Who makes the lowest distortion speaker drivers

Electrostats???? No way. :) high in distortion, limited dynamic range and wayyyyy off in FR. GREAT in terms of detail and having that "headphone" level of detail.

So, I refuse to read the rest of the thread, but I thought among the top were AMT's from Mundorf and other top vendors, as well as Accuton. Am I wrong?

Best,

E
 
You're pretty much barking up the wrong tree, if you think low measured distortion equals good sound quality then you really need to do some listening, to both power amplifiers and loudspeakers (and everything else of course). However you will probably find power amplifiers make a bigger difference, forget the numbers, use your ears, Hi-Fi is the only hardware that some want to judge in an entirely different way than which it is used. How do you judge a new TV? do you look at the numbers or do you actually judge it subjectively? Trust your own judgement, it's you who has to live with it.
I assure you there will be certain idiots here who will assure you that the numbers tell all, don't believe it! Don't believe me! Listen for yourself, don't listen "critically" you won't hear anything, just relax and enjoy the music and I promise you will hear everything, do NOT try, relax forget all the stress of making that decision, the stress of it is where the problem lies.

PS For the "numberers" try the above method, but lay aside your preconceived notions, your mind is getting in the way, it is really easy and really effective, just enjoy the music and don't try to analyse anything. I promise this really works.
 
PPS Of course if you are not a music lover and just some kind of technology geek then this won't work for you, in which case keep the specification sheet in front of you while listening, then you can see it's the best, just don't take up photography, then you will have to judge something subjectively. Looking forward to the rebuttals BTW!
 
The matter of the hifi seems to reprocduce an complex audio signal.
(the THD measurements are a pertinent tool IMHO when used in a precise contex and detailled measurement protocol)
And people here are apparentely trying to judge the quality of thier stuff with sinewaves measurements… or other strange noises.
The basic logic seems to indicate that we must compare the real audio complex signal of the input to the output one.
Why ignore reality ?
 
To come back to the original question:

Everything that prevents the cone to move in a way exactly proportional to the signal driving it may cause non-linear distortion. Some sources of these non-linearities that come to mind are:

1) suspension linearity;
2) linearity and robustness of magnetic field in gap;
3) linearity of EMF of coil in gap;
4) break up and rattles; buzz and rub;
5) eddy currents in VC-former and pole piece;
6) frequency dependent energy storage (resonances).

Now, much in a loudspeaker driver is a compromise between one thing or another. For example, a perfectly linear suspension is good for distortion, but increases the risk of the driver bottoming out. A non-conductive VC-former is good to fight eddy currents, but will also have worse thermal conducting properties than a conductive one, thermal modulation being the 7th form of potential distortion.

Some other aspects of a driver that define quality are a matter of cost. For example, a conducting layer of a non-magnetic conductor over the pole piece will fight flux modulation, so improves distortion source #2. However, it increases the magnetic gap, so to achieve an identical BL-product, you will need a stronger magnet, increasing cost. The leads feeding the VC are usually flapping around, forming a potential source of buzz and rub, #4. One improvement can be to have the leads symmetrical, but this requires a second terminal. Another is to weave the leads into the spider. Both solutions cost money. Since symmetry is important for linearity (#1,2,3), the precision with which the driver is constructed also comes into play. This also will drive up costs. The same goes for the linearity of the BL product; it costs either efficiency or magnetic power to improve. In order to give their cones the right kind of resonance profile, manufacturers sometimes resort to costly steps such as thick oxide layers, surface treatments or inventive terminations to the surround. All cost money.

When selecting a driver to fit into a speaker system, a lot of these factors come into play and it is impossible to recommend this loudspeaker manufactured over another. What is the best driver starts out by defining what its intended use is. Without knowing that, nothing absolute can be said about it. I can just give you some elements to take into account.
 
To come back to the original question:
Everything that prevents the cone to move in a way exactly proportional to the signal driving it may cause non-linear distortion. Some sources of these non-linearities that come to mind are:
1) suspension linearity;
2) linearity and robustness of magnetic field in gap;
3) linearity of EMF of coil in gap;
4) break up and rattles; buzz and rub;
5) eddy currents in VC-former and pole piece;
6) frequency dependent energy storage (resonances).
Now, much in a loudspeaker driver is a compromise between one thing or another. For example, a perfectly linear suspension is good for distortion, but increases the risk of the driver bottoming out. A non-conductive VC-former is good to fight eddy currents, but will also have worse thermal conducting properties than a conductive one, thermal modulation being the 7th form of potential distortion.
Some other aspects of a driver that define quality are a matter of cost. For example, a conducting layer of a non-magnetic conductor over the pole piece will fight flux modulation, so improves distortion source #2. However, it increases the magnetic gap, so to achieve an identical BL-product, you will need a stronger magnet, increasing cost. The leads feeding the VC are usually flapping around, forming a potential source of buzz and rub, #4. One improvement can be to have the leads symmetrical, but this requires a second terminal. Another is to weave the leads into the spider. Both solutions cost money. Since symmetry is important for linearity (#1,2,3), the precision with which the driver is constructed also comes into play. This also will drive up costs. The same goes for the linearity of the BL product; it costs either efficiency or magnetic power to improve. In order to give their cones the right kind of resonance profile, manufacturers sometimes resort to costly steps such as thick oxide layers, surface treatments or inventive terminations to the surround. All cost money.
When selecting a driver to fit into a speaker system, a lot of these factors come into play and it is impossible to recommend this loudspeaker manufactured over another. What is the best driver starts out by defining what its intended use is. Without knowing that, nothing absolute can be said about it. I can just give you some elements to take into account.

linearly related to what ?
 
Linearity with regards to VC-position (suspension, mag field) or current applied (magnetic field in gap).

Coil displacement relative to the basket ?
That displacement is not linear, there is 3axis of rotation and 3axis of translation.
Imagining a perfect linear displacement of the whole parts of the mms on a large frequency range is a fantasm.
The suspension preloading is not linear (basket to mmd) and the two air loading problems plus all others disturbances add a lot of complexity to it, the mag field is barely homogenous (in the gap without the coil :D).
We can be approximative, but in that case, don't have to be surprised if the result is not the expected one.
 
Exactly, whereas the perfect amplifier has already been designed (many times over)

IMHO the perfect amplifier will never and had never been existed in this world.
I don't know how far from perfection the best amplifiers are but the loudspeakers are extremely far from perfection... million miles from it... probably light years from perfection.
The reality is that they are all very, very bad due to our limited technology.