View Full Version : Other brands vs Harbeth
14-06-2008, 09:50 AM
Am new in this forum. Having lived with my LV for 2 years & having read so much abt Harbeth, I'm getting interested in getting a pair of Harbeth - SHL5 or most probably the M40.1. My listening space is 3m x 5m, it's suitable for SHL5 but if I'm getting the M40.1, then I'll need to shift sys to my living room.
I heard the SHL5 in our local dealer's place, the sound is clean & clear but have yet to demo it at home.
BTW, I'm just curious, has anyone compared the 2 loudspeakers?
Thanks & best regards
As I've respectfully mentioned before (recently) a serious cross-comparison between a commercial loudspeaker and ours with a 50+ year design legacy behind it is, quite honestly, illogical. What could one reasonably expect to conclude? Bought-in commercial 'off the peg' drive units made to a price and to a wide tolerance can not, under any circumstances, be compared with British engineered, Harbeth designed, Harbeth made, Harbeth tested, Harbeth matched and paired drivers*. The piece to piece tolerance of bought-in woofers is about 5 -10 times wider than that of the drivers we make ourselves, and hence, no two commercial speakers sound or measure alike. And, I do not need to add, there is no equal to the low-coloration of the Harbeth RADIAL driver in general manufacture.
Please apply this golden rule to evaluating speakers before the music starts to play .... look at the drive units very carefully, and determine answers to these questions ...
1. Do the woofers have foam surrounds? If so, the surrounds will rot and the speakers will reach the end of their life in 5 years or so. The manufacturers freely admit this - its been known for 50 years. Using foam surrounds just proves that the design is targeted only at the consumer and is not capital equipment (http://www.harbeth.co.uk/usergroup/showthread.php?p=3284#post3284)as all Harbeth's are, with their 20+ year working life.
2. Look carefully (and touch if you can) the rubber surround of the cone. This is one of the most important parts of the speaker, humble though it looks. Is it thick and dense with a good springiness? If thin and feeble - forget it ... the speaker will not have enough damping.
3. Take out (if you can) the woofer and have a good look the quality of the chassis and magnet. If the chassis is thin and the magnet metal parts look ugly and hurriedly stamped out of metal - forget it. Harbeth woofer magnet parts are individually turned in a lathe. They are engineered. They look beautiful and they wont rust.
4. The bass/mid cone .... what's it made of? All materials (except Harbeth's unique RADIAL) have a sonic signature. The acoustic signature of the bass/midrange drive unit defines the sound of the speaker. The speaker overall can not be better than the inherent sonic quality of the bass/midrange unit.
5. Cone shape. Straight sided or a carefully profiled curve like a Harbeth? Clean, smooth lines or a strange uneven surface with ridges and bumps? How can two be made exactly the same?
6. Vacuum formed (99.99% sure they will be) or injection moulded like Harbeth? If the cone is vacuum formed (from a sheet of plastic film) it will have a matt finish. Only injection moulding gives a glass-smooth high-gloss finish.
etc. etc. etc.. I could go on, but I think you get the point. Bought-in, off the shelf drive units can not give you the Harbeth sound. The Harbeth sound IS the drive units. Attached pictures of our RADIAL 8" driver designed and made in Harbeth UK. The most accurate and natural sounding bass/mid driver in the world.
* I have been negotiating with a new supplier for nearly two years and have traveled several times to their factory. They consistently moan that "Harbeth QC is too high" and that "all other customers in the world can accept (their) QC level - why not Harbeth?". Studying the reality out on their production line, it's self-evident how few speaker manufacturers really dig-in to production QC tolerance issues with their suppliers. Buyers are so utterly focused on 'cost down' at the expense of excellence. The last meeting was brought to a stunned silence when I volunteered to pay more to get what I feel comfortable with. "We thought you were joking - you Harbeth think so differently to others ...". Yes, we do, and we're determined to protect our customers and our brand reputation. Engineered parts should be beautiful in design and execution - and that cost money.
I thought I'd take a quick look at a few speaker brands to see how they specify their speakers. I'm astonished - despite the fact that there is an international standard way of writing technical numbers, some of the best known brands can't even present their product data correctly.
According to the SI (http://en.wikipedia.org/wiki/SI_prefix) system this is the way technical numbers should be presented in a list of technical features ...
Hz (named after Herr Herz) not hz or HZ or hZ
dB (deci-Bell, named after Alexander Graham Bell), not db or DB or Db
kg (kilo grammes) not kG or KG or Kg
W (Watt, named after James Watt) not watt
OK, we all make occasional typing errors, but to have entire brochures riddled with these disrespectful errors from top to bottom raises serious questions in my mind about the product and management.
14-06-2008, 05:29 PM
Thanks for the reply! Will take note of your comments.
Just to reiterate this point ..... which is at the core of why speaker brands sound different.
If the characterless natural sound of Harbeth is what you enjoy, then it is the inherent sonic characteristic of the Harbeth RADIAL cone material that you're experiencing. There can be no other brand that will give you this experience because no other brand has Harbeth's unique RADIAL cone. So you can say, quite simply, Harbeth is in a class on one.
The "Golden Fact" of speaker design
If you are not a technical person then one fact alone is enough technical training to enable you to make sense of the tonal qualities of different speakers when you audition them. Ignoring styling, cosmetics, number of drive units, crossover design, shape, size, colour, brochures, reputation, reviews etc. etc and before you even listen to the first note over the speakers, the core fact is this ....
At the most basic, fundamental level, the characteristic sound of a speaker system is defined by (and can never be better than) the material from which the bass/midrange cone is fabricated.
Are we completely clear how coloration and cone material are opposite sides of the same coin? Do we see that Scotch whisky cannot be made from tap water - the defining character of whisky is the quality of the water. And for speaker cones it's the structure and bonds of the molecules in the cone. To make a better whisky you must use better water. To make a better loudspeaker cone you must use better molecules.
Think about this - if you attempt to reproduce a piano over a loudspeaker cone made from a sponge how do you think it would sound? The inherent quality of the sponge would define the sound we'd hear which would be soft and foggy; there would be a total absence of the bright, crisp clean tone we would hear from the live piano. All the micro-tones would be lost inside the sponge (as heat) never to be released as sound waves. All speaker cones fall somewhere along a line from the sponge to the metal bell. Harbeth's RADIAL is perfectly optimised for just the right transient attack and sustain. There is no better solution for clarity and detail in the entire speaker industry, hence nothing else can or does sound like a Harbeth!
16-06-2008, 01:53 PM
Hearing from you that the cone material of RADIAL contributes to the sound that we hear Harbeth speakers as (great as) they ar, could the a loudspeaker be further improved by possibly:
1) studying the enclosure to see if there could be an even better design?
2) crossover to see there could be an even better design?
Taking these 2 into consideration, it is certainly a totally new design from ground start, balancing the virtures of material science, enclosure acoustics and crossover sophistication and complexity.
It guarranties investments of much resources, but potentially, do you think such an investment could result in design(s) that could be better the current ones.
I sometime can't help by wondering if there could be RADIAL cone that could cover such that a range that a single fullrange of RADIAL driver could be possible.
The key sonic (and technical) advantage that Harbeth RADIAL cone (material) brings is in the reproduction of musical fundamentals in the presence region that is, towards the top end of the bass/midrange driver's working range, at or just below the frequency region at which the bass/mid hands over to the tweeter - say 1-3kHz*. This is where other cone materials are sonically at their very worst - they are lifeless and airless in the transition range from the midrange to the top, around the transition to tweeter. (Subconsciously speaker designer may be aware of this - they then crank up the tweeter's output to introduce a fake clarity; the issue is not the harmonics but the suppressed fundamentals being reproduced by the bass/mid driver alone. Most speakers are far too bright).
Now to answer your points ....
1. Yes, the enclosure can be tuned in many different ways, but the contribution from the cabinet is in the low/middle frequencies (say 300Hz, approx. middle C) only one tenth of the frequency where the cone's inherent characteristic comes to the fore.
2. Better? How? I don't understand the question. A properly executed crossover blends the two drive units seamlessly, balances their efficiency to give the desired overall frequency response and ensures that the amplifier load is benign. Beyond that, what else can the crossover be expected to do?
3. I've shown here before that even the Harbeth 8" RADIAL driver can be equalised to work without a tweeter! And it sounded very good too! Other cone materials have either such a weak HF output or one which is hideously colored that a tweeters is absolutely essential.
* Roughly, if you count the piano keys from the left, you could say that Harbeth's RADIAL really shines in comparison to other materials from about the 64th key upwards. (http://en.wikipedia.org/wiki/Piano_key_frequencies) In that region, other cone materials either have a dead, lifeless quality (http://www.harbeth.co.uk/usergroup/showthread.php?p=3303#post3303) (where you cannot hear the microtones of the piano, and you can not judge the size of the acoustic pool in which the recorded piano is playing) or are brittle and ringy (woven glass fibre). Put this to the test yourself. Human voice clearly and easily demonstrates this if recorded in a slightly live acoustic - listen to the tail of the words as they decay into the silence. RADIAL captures that detail down to a very low level - conventional materials dry it out - as heat actually in the cone itself.
Brass instruments (trumpet etc.) never, ever sound right on other polymer materials which rounds off and the attack and darkens the decay. A live brass instrument is a formidable beast with frightening presence - quite impossible to faithfully reproduce that energy on a conventional speaker cone, regardless of brand name.
I wish I could give you the hard and fast technical information (graphs etc.) prepared by the scientists and engineers behind the British Govt. funded cone research project which led us to RADIAL but I am prohibited from doing so. I will discuss this with my PR colleague to see if there is some way that we can present the technical comparisons whilst retaining confidentiality and protecting the finer technical aspects. Suffice to say, there are no off-the-shelf materials specifically engineered for loudspeaker cones and there never have been. Harbeth RADIAL is a blend of the best acoustic and mechanical properties of many other materials, and deciding on what materials to blend and in what ratio and in what way took 10 man-years and much Harbeth money. And tears. But RADIAL defines Harbeth.
You can read more about Harbeth's unique RADIAL cone, here (http://www.harbeth.co.uk/usergroup/showthread.php?p=3303#post3303).
16-06-2008, 04:48 PM
A detail posting in such a short time. My salute to you!
Looking at point no.3 in your post, could a RADIAL cone be created to cover such a wide range that there is no need for a crossover anymore?
To say crossover, there is a "step" to cross over so that frequency hands from one drive to the other. Will it be possible to a RADIAL cone be created to cover such a wide range that there is no need for such a "step"?
A wild idea(I'm not qualified to measure/indicate the feasability), say...to borrow a full range design that has no crossover, and replacing the original 8" drive with a RADIAL drive capable of covering the same frequency range?
Some days, I'm waiting for sample parts or glue to dry so I'm desk bound - today is one of those days!
The 'step' you mention has to be gradual as the output from the bass/mid driver gently hands over to the tweeter at and around the 'crossover frequency'. In my experience, 80% of the entire design time for a Harbeth speaker (I can't comment on how other designers allocate their design time because I don't meet them) is juggling with the energy in this crossover region so that what we hear sounds natural and seamlessly integrated. Not just at the sweet spot but over a wide arc.
The point you make about a wide-range RADIAL driver touches one fundamental, but unavoidable truth - that to divide any sonic spectrum into 2, 3 4 or more drive units is counter-intuitive when all musical sources - including our mouth - are just one point in space. Yes, the RADIAL driver can be pushed so that a tweeter is not necessary but - a big but - at a price so to speak. We know that as frequency increases, the effective working part of a cone drive unit progressively withdraws nearer the voice coil. Put simply, if you took a pair of scissors to any normal, working 8" driver's cone and you trimmed away 7 inches leaving just a stub 1" in diameter, you'd have made a tweeter of sorts. But, here's the rub: in reality, that tweeter-like stub is connected to the remaining 7" of cone and 7" versus 1" is a huge increase in weight. So the tweeter bit has to try and drag around with it the heavy 7" bit which is not at all need for generating high frequencies, but is needed for low frequencies. So, the HF output would be extremely weak and would need a substantial electrical boost (+10-20dB) to bring it up to the same efficiency level as the bass. That can only be done in an active speaker or by using the maximum HF boost on the amp's tone controls - if it has one.
And then that would leave one serious other unresolvable issue: the 1" part of the cone we do need for high frequencies is in fact in the shadow of the entire cone, set back from the cabinet down by the neck, the voice coil. That would be like holding a megaphone to the mouth; it would give the listener an extremely narrow listening sweet spot just a degree of two wide, vertically and laterally. He'd almost have to have his head clamped (shades of Clockwork Orange) to hear HF reliably.
The crossover is intimately married to the drive units acoustic and electrical characteristics and dropping one driver into a circuit designed for another sill not give good results.
I think we should release some little screen-cam videos of how I design crossovers and what the issue are?