room. You might have to be creative, you might have to get something custom made, you might have to resort to bribery and corruption. No excuses!

You might also consider a speaker which is probably very different to what you have now - Gedlee Summa or Abbey. For some of you high end guys, those two speakers are too cheap to be taken seriously. I consider this speaker to be state of the art.


  1. Hi Paul,
    You have the best DIY audio site ever.
    Would love to know your take on bending wave drivers like German Physiks, Manger, Walsh.

    If you had unlimited funds what kind of DIY speaker would you build? Would you use DSP in the mix

    I follow Linkwitz, Gedlee, Danny Ritchie and Derek Wilson (Overkill audio). They all have great ideas.

    Thanks for sharing your knowledge with all of us.


  2. Thanks Mike! I have no experience with bending wave drivers, they do sound intriguing but I can't offer any real insight on them. I would need to hear them first. They do have an exotic appeal, that's about all I can say about them.

    If I had unlimited funds and wanted to put together a system that was as good as money could buy, I would commission a range of experts to design custom speakers and the room for them. I'd have in mind guys like Geddes, Linkwitz, Tom Danley and Dennis Erskine. The challenge could be that I would want them to do some things a little differently than they would prefer.

    Would it be passive, active, analgoue, digital? Good question. If funds were truly unlimited there would be no need to choose up front, but different versions could be tried. It's quite likely that I would investigate all three, perhaps a hybrid, and it's also likely that I would want some of the electronics to be designed. It's likely I would want fully active and it would be multi channel with a digital source and only one digital conversion in the chain. So that would mean the electronics in such a system would be custom made as well. That could be developed with DEQX.

    Where funds are unlimited, options are available I wouldn't normally consider. It could start with perhaps a Plasma super tweeter (they are exquisite). Then a compression driver - not an off the shelf unit necessarily as I feel most of them could be improved were they more commonly developed for high end audio. Then dedicated midrange drivers in a Synergy/Unity horn arrangement loading into foam loaded oblate spheroid elliptical waveguide. The drivers would need to be designed as the Unity mids are a little on the cheap side. Perhaps some Unity mids developed by Acoustic Elegance or even Seas. Then an open baffle array of low mids - again custom drivers similar to some AE speakers units. Then an array of custom woofers, with the lower bass covered by some built in bass horns in a multi sub arrangement based on Rythmik servo technology.

    The room would be a big part of the sound and although I'm not a huge believer in focusing too much on electronics, in such a system I would audition carefully each part. If I had the chance to create such an extreme system with no limits, it would be a full time job for a number of years with money being thrown in many directions and millions spent.

    If I was that kind of wealthy, I'd probably put on a big yearly audio event open to all audio enthusiasts.

    Having said all this, I enjoy the challenge of working within modest limits and designing the speakers and room myself. I'm slowly working towards creating a system that I expect will be spectacular. When people hear it, I believe it will challenge many audio beliefs. When you put together a sytem and spare no expense, everyone expects it to be exceptional, but some of these system also are disappointing in that many favourite recordings have their flaws exposed a little too much. However, when you put together a system with all the focus on the big issues, and skimp a little on the lesser issues like source and line components and amps, some people can't believe how good the result is. Many audiophiles are so entrenched in the view that everything has to be expensive, that they refuse to entertain the idea that you can get music out of a cheap DVD player.


All comments are moderated.