PDA

View Full Version : Harbeth RADIAL v. other cone materials ...



A.S.
03-01-2008, 11:15 PM
This thread concerns the basic technical reason that a Harbeth sounds so much more open and fresher, with more micro-detail than other speakers. The original post was elsewhere (http://www.harbeth.co.uk/usergroup/showthread.php?p=1734#post1734) on the User Group, but as this is such an important subject, I've decided to move the subject here.

I said this in response to a prospective customer who complained that he was was not hearing enough detail with his existing speakers, the cones of which are advertised as being vacuum formed from a general-purpose plastic of the sort used to make squeezy cosmetic bottles.

My comments were:


... I feel like they are lifeless. Perhaps it is for the amplifier... I?m looking for a speakers with the same magical midrange and more punchy and vitality as Harbeth?sThe reason you have a dead, lifeless, airless lacklustre sound is very simple and has nothing to do with your amplifier. What you are hearing is the sonic signature of vacuum formed homopolymer polypropylene as used in the bass/mid cone.

Nothing - I repeat - nothing - inside the speaker system, outside in the electronics, the room or even the music can restore what V.F.H. polypropylene robs from the micro-detail in the music. Once music has passed through V.F.H. polypropylene cones on its way to your ears as sound, something will have been removed from the music. In fact, as you may know from our Govt. funded cone research project, what actually happens is that the low-level sonic detail* is converted to heat inside the cone (due to friction in the V.F.H.polypropylene molecules) and there is no return path from friction back to sound. Once the sound has gone, it's gone. The process is exactly akin to recording the finest digital recording onto cassette tape; the detail is lost below the granular noise of the tape particles and merges into the background noise as a continuous mush. This seems to be a serious problem in V.F.H.P 8" 200mm driver cones: the missing detail can never be restored and what's left is certainly smooth - but foggy.

The reason only a Harbeth sounds like a Harbeth (that is, it has incredible detail) is because of our RADIAL cone material. Simple as that really! You could spend years chasing this speaker or that and wasting much time and money in pursuit of the resolution you are missing. Please don't waste your time and money. You can't out-Harbeth a Harbeth by using yesterdays speaker technology.

* The best examples of this fogging process at work -

a) listen to a clean recording of a soprano in a large hall on V.F.H.P. cones and observe how, in the gaps between phrases as her voice decays into the acoustic space, it is as if the walls concert hall walls are covered in cotton wool. The micro detail, just above the recording's noise floor has been rubbed out. You can not really determine how large the hall is because the reverberation tail behind the voice has been corrupted.

b) listen to brass instruments on V.F.H.P cones - it is not capable of resolving the complex harmonic 'rasp' of brass, which in real life, and on Harbeth's RADIAL cone, has the correct 'bite'. Yes, V.F.H.P. is definitely smooth but has the sonic corners rounded off; it acts like an acoustic filter.

A.S.
03-01-2008, 11:26 PM
I can further add to this - although I readily admit that I am not a polymer engineer - that it would seem that the acoustic limitations of milky coloured V.F.H.Polypropylene seems to lie in the nature of its molecular backbone. Here is a micrograph (http://en.wikipedia.org/wiki/Image:Polypropene_migrograph.png) of PP from which you can get an impression of how tightly interconnected the molecules are. This probably explains the rubbery or waxy nature of PP, which is notoriously difficult to glue things on to. You will note from the ExxonMobil website that polypropylene is listed as ideally suited to the manufacture of plastic buckets (http://www.exxonmobilchemical.com/Public_Products/Polypropylene/Polypropylene/Worldwide/Applications/PP_App_ApplicationInfo_Consumer_HouseHoldProd.asp) or car battery cases and bumpers (http://www.lyondell.com/Lyondell/Products/ByCategory/polymers/ByType/Polypropylene/) which perfectly demonstrates the waxy nature of PP.

To have a loudspeaker cone material which is inherently waxy or rubbery is not a good idea at all since rubber readily absorbs sonic energy! PP does have useful properties - it is (almost) the lowest density polymer available which makes for lighter cones and more sensitive speakers. It is relatively cheap. It can be moulded at normal temperatures with relatively simple equipment. It is colourless and will accept dye. It's molecular damping properties are interesting, but not 'raw', in bulk as used in V.F.H.P cones. The material itself is out-of-patent and available from many global suppliers, such as ExxonMobil. It has no known health issues. It's damping profile is interesting but highly frequency dependent (undesirable for an speaker cone).

Harbeth's Govt. funded research programme which led to or exclusive RADIAL cone material took our three full-time in-house research engineers over three years. The flow chart of the RADIAL project is attached - that's the decision tree you need to follow to invent a better speaker cone material! Little wonder no other company has bothered!

And. let's not forget, our founder, Dudley Harwood discovered and patented polypropylene as a speaker cone material when he was at the BBC and the copolymer polypropylene cone featured in the original Harbeth Monitor of 1977(see attached leaflet).

>

Pluto
04-01-2008, 05:21 PM
From reading the adverts of other speaker manufacturers, how accurate is my gut feeling that there might be a resurgence of paper-coned drive units?

A.S.
04-01-2008, 05:34 PM
Paper cones - that's interesting. Actually, well engineered paper cones, as opposed to those made in dubious back street brewhouses with any old cardboard found lying around and heaved into a bubbling vat, have potential. If engineered properly, they have fibre strands that lie in the 3D thickness of the cone at random angles. This then, is the exact opposite of vacuum formed polypropylene in which there is a definite orientation to the molecules, akin to a grain. When forming a conical speaker cone you do not want a grain in the material because, as all DIY woodworkers know, the cut along the grain has fundamentally different properties to across the grain. Another reason for the clean Harbeth sound is that our RADIAL cones are injection moulded with the liquid plastic injected from the centre and flowing outwards - hence the name.

To explain the enormous differences between the vacuum forming process which the global speaker industry uses to make their speaker cones, and the way we make our Harbeth RADIAL cones by injection moulding look here:

Vacuum forming method: (http://www.technologystudent.com/equip1/vacform1.htm) Cheap, kitchen table equipment of type used in school science labs.. Labour intensive, the object (cone) needs to be trimmed from the sheet by hand with scissors when cool enough to handle. Throw away unmoulded rim. Wasteful. No two cones are exactly alike. Danger: no control over plastic sheet film material - can sheet supplier control QC? Can you trust plastic supplier to use only virgin plastic or will he mix with reground/recycled (http://www.gipo-rpi.com/sheets.html) material to cut cost? This will change sonic performance, batch to batch.

Technical cost of equipment to make vacuum formed cones: only about USD 4000.

Injection moulding method (hthttp://www.technologystudent.com/rmprp07/inject2.html) (used by Harbeth). Extremely expensive moulding machine and custom made mould tool. Zero waste. Liquid plastic squired into mould tool under huge pressure. Shape is perfect. Every one identical. Shiny surface, blemish-free. Complete in-house QC of plastic material. See attached picture of power press to make Harbeth cones which are made at the rate of about one per minute, with just one minder.

Technical cost of machine + inserts to make Harbeth injection moulded cones: about USD 400,000 (set-up cost one thousand time more expensive than vacuum forming).

>

pranderos
05-01-2008, 07:22 PM
Alan, if I understand correctly, the proprietary Radial material is not used for the driver of the HL-P3ES-2. If that is correct, how have you managed the shortcomings of lesser plastics in executing the P3 driver, which provides a generous helping of the Harbeth sound?

regards,
-Paul-

A.S.
07-01-2008, 07:44 PM
I may have mentioned in a previous comment in this thread that the problems with PP become significant as cone size increases. The P3 with its 5" cone is about the limit of what you can achieve with PP - in my opinion.

pranderos
07-01-2008, 10:26 PM
Thank you Alan. Does it follow that conversely, no great benefit would accrue to the P3 with the introduction of a 5" Radial driver? Just asking; I was quite enchanted with the P3 as it is...

regards,
-paul-

A.S.
08-01-2008, 05:16 PM
It's an interesting thought ..... but the tooling cost is quite considerable. The P3 is a great speaker - but it's so expensive to make, hence the selling price seems high. So we have to be careful not to add more cost!

A.S.
16-06-2008, 04:15 PM
Polypropylene, a milky-white material was first used as a speaker cone material by Harbeth's founder, Dudley Harwood, when he was at the BBC. He was granted a patent on the use of PP as a speaker cone, and the Harbeth company was founded upon his retirement from the BBC to maufacture (and sell to the BBC) these new PP-coned monitors (HL Mk1-3). Polypropylene can be colured (usually near-black) by the addition of dye, and worldwide most medium quality speaker cones are made from PP.

It's interesting to note that unlike Harbeths' own RADIAL material, PP was not designed as a speaker cone material, nor is it listed as being suitable or even useful as a cone material. PP is a low-cost general purpose plastic, and "most commonly used for plastic mouldings at relatively low cost and high volume, include bottle tops, bottles and fittings. It is produced in sheet form and this has been widely used for the production of stationary folders, packaging and storage boxes." "Also includes housewares, rigid packaging containers, toys, disposable medical syringes, video cassette cases, appliance housing/outdoor furniture and luggage. Non-woven applications include insulation wrap, disposable diapers, automotive interiors and medical textiles".

PP is a most useful general-purpose and low cost material but it does not boast specific acoustic properties, nor would you expect it to, as it is not really an 'engineering plastic'. But it is extremely easy to mould using nothing more than a kitchen grille. And cheap. It ihas lower mass than previous generations of material (bextrene) and hence will make a louder, loudspeaker and some say that it does not need to be doped to suppress coloration. Bextrene always needed a thick layer of dope to tame its 'quacky' coloration.

Data sheets showing applications ....

Wiki (http://en.wikipedia.org/wiki/Polypropylene)
Datsheet (http://en.wikipedia.org/wiki/Polypropylene)

Attached is a diagram which shows what we believe is the main issue with PP - the cross linking of the molecules. Each one of those billions of 'tethers' where the molecular chains curl up against an adjacent chain results in a tiny sponge, and this strips the fine detail from the sound wave as it travels through the cone.

>

harbethpr
19-06-2008, 05:01 PM
Harbeth?s secret in the quest for an ultra-pure sounding loudspeaker:....

As a highly innovative British-owned loudspeaker brand, we have decided to tell the world the secrets to our unique cone material after almost two decades of secrecy. While conventional loudspeakers depend on materials originated for other purposes, RADIAL? is the first purpose-designed for speaker cones.

Harbeth Audio developed RADIAL? cone technology after a substantial grant from the British government. The ten man-year project, begun in 1990, resulted in what has proved to be a world-beating solution to the many problems inherent with conventional speaker cones which rely on paper, shampoo-bottle plastic (polypropylene), or woven material. None of these were tailored for the audio industry but have been adopted because of their cheapness. Now Alan Shaw, Harbeth?s designer and MD can spill the beans on his secret formula. It?s all about ensuring the polymers in the cone material are kept separate so that sounds are not lost as heat from friction between the molecules.

?Turning sound into heat is the worst possible scenario for a hi-fi listener?, admits Alan. ?Because you can never change the heat back into sound ? it is lost forever.?

RADIAL? is a complex blend which contains a special ingredient which prevents the plastic becoming waxy with a ?rubbery? feel ? as is inherent with soft and squidgy polypropylene.

?It?s a bit like comparing a soft jelly baby to a hard, boiled Foxes glacier mint?, explains Alan. ?One is soft and malleable, the other rigid and can?t be bent out of shape by the sound waves passing through.?

You can see the effect by squeezing any empty polypropylene container found around the home ? they are all marked with the ?double P? symbol and a triangle. The highly stiff RADIAL? cone produces a long, clean decay time ? just as in the concert hall. Conventional cone materials, on the other hand, make the walls sound cotton-wool covered ? ?an acoustic fog?, says Alan.

?Take a single piano note as an example?, says Alan. ?With a Harbeth RADIAL? cone the note will sound realistic, as if in the concert hall with the actual piano. However, when played through a speaker with conventional cones, the note dries out prematurely as the decay is shortened. With a woven cone there is another problem from unwanted ringing as the note dies away.?

Quite simply ? RADIAL? technology is what makes Harbeth speakers unique, and such a hit with broadcasters and audiophiles around the world.

Gan CK
20-06-2008, 05:27 AM
I recall playing a piano solo track to my audiophile friend & his wife some time back. His wife being a piano teacher, was immediately struck by how my SHL-5 reproduced the tonality & timbre of the grand piano with all the natural decay & harmonics fully preserved. She said she never heard piano sounded more real & natural on any other loudspakers. My friend often brings his wife along when they go listen to various setups & my SHL-5 was the only loudspeaker she heard that could move her & the most ironic thing is that it is also the least expensive loudspeaker they heard.

I brought another friend home to listen to my SHL-5 & this time the guy is an experienced acoustic guitar player. His mentioned that listening to acoustic guitar on my SHL-5 made him feel as though he was strumming his expensive Takamine guitar right in front of him. He didn't really know how to describe the wonderful experience but commented that the SHL-5 gave him the appropriate wooden feel that other speakers didn't. There are many other examples that i can quote.

Truly indeed, Harbeth loudspeakers are made like the finest musical instruments & the Radial cone is definitely a big contributor to that extremely tone & timbre correct sound. No other speakers sound more like music than Harbeth. That's why they are my final speakers in my hifi journey after having gone through many many others.

leoyeh1
03-11-2008, 06:53 AM
Dear Alan,

I see the Radial is used as PMC MB2 XBDi LF drive units, is it the same with Harbeth?

http://www.pmc-speakers.com/product.php?mode=view&pid=201&show=critics

Best regards,

Leo Yeh

A.S.
03-11-2008, 08:43 AM
Of course not. We've been using RADIAL as a descriptive name since the early 1990s. We make our own drivers and I don't think many companies have the skills to design and manufacture their own drive units - because if they could, surely they would. In-house manufacture gives you exactly the performance you want and far more control over QC. Setting aside for a moment the superior sonic quality of Harbeth RADIAL, the RADIAL drivers we make ourselves are so incredibly consistent that there is nothing that we could buy-in that has anything like as good a QC.

Might be worth you enquiring as to what that driver material and supply source actually is. We do not make Harbeth's crowning glory available to any other company.

A.S.
04-11-2009, 04:23 PM
Whilst investigating for my own curiosity what they've technically done in the remastering of The Beatle's Sgt. Pepper's Lonely Hearts Club Band using software that allows an instantaneous comparison of the old v. new recordings I've found a very useful software facility. (We'll return to the Sgt. Pepper later).

As you now, we and Harbeth users expect the midrange clarity that only our RADIAL? cone material can yield and we've talked about this many times over the years; superior cone technology is the core concept behind the Harbeth reputation for clarity. "Conventional cones mask detail" - I've said that for over twenty years. Our entire commercial success has depended upon that fact. Anyone who has made a side by side comparison between a conventional (polypropylene) cone and a Harbeth RADIAL? cone doesn't need me to tell them that you can hear details on the Harbeth that are lost in the acoustic fog of a conventional cone. It's subtle, but quite definitely a real perforance advantage and the more you are exposed to the clarity of RADIAL? the more stark the comparison with polypropylene, or other materials.

Now the opening statement: most good speakers can sound reasonably similar in the lower and middle registers; the differentiation of quality generally becomes more apparent in the presence and low-treble region, where the human ear is especially - I means extremely - sensitive. Read here. This presence region, covered by the top end of the bass/midrange unit is where I detect the defining differences between speakers, and specifically, between speaker cone materials. Polypropylene cones, patented by our founder Dudley Harwood, used in the Harbeth Mk1-3 and most commercial speakers these days, was abandoned by him for his last monitor, the Mk4 and I continued with that concept into the now current Harbeths.

So, what we're looking to demonstrate if we can here, without you having to actually listen to a Harbeth, is how a conventional speaker cone fogs the detail especially in the presence region. I've found a way to synthesise that, but first I'd like to give you some more technical background to make the demonstration understandable. If this is demonstrable over the internet as it is here, then I have achieved a life long ambition!

Continued in next posting ... please bear with me as we work through this together. I'm fleshing-out the concept in my own mind and I need to do that in an incremental way.

A.S.
04-11-2009, 06:01 PM
Concept One: sound waves decay with time and we can plot that decay.

The starting point is to appreciate that sound waves have a finite amount of energy. That energy dissipates with the passing of time, the further the sound wave travels from its source and the more surfaces that it contacts. Were that not true, we would be able to hear voices from years ago and we'd be deafened by accumulated traffic noise over the past 100 years! Sound energy decays, and decays in seconds or fractions of seconds not in minutes or hours. It's just occurred to me that the way sound waves behave in rooms, musical instruments and loudspeaker cones has a common basis. Let's start with rooms since we all experience sound in rooms.

Let's position a microphone in an ordinary room, and then fire a starting pistol or clap our hands together really smartly. We create a sonic impulse and the sound waves radiate from the source - our hands - away from us in all directions. The first sound wave (the direct sound) to hit the microphone must be the one that takes the shortest path from the source. All the other sound waves must travel to a surface and then reflect off that surface perhaps to another and another eventually reaching the microphone. This process continues until there is no stamina left in the sound waves and the room falls completely silent. The total amount of energy (calories) that we'd used making the sound equals equals the total sonic energy picked up by the microphone over time + the sound that leaked out of the room and never reached the mic + the heat that was generated as the sound waves hit the surfaces. Now, we can draw a graph of the sound energy that the microphone detects, as time passes and as we'd expect, there is a decay line of energy v. time. See attached scan Example 1 (decay_with_time.jpg).

Example 1's blue trend line shows the sound decay gradient, and in this simplistic example, it's a straight line. That may be unrealistic. For example, if the part of the room where the microphone is located has soft, absorptive surfaces but the far end is hard, reflective and poorly damped the microphone would measure not one overall room decay gradient but two or more. See Example 2 (typical_doubleslope2sc.jpeg). Now we see the initial purple sound-energy slope showing a rapid drop in energy soon after the impulse (as those absorptive surfaces near the mic soak-up sound that would otherwise travel to the mic) but then, after the sound waves eventually travel to the 'live' end of the room and then back to the microphone the orange energy curve doesn't decay so fast, because all those little reflection from the hard acoustic there would 'top-up' the microphone.

Imagine we were blindfolded and now standing in that room where the microphone was. We fire the pistol - what do we hear? We hear what the microphone heard; the initially 'dry' sound and then a fraction of a second later, the 'wet' sound of the reflections (mini echoes) from the live end of the room. This duality of sonic character would sound very odd indeed, unpleasant, unnatural. Suppose we relocated to the untreated, live end of the room and then created the impulse. What then? Now we'd hear a blast of slow decaying reflections, and then virtual silence from the dry end. Those sound waves that did travel down to the dry, damped end of the room would be significantly absorbed and so enweakened that even if they did make it to the microphone, they'd be almost insignificant.

So, we now have the concept that even blindfolded and without any test equipment, we can grade any room's acoustic decay characteristics - it's sonic signature - by plotting how sound diminishes with time. What we definitely do not want are certain notes that hang-on long after the musician has played them so that they wrap-around, lingering in the air, still present when the next note is played. But conversely, as mentioned here (http://www.harbeth.co.uk/usergroup/showthread.php?p=6298), if we take this damping concept too far the gradient line drops vertically like a stone and all reflections are stifled. Then we'd be listening in an anechoic chamber, which we know sounds extremely unnatural indeed. There is an optimal gradient as we'll see.

>

A.S.
04-11-2009, 07:10 PM
Concept Two: The sonic decay curves of musical instruments define the tone of those instruments.

We've know from our personal experience of real rooms that there is something about the way that sound decays in those rooms that makes them enjoyable listening spaces or not. At one extreme we have the carefully considered, even decay gradient line of a world-class concert hall; at the other we know how buskers sound strumming in an underground station*.

Not only do rooms have their own characteristic decay curve, but musical instruments do too. If they didn't we'd be listening not to a rich mixture of harmonics with attractive sustain and decay but to frequency generators (synthesisers) producing laboratory tones. So, the timbre and tone of real musical instruments are defined by their attack and decay (blue-line curves) and those characteristic curves vary not only from instrument to instrument - say from a church organ to a violin - but for each note of the instrument. A well crafted violin, example a Stradivarius, would have a very unique decay profile, note by note, and the very smallest change in the materials used, the construction process and drying time right down to the varnish and glues will change the gradient of our blue lines, one for each and every note and to make matters more complicated at every loudness!

Example 3 (piano_decay.jpg) shows the the decay gradient for a few notes across one type of piano. We see in the lower notes that there is one well defined gradient, and that as frequency increases the gradient takes on the twin-slope we observed in the room decay curve. This must imply that inside the piano body there are two hopefully, harmonious structural resonance systems, one associated with the low frequencies, and the long decay, right hand side second, shallow gradient (probably from the the wooden body itself) and another, possibly the sound board, strings and iron frame responsible for the initial fast decayinging gradient. The master instrument maker balances these characteristics at the design stage then works very hard to hold all these variable in production to offer a consistent sound to his pianists. A Steinway (http://www.steinway.com/factory/) has a different construction method, different materials to a Bechstein (http://www.bechstein-centren.de/america/__2/main/pianos/c_bechstein/grand_pianos/d_280). The sound board in Peter Katin's Steinway gave me the idea for the damping arrangement in the C7.

*Incidentally, I went to a lecture a few years ago held by consulting engineers in the aftermath of the terrible fire at King's Cross underground station in London. Many lessons were learned and one concerned the problem of public address speakers in the highly reflective acoustic environment of a ceramic tiled tube station.Due to negligible absorption on the platforms, the more anxious and urgent the PA evacuation messages became, the less intelligible they were. The most absorptive objects in that acoustic space were the travellers themselves and as some of them evacuated, the extremely long reverberation time increased the cacophony for the remaining victims until the warning messages were completely jumbled - one echo overlaying another. The result: confusion and inactivity which cost lives. Down there, our blue decay-gradient line would have been very shallow and extended many seconds to the right.

>

A.S.
05-11-2009, 11:24 AM
Concept Three: Musical instruments have their own unique sonic decay profiles - what about the ideal room?

In my last post, we picked a few notes on the piano and plotted their decay characteristics. This showed that especially in the upper register, the piano has an extremely complex decay behaviour, which gives a concert grand that delicious clean brightness.

We also plotted the room on an averaged basis - that is, we ignored individual frequencies and just plotted the room's overall decay characteristics right across the audio band. Again, this is a gross simplification. Real rooms, including acoustically hard, poorly damped modern living rooms have some frequency bands where the damping is acceptable and other bands where there is a definite twang due to reflections from inadequate damping. An acoustic specialist would attempt to reduce the reflective energy in just those problematic frequency bands without killing the overall 'air', which would result from covering every surface with absorber. The rule is 'just enough treatment and no more'.

In the recording studio or concert hall, the optimal amount of damping, and hence decay gradient is very carefully considered - too much and the performers and audience complain about the dry, lifeless acoustic; too little and they have trouble following the music because of the echo. For example, the Royal Albert Hall 'suffered from acoustic problems solved in 1969 when a series of large fibreglass diffusing discs (http://www.flickr.com/photos/pikerslanefarm/2710789933/), looking like mushrooms or flying saucers, were installed in the roof to cut down the notorious echo. It used to be said that the hall was the only place where a British composer could be sure of hearing his work twice'.

In Example 4 (decay_frequency_bands2sc.jpg) we can see not a global decay curve for a recording studio, but the decay gradient for bands of frequencies, centred at the stated frequency. You can see that the decay gradient of sound decay v. time is almost the same across the audio band, and this room would be considered acoustically neutral, contributing little to the tonality of the instruments being played therein.

OK, we can now put aside the business of sonic decay in rooms and instruments. Will return to this later.

>

A.S.
05-11-2009, 12:05 PM
Concept Four: Now we should take a look at the harmonic output from an instrument We'll see that the quantity and loudness of harmonics is a critical component of the sound of a musical instruments.

The same note played on an organ and a violin will both be in-tune, but the character of the sound will be completely different and could not be confused. They are both playing the same fundamental frequency but the contribution of harmonics (or overtones) will be very different. Considering the violin, see Example 5 (violin_harmonics2sc.jpg) we can see that when the violin is playing note E5, a frequency of 659.26Hz that some (marked) harmonics are slightly louder than the fundamental note itself (!) and that all harmonics excepting the very extreme one are substantially as loud as the fundamental note.

Now the next critical mental jump: we've seen how those harmonics are far removed in frequency from the fundamental note (659Hz), unquestionably in the midrange where the body and power lies in music. We've seen how loud those harmonics are. We can see from Example5 (keyboard_harmonicssc2.jpg) that the two loud harmonics of the violin fall at and beyond the extremes of the piano's fundamental range; one reason that a violin doesn't sound like a piano. The 3rd harmonic of the note falls at 1997Hz, the 5th harmonic at 3295Hz and 6th, 7th and even the 8th harmonic at 5272Hz, itself louder than the fundamental note, will have to be reproduced solely (or for the 8th harmonic in part) by the very top octave of the bass/midrange driver below, through and just above the crossover frequency. But mostly below crossover, so not substantially fed to the tweeter.

Key point: the tweeter is not involved in the reproduction of the critical lower harmonics that define the unique characteristics of the violin. The reproduction of these harmonics is in fact primarily the responsibility of the very top end of the bass/midrange driver. And that's the core of the coloration problem ...

>

A.S.
05-11-2009, 12:52 PM
Concept Five: Speaker maths concentrates (almost entirely) on the bass end ...

We've seen that the primary focus of the bass/midrange unit is to reproduce the lower musical fundamentals and then their lower-order harmonics. A tweeetr is not needed to replay these, but of course, is desireous to reproduce the upper harmonics.

The bass/mid driver is fundamentally specified to reproduce the lower and middle frequencies without which, music has no warmth and weight. Until very recently all freely available academic research and software modelling tools available to loudspeaker designers concentrated solely on achieving a target bass response for a given drive unit in a given enclosure*. mathematically predicting and influencing the behaviour of the speaker drives in the middle and upper frequencies was/is at the very edge of acoustic research and with poor certainty. This means that the speaker insudtry has available the software maths tools to model and predict the speaker's low-end really very accurately, but in the middle and upper frequencies, the models are useless. These critical frequencies are in the hands of the speaker designer, drawing upon his experience of cone shape, material, thickness, allowable costs and complexity, glues and the rest. And a criteria deemed of essential importance to one designer may be of little consequence to another. In HArbeth's case, low coloration and excellent clarity in the harmonics are critical to us - but to another, using off-the-shelf conventional materials, just achieving a s smooth response may fulfil his design brief.

This is why there are still substantial differences in the quality of speakers in the 'presence region' of the low-order harmonics.

*I am not aware of any formal research into cone coloration since our Govt. funded RADIAL? project of the early 1990s. This means that the bought-in drive units available to other manufacturers are using obsolete technology. In fact, if those drive units use the polypropylene cone material (most bass/mid drivers are made from polypropylene) that our founder Dudley Harwood patented, then those drive units and hence those speaker systems are indeed trapped by the technology of the 1970s.

(to be continued)

A.S.
14-11-2009, 10:41 PM
First some background about the piano and its extremely complex sound, rich in harmonics:

The cone and the piano are complex mechanical structures whereby every small detail right down the the glues used will and do effect the tonality of the sound. The same with loudspeakers, specifically loudspeaker cones which are analogous to the pianos sound board, without which the piano would have no loudness.

Here is a video overview of an operational piano (http://revver.com/video/492509/music-instruments-the-piano-inside-a-piano/) from the outside. Now we can disassemble the piano and look at the all-critical sound board (http://videos.howstuffworks.com/science-channel/37435-deconstructed-piano-soundboard-video.htm) (I winced when the circular saw started up). And here is what the masterful Steinway piano company have to say about their sound boards 'the sound of the piano' here (http://videos.howstuffworks.com/discovery/30268-some-assembly-required-piano-soundboard-video.htm). And why thin woods (of the type we at Harbeth have so much confidence in) lend themselves to careful tuning, in the piano case here (http://videos.howstuffworks.com/discovery/30266-some-assembly-required-making-the-piano-rim-video.htm).

And fitting the strings, under immense tension here (http://videos.howstuffworks.com/discovery/30270-some-assembly-required-pianos-and-strings-video.htm). Note the mention of Steinway's patented overstringing which produces a rich harmonic tone. And finally, the genius construction of the hammer mechanics here (http://videos.howstuffworks.com/discovery/30269-some-assembly-required-piano-key-construction-video.htm). An animation of the motion of the sound board when the piano is being played is here (http://www.youtube.com/watch?v=GL5f-EcqPOc). And here, to prove that the sound board is just a push-fit in the Steinway is the piano on its back, and the sound board being tapped-out (http://www.youtube.com/watch?v=gCadaiKvDvQ&feature=related).

Then to my video:

I have created a new item in Tech Talk. I'm attempting to illustrate a subtle sonic effect, and hopefully the Flash video will speak for itself. just listen with reasonably good PC speakers or headphones and I hope (perhaps after one or two play throughs) the magic of the Harbeth cone material will be clear. For simulating our RADIAL? cone the piano is unmodified: to fake the sound of a conventional polypropylene cone I damped the sound board. Don't worry if you can't hear the difference immediately: we're dealing with a very subtle effect which we, and our users take for granted. It's akin to sampling fine wine: until your palate has experienced the very finest, you cannot imagine it from nothing; but once you have experienced it, it is unforgettable and instantly recognisable. It's definitely worth listening a few times.

Video is here (http://www.harbeth.co.uk/uk/index.php?section=products&page=designersnotebookdetail&id=16).

>

A.S.
18-11-2009, 12:27 PM
Test narration using new Flash facility. Please ignore - for concept proving only to reduce my time input.

/library/flashfiles/beat.swf