One thing that comes to mind is Geddes opinion that second and third harmonic distortion from loudspeakers is not as objectionable as higher harmonics. I think he's probably right. I think measurements of second harmonics are still pretty good information, since they are easy to make and they indicate asymmetry. That will also introduce other even harmonics and the ones higher up in frequency may be the ones that are really noticeable, even if at lower amplitudes.

Something else that comes to mind is the matter of nonlinear distortions verses linear distortions. That is something Geddes brings up a lot, and was also mentioned recently by Dave Williams. My thinking is that efficient, powerful speakers used at home hifi levels are probably used pretty much in the linear regions, at least where electro-mechanical properties like suspension stiffness and voice coil resistance and motor strength are concerned.

I think the mathematical models are useful. I don't think it is wise to assume perfectly pistonic behavior of the diaphragm, perfectly resistive impedance of the voice coil or perfectly linear excursion of the motor. It is good to realize that isn't the case, and to take it into account when making a design. But the models that make those assumptions are useful tools.

I also think that it is worthwhile to design systems that are very tolerant of parameter shifts because they're going to happen. And I think that doing things to reduce the things that cause nonlinear distortion is always good too. But first, you must define them and know what they are to begin dealing with them.

There are many other distortion mechanisms too. There is throat distortion, caused by the nonlinearity of the air in extreme compression/rarefaction cycles. There are the high order modes that Geedes often focuses on. There is doppler distortion from moving diaphragms. There is intermodulation distortion. There is compression. Then there are other nonlinearities that don't have specific names. There is the modification of waveforms in some capacitors such as electrolytics and ceramics. There is the effect produced by a coil or transformer in magnetic saturation. There are the changes of a conductor when hot. Or of a semiconductor, like the carbon in resistors or the silicon in transistors. There is the non-uniformity of magnetic flux at the edges of travel in a loudspeaker, and the asymmetry produced by flux modulation.

What do you guys think? Lots of things to consider here.

What we have at the moment is an interesting market. The 'hifi' market that you see in the stores,and the 10% of those who are simply not happy with those products,be they engineers,enthusiasts,musicians or anything. The vast majority are happy with these cheap system until they hear better, so the real 'hifi'owners will always be a minority group. I feel that the real improvements will always come through the prosound area and that there will always be the people who are happy with their technically bad, but subjectively good {to them atleast}sounding systems.

Once into this group theres so many ways to go,but to me theres 2 main branches each aiming at different goals.
the low efficiency omnidirectional guys,and the directional high efficiency guys. These to me seem to be opposite and equal but just 2 extremes of the audio pie full of varying flavours! One cant be better than another,and aslong as people admit where they are measurement wise and arent operating in a physics free world I dont mind if they listen through 5" fullrangers or MTM d appolito 2ways..

Cheers!
Mike.e

So, no, I don't agree with your comment.

Certain alignments are very sensitive to electro-mechanical parameters. It's like they're close to the edge, often providing maximum extension but also needing tight tolerance to prevent underdamped response. Other alignments are much more tolerant of shifts and could not become underdamped except from extreme changes which are completely out of range.

Here is a good primer on distortion in general, with a cool applet.
http://www.mindspring.com/~j.blackstone/dist101.htm

-akhilesh

The one at http://www.mindspring.com/~j.blackstone/dist101.htm is really cool, because it lets you actually hear tones with various amounts of harmonics. You can start with a single tone sine wave, then add in components of whatever harmonics you like to hear the resulting sound.

The other link didn't work for me. So I backtracked to find the host at Pete Millet's Audio Page, http://www.pmillett.addr.com. That looks like an interesting site as well.

My intention was to say that the two approaches are opposite and produce rather different results in room. Is this your opinion also?
Whether one is more correct than the other...
Perhaps my NZ english prevents me from understanding quite what you mean in this sentance below

Regards
Mike.e

omni has no reflection free time lag after the direct sound while the directional speaker does - if pproperly aimed

Picture an omni source - it hits all the walls, near far what ever, on the first wave. Now picture the directional source, picture it very narow, like a beam of light. It hits only one wall on the first pass, another on the second, maybe two, its at least three or four reflection before the reverb field begins to build. The omni happens immediately.

This takes a lot of thought, but think it through and you'll see what I mean.

So I am saying that one type is definately better than the other.

The ear integrates over a period of time. The time is variable with frequency but it has a typical value of 10 - 20 ms. The more reflections that arrive in this time period the more difficulty the ear has determining an unambiguous sound localization. Thus this first few ms. is critical to imaging and coloration of a loudspeaker. Sure the room can help, but in small rooms the lodspeaker and room must both be done correctly to get the best imaging and sound quality. Neither one alone can do it.