[ Home | Staff & Contacts | DIY & Tweaks | Listening tests | HiFi Playground | Music & Books ]

October 2006 editorial

"Tell Me Lies Tell Me Sweet Little Lies"... - Fleetwood Mac...

[Italian version]
Writer: Geoff Husband - TNT France

Imagine for a moment. You've just bought a new car - let's say something middle-of-the-road (sic) like a Ford Focus. You've read the Ford ads and it's just what you need, plenty fast enough, 110 mph top speed and 0-60 in 10 seconds; economical - 45 mpg and with room for five adults.

How would you feel if when you got the thing home it only reached 70 mph, 0-60 mph in 15 seconds, did 25 mpg and only had room for 3 people?

Or you get your latest PC delivered from Dell and instead of the 19" monitor it's only 17" and the hard-drive half the size claimed.

Or you buy a new Canon digital camera, on using it you find it's a 3 megapixel not the 6 it said on the box?

All rather far fetched I think you'll agree, in these days of consumer protection the manufacturers would be in court as quick as you could say "lawyers fees" and rapidly brought to book.

So why I ask myself, do audio manufacturers routinely get away with exactly the same sort of misinformation designed specifically to fool the purchaser into buying their product?

Here I'm not talking about such drivel as "300 Watt" PMP on your teenager's ghettoblaster, or even the plainly nonsensical "Perfect Sound Forever". I'm talking about well known, major hi-fi manufacturers deliberately deceiving the public.

You want examples?* Well ask yourself why some DAC chips automatically mute when presented with a "digital black" test signal. Is there any technical reason? No it's purely so that the manufacturer can quote a huge S/N ratio... In the war of numbers this might give an advantage, but it's a cheat - a lie told to the consumer. In the grand scheme of things it doesn't make much difference to how the thing works, but it shows the sort of mentality the industry has. It's nothing new; do we really think all those quoted rumble figures for turntables were correct?

Such technical 'cheats' rely on the fact that few people, few reviewers even, have the ability to check esoteric technical specifications and so they can get away with it.

I can live with all that because most of these cheats have little bearing on the actual performance of a component. What I do find hard to forgive is the industry's addiction to lying when it comes to amplifiers and speakers. The reason is that these two items are totally dependent on each other, the technical spec of a speaker and amplifier having major effects on how they will work (or not) together.

Speaker manufacturers typically quote three specs** that are significant - frequency response, efficiency and impedance.

Let's dismiss the first of these. Frequency response is often quoted as +/- 3 Dbl "in room". This is of course bollocks. Some speaker manufacturers design well-engineered speakers that measure flat (say +/-2 Dbl) between 60 and 20,000 Hz; I've seen Cabasse speakers that measure this well. "In-room" the same speakers are all over the place, and lucky to hit +/-6 Dbl from 100-10,000 Hz. It's the mass of resonances and reflections in the room that does it, and of course every room is different and the position of the speakers in the room vital. This is the smoke-screen that the manufacturer can hide behind; which room? Where are the speakers placed etc? A good manufacturer should quote "anechoic response" and provide the graphs to prove it - anything else is smoke and mirrors. Happily this is the least important spec of all...

Efficiency. This is quoted in Dbl per Watt input - a logarithmic scale where an increase of 3 Dbl is a doubling of efficiency, 6 Dbl four-fold, 9 Dbl eight-fold, 10 Dbl a ten-fold increase and so on.

I've spoken to several people, including two speaker manufacturers (one a world leader), who have confirmed that it is "industry standard" to exaggerate efficiency by 3 Dbl. They do it "because everyone does it". In fact they have to do it for the same reason. The 'cheat' that some use to justify it, is to measure the speaker in a very small, reverberent room so that all reflections are summed with the speaker's direct output - clever eh? That means that those shiny new speakers you've just bought produce half the sound pressure levels you paid for. Or to put it another way - turn your 50-Watt amp into a 25 Watter.

Hey but who cares? In these days of cheap transistor Watts you just pump the volume up a notch...

Now to the most technical spec, and the one most likely to be misleading - the impedance.

Have a look round the back of your speakers and you'll see the impedance of your speakers written on the spec plate - usually 8 ohms. The impedance is the resistance of the speaker to the amplifiers electrical input. This is very important because the lower the impedance the more current the speaker will draw from the amp. This makes it demanding. For most speakers this is at best a very loose average, at worse a work of fiction. Many speakers that claim 8 ohm impedance drop to 2 ohms or less, and this is usually in the bass region where a hell of a lot of music energy is to be found.

These "economies with the truth" really come home when the similar antics of amplifier manufacturers are taken into account. Amps are sold on their power output and many amps simply exaggerate this, though with big outputs being relatively easy, the transistor amp manufacturers tend to be more honest here.

The worst offenders are the manufacturers of valve amps, and here the combination of spurious figures can cause real problems. We all know that valve amps sound wonderful, but in a world where Watts sell, manufacturers will insist in inflating the figures for power output. This is also complicated by the characteristics of a typical valve amp.

Lets take a nice EL34 push-pull amp pushing out a claimed 30 Watts. Well let's be generous, close our eyes to the distortion figure and accept that as the truth. But, and it's a big but, that 30 Watts is midband output, circa 1 kHz, at 20 Hz you'll be lucky to get 20 Watts. Then of course most valve amps don't like low impedance loads - so into 4 ohms the output actually drops (a big transistor amp will increase output into low impedance) to maybe 15 Watts - into 2 ohms less again. It's not credible to think that the manufacturers are ignorant of the performance of their amps - it's just that "everyone does it"...

Starting to see the problem? If the speaker manufacturer is lying through his teeth and the amp manufacturer doing the same, two components which on paper are an ideal match, and which you might well buy as a pair, simply do not work together.

A worse case scenario (these are actual figures from real components)... You've just shelled out 6 months wages on a pair of wonderful 300b SE amps from a top high-end manufacturer. These babies only claim 10 Watts, but you are sensible and so buy some lovely floor standing speakers with a decent efficiency of 95 Dbl (for one Watt). This combination should be capable of 105 Dbl which is LOUD, certainly enough for an average room.

Great match. But hang on a minute, if the speaker manufacturer is like most others, those 105 Dbl are going to be 3 Dbl shy - 102 Dbl, still loud but half what you paid for.

Now the amp - those 10 Watts are only in the midband, in fact at 20 Hz you'll be lucky to see much over half that, so we need to lose another 3 Dbl in the bass - only 99 Dbl left. Now we're starting to get to the point where headroom for bass transients gets limited, but things are much worse. Why? Because those nice high-efficiency speakers actually drop to under 2 ohms at 100hz, at that level the amp is struggling to push out over a couple of watts and has very little control of the speakers, similar things are happening at the extreme treble end of the spectrum.

And so we get the classic single-ended "faults", poor bass extension and control, and weak treble. The poor punter damns the amps and buys the big muscle amp that the speakers always needed, or maybe decides to buy that wide screen telly and sod music...

So an unhappy customer, not because he/she'd bought poor components, but because the manufacturers had lied about their spec. Those 300b's would sound magical (including the bass) with very high efficiency speakers with a flat impedance curve, the speakers wonderful with a current-pumping muscle amp.***

We're all complicit in this - manufacturers, dealers, reviewers, have lived with this truth for so long that it's just taken for granted that many specifications are fiction and we ignore it. But should we? Does it hurt anyone or is it all harmless hype? I hope I've shown that in some cases the result is a total missmatch of equipment. In a world where an increasing number of people buy things 'sight-unseen' from the internet this bad habit will cause more and more misery and disillusionment amongst hi-fi buyers who have only published specs to go on when matching equipment.

And in the end are we to blame? Too many people buy things by numbers. Let's say that tomorrow an honest manufacturer brings out a big, Lowther-based horn and quotes a true 98 Dbl efficiency. Would they sell? No, because next door will be someone selling the same sort of thing with claimed efficiency of 101 Dbl - QED... And of course what's to stop another manufacturer claiming 103 Dbl? Who's going to check? Will that claim increase or reduce sales? In fact the only limit is where such claims are obviously ridiculous, but that still leaves quite a bit of room for manoeuvre.

This is where we need manufacturers to be honest, not with figures - that genie is out of the bottle - but with decent advice on what works with their equipment. That speaker manufacturer should have warned that those speakers, despite their efficiency, were unsuited to low-powered valve amps, recommended good matches, minimum requirements etc. The manufacturers of weedy SE amps could stop claiming that "because of the quality of their output transformers" that they can drive "normal" speakers and actually tell the truth, because otherwise the long-term result is a lot of people buying gear that simply doesn't work together.

Ours can be an expensive hobby, but it's also one that can be discouraging - the shelling out of a large amount of money to get a poor result. Long term, if manufacturers make honest and realistic claims and give informed advice it can only help the industry. This is where we need the help of good dealers, honest reviewers and informed forums. But here's an idea - Maybe someone could introduce a 'truth' badge (like TNT's 'real stereo') where a manufacturer's claims are tested and verified independently? That way us poor punters could buy with at least some confidence. Here's hoping.

*I've not mentioned manufacturers (name and shame) because as far as I can discover these practices are widespread and I don't want to single out one over another.

**I don't have much in the way of measuring equipment and so am indebted to my pile of old Hi-Fi Choice magazines, their technical group tests show trends very accurately. For example in one group test of valve amps the figures were as follows - Amp 1 - Claimed output 35 Watts, actual output 25 Watts (midband). Output at 20kHz into 4 ohms - 6 Watts. i.e. less than 20% of claimed output. Amp 2 - Claimed output 35 Watt, actual output 37 Watts (midband), but only 18 Watts into an 8 ohm load at 20 Hz. Amp 3 - Claimed output 80 Watts, actual output (midband) 80 Watts, but only 10 Watts into a 4 ohm load at 20 kHz (12% of rated power). All of these were relatively meaty push-pull amps, single-ended amps would be much worse - but you get the idea.

*** It's interesting that manufacturers who make both valve amplifiers and high- efficiency speakers tent to be more honest as they want you to match the two - would that continue if they only made one link in the chain?

Copyright 2006 Geoff Husband - www.tnt-audio.com

[ Home | Staff & Contacts | DIY & Tweaks | Listening tests | HiFi Playground | Music & Books ]