How come that when most audiophiles


follow the philosophy of "shorter, less complex signal path is better", they then wire their carefully chosen equipment to speaker cabinets filled with a boatload of transformers, capacitors, resistors, and drivers which exhibit gross non-linearities which are only compounded by adding them all together? I believe that the reason is the "specification game" again, where people believe that speakers must have a frequency response from DC to light +-3db, and as a result, speaker systems must have many drivers to cover the range. Notice the specs only show freq. response, and nothing about phase non-linearity induced by multiple crossover components. This seems to be a non-linearity in system philosophy where short signal path does not apply to speaker systems, but is paramount in all other aspects of the system. I use a direct input from source to OTL amp and DIY Fostex based 1-way speaker cabinets. The result is very natural, dynamic, phase-coherent,detailed, and revealing. The only non-linearities I have to deal with are the ones inherent in the driver/cabinet combo. With some careful design and impedence curve mods, I get a more musical sound than any "high end" speakers I have ever heard(and I've heard alot) as well as any of the multi-way speakers I've ever designed and built(also alot). Why do you think that there is this disconnect in thinking regarding short signal path as it relates to speakers?
twl
I agree, and that is why the theoretical "best" way to go is a low-distortion (e.g. Bryston, Westlake, etc.) external crossover, remove ALL crossover components from the speakers, connect each driver to its own set of binding posts, and drive each with a dedicated low-distortion monoblock amplifier (e.g. Bryston, Boulder, etc.).

Or, the easier approach would be to use a pair of quality pro monitors like the PMC AML-1, which is essentially a Bryston 10B / 2B / 3B built into the cabinet with custom drivers. They retail for about $5k, but you probably could not beat them for the price with "audiophile" gear.
Interesting topic, twl. I would like to hear a speaker like yours to better understand what you are talking about.

At last year's Home Entertainment show, I spent quite a bit of time listening to the Roman Audio speakers, which use the Diaural crossover. I felt there was a seamlessness and overall "rightness" about the sound that was different from most other things I had heard. The Diaural is a very simple crossover, and I wonder if what I heard is along the lines of what you are talking about.
OK, your looking for the impossible, how about this?

A single driver speaker that is pretty much flat from 24 Hz to ultrasonics, and can be driven by amplifiers from as little as 21 watts to over 700 watts?

This speaker would have no crossover, less than 36 inches of wire and exhibit an impedance of 8 Ohms.

It must have a radiation pattern that remains IN PHASE from 24 HZ to ultrasonics, and yet produce between 88 and 95 BD sound pressure level at 1 watt, measured from 12 feet, rather than the industry standard of 39 inches.

This speakers single driver must be engineered where it is so lightweight and quick, that it measures and weighs less than the air it moves.

It must use a magnet that will deliver equal force and power distribution over it's single driver at all times. This must remain true, regardless of the total power applied to the driver, the sound pressure level required, and regardless of the frequency it is producing.

Last, it cannot occupy any more floor space than a Vandersteen 2C speaker.
It seems to me that about 99%+ of all audiophiles are at the "mercy" of speaker manufacturers. Why speaker makers don't use simpler electronics I have no idea, but if what you suggest is really superior, it should show up in music quality and character; it should show up in "reviews", and it should show up in advertising. IMO, audiophiles purchase by reading, by looking and especially listening-- no?. Some enterprizing speaker maker could make a lot of money from the concept "simpler is better" if it's really valid (and significant)-- and if he's able to "sell" it.

Speakers are certainly the most "colored" of all components in a stereo system, and it seems to me that the "simpler is better philosophy" regarding speaker electronics, may not be the biggest reason(s) for the colorations-- just my opinion. Cheers. Craig
I reject the notion that there is a single technical and/or equipment path to audiophile nirvana. What works for one person may not work for someone else. While the system you describe might work wonderfully on several types of music, I doubt it would really excel at raggae, large scale orchestral, rap or electronica. At the same time these other systems may not have the textural magic and transparency that your system possesses. Any system involves making trade-offs. The success of each system will depend heavily on the biases of the listener. To paraphrase another Albert - it should be as simple as possible, but not more than necessary.
So Onhwy61, are you directing your comments to the topic posted by Twl, or to my post about the impossible speaker?

My post described a speaker that appears to have all the attributes that Twl was searching for.

Your argument as to the possible merit of that technology should be directed at him.
As the person who posed the topic here, I think all of your feedback is very thoughtful and correct in the aspects you all approached from. It is good to hear knowledgable people and their valued opinions. If any of you are interested in my speaker design and its strengths and limitations, I will give you further info here. First, it is not perfect as all of you well know, and since you cannot hear it and I'm not selling it, you can only go by my description. I'm sorry Albertporter, I cannot meet all the criteria you posted, but in many areas I came surprisingly close. First, I used a Fostex FE103 driver(4") on each side.Don't laugh yet. It posesses an excellent moving mass to spl ratio(efficiency) of 92db unloaded.It's very quick and has much less phase shift on transients than larger cone drivers. It is quite flat in response from 100-6000Hz, outside of which it suffers from impedance rises that I dealt with to improve the response. I designed a TL tube cabinet tuned to 41.7Hz - exactly half the driver's natural Fs which gives bass to that freq within 3db. I reinforced the lower mids with a disc shaped baffle behind the diffraction plane. The highs were cured with the impedance mod. The Z mod was simply a 24ohm non-inductive resistor paralled across the terminals. This reduced the 32ohm rises at the Fs and the gradual rise above 6kHz to less than 13.7ohms while only reducing the nominal system Z to 6ohms. (Simple product/sum theory) By reducing the Z rises, the driver efficiency at the frequency extremes was increased to within the 3db down level of the OTL amp from 41Hz to 20kHz(in room response at listening position). I now achieve 100 db peaks at my chair. The TL tube is loaded by the floor(front firing)at the base of the unit. Judicious tuning was required to get the correct tube/direct radiator balance to reduce midrange contamination. It stands 6ft. tall and 18" across(cylinder shape)- not much different than Van.2c's. Sweet spot is smallish due to beaming of highs. Bass is full, solid and quick. Midrange is golden. Transient dynamics are tremendous and all detail is revealed. Rock, jazz, classical, all sound very well given good recordings and imaging is pinpoint and concrete. The downside is spl of only 100db/1 watt peak. Max power handling is 3w continuous, 8w peak. My listening habits are not affected by this. Strauss's Also Sprach Zarathustra, Sade, Weather Report, Kinks, Joni Mitchell, etc. are all handled with aplomb. They look like 6' tall stick telephones from the 1920's. Very unique. I welcome all comments. BTW, I'm driving them with a Berning MicroZ OTL - 1watt output.(NOS tubes)
Post removed 
Other than the reference the other Albert (Einstein's description on how complex the laws of physics have to be), who's exact comment escapes me, my earlier post was solely directed towards Twl's excellent post. To be referenced in the same sentence as Albert E. is no disparagement.

To Twl: it sounds like you are really happy with your setup. It seems to accomplish everything you are looking for in music reproduction. My point is that someone else may have a slightly different set of musical reproduction priorities which could be better met with a different type of system. For instance someone with a large room who likes pipe organs may not be happy with your system. To quote that great American composer S. Stewart - "Different strokes for different folks, and so on and so on..."
I agree Viridian, that the above mentioned non-linearities should not be minimized. However, driver/cabinet non-linearities are present in all dynamic speaker systems regardless of the number of drivers. So are speaker/room interface problems and amp/driver impedance matches/mismatches. In my above design description, I attempted to explain the methods used to overcome some of these problems the best I could. I have admitted to max. spl limitations, and, at higher volumes, the attendant dynamic range limitations. These are not conflicting with my normal listening volume requirements. Of course, as you mention, IM distortion is present inherently in any driver that produces more than one frequency simultaneously as is also the case in multi-way systems, although I will agree that it may be more problematic in single driver setups. However, if you have ever done any speaker design work, you are aware that the above problems are not solved by multi-way systems. In fact, crossover slope overlap, phase shift,signal loss, even more complex impedance mismatches at diverse frequency points, image shift, and sometimes efficiency level are added to the problems you mentioned. I am not claiming perfection for my speaker design, but am merely pointing out the benefits of short signal path/low complexity philosophy as regards speaker systems in a properly designed one way system.
And, I might add, that since I have no commercial expectations, and have only done these speakers for my own pleasure, I have selflessy divulged virtually all of the design parameters to all on this board so that if anyone is so inclined, they may have an easier time of DIY. My sweat can be their gain. I have found the people here to be knowledgable and genuine and I am happy to share any of my experience, designs, tweaks, etc. with any and all.
Drubin is right, my punch line is that the speaker I described already exists, the Millennium One by Soundlab.

I might have fudged a bit on one part of the description. The Soundlab employs dual transformers separated by resistors, making for "sort of " a crossover. Other than that the specifications are correct.

Twl's speaker achieves it's musical magic by eliminating much of the multiple driver / crossover errors, producing a purity that attracts many to this type of design.

In addition to the designs high efficiency advantage, it allows use of ultra low powered SET amps, many of which have their own magic.

As Onhwy61 says, all of this is subject to each persons personal preferences. However, I admit I have been attracted to Hammer Dynamics and a low powered SET as a second system for many of these same reasons.
I can relate to what TWL states regarding the benefits of a wide bandwidth point source. I can also relate to his comments about "direct drive" from a power amp to driver. With that in mind, I don't know of any such designs that are capable of deep bass with authoritive weight while being driven by low power amps. For that matter, i don't know of any that are capable of high spl levels without compression when being driven by more powerful amps. As such, one simply chooses the trade-offs in a system that they personally find the least offensive and progress from there. Tis the beauty of having multiple systems. You can experiment in several different directions all at the same time. Sean
>
Not all speaker manufacturers avoid the phase issue. See, notably, Dunlavy which documents both frequency and time domain performance. Vendors such as Thiel and Audio Physics (IIRC) and some others also place emphasis on this as well.

Whether "musical" and "accurate" are the same or different objectives is another story, and reflect on the individual designer's (and listner's) preferences.
I don't see any reason whatsoever that an ideal speaker "must" -

- have SPLs from 88-95 dBW at 12 feet?
- have an 8 ohm impedance?
- have a driver that measures and weighs less than the air it displaces?
- use a magnet that delivers equal force, etc. etc. etc.?
- must occupy less space than a Vandersteen 2C?

Some of those specifications are essentially irrelevant to the end sound. But if we're looking for ideal technical accuracy, including phase response, it'd be extremely difficult for any speaker to match or exceed the Dunlavy SC-VI, which near as I can tell doesn't meet any of the above except for sound output.
Here is an excerpt from Twl, the creator of this post.

Notice the specs only show freq. Response, and nothing about phase non-linearity induced by multiple crossover components. This seems to be a non-linearity in system philosophy where short signal path does not apply to speaker systems, but is paramount in all other aspects of the system. I use a direct input from source to OTL amp and DIY Fostex based 1-way speaker cabinets. The result is very natural, dynamic, phase-coherent, detailed, and revealing. The only nonlinearity I have to deal with are the ones inherent in the driver/cabinet combo.

MY POST BEGINS HERE, ON THE TOPIC AS BEGUN BY TWI:

His point being that simple, single driver systems retain more of these specific properties than the more complex designs.

My "impossible" speaker posting was exactly on topic because it listed ideal standards for that single driver system, including the crossover design and ability to be driven (impedance load and efficiency specs).

If you do not believe that these are important to a speakers design and performance, you have much to learn. There is no free ride in the realm of high performance audio, and the better and more coherent the speaker, the more likely the amplifier will deliver the information from the source and electronics that follow.

The speaker is the most flawed component in any system, it must convert electrical energy to mechanical energy. This is an extremely difficult task with the best conditions, and if the speaker has any design flaws to overcome, the ability to produce near believable results become impossible.

Phase coherency, driver mass, crossover phase distortion, notch distortion, cabinet resonance and wildly varying load impedance are but a few of the obstacles to making a speaker behave properly when driven by the electrical signal.

My "ideal " speaker addresses more of these problems than most designs and may be proven in the real world by auditioning the speaker these specifications represents. Twl's ideas are not new and they are not incorrect. He is seeking to accomplish this goal with dynamic drivers, I simply pointed out that an electrical version already existed that conformed to these ideals.
Thank you Albert. By the way, I checked out the Millenium 1 on the web. They look very nice and the TAS review was very interesting, however I couln't afford them and my amp wouldn't drive them. I can't help feeling that, for my needs, I saved the $14,500.00 difference and still achieved a very satisfying result.
No doubt, and I am pleased that you found happiness with your system. The important thing is finding the special combination of equipment and speakers that allows you to listen to the music without constantly thinking about how the system "sounds".

If you are there then you are successful, regardless of how much money is spent.
Two comments for Twl:

Where does the Quad 63 series fit into your schema? It's a phase coherent point source design, yet at the same time with its miles of wire it definitely ignores the shorter is better philosophy.

Shorter is better may only apply to the analog world. Correctly implemented DSP processing is essentially transparent and opens up the possibility of digital crossovers driving multi-amp/multi-speaker arrays with phase coherent in-room results.
The key, as pointed out by Onhwy61 and Craig is "Everything should be made as simple as possible, BUT NOT SIMPLER". Not all speaker designers are willing to compromise phase and time performance just for simplicity's sake, even in their wide bandwidth designs.

As TWL has shown, with a different set of design goals, one can make a very satisfactory "short wire" speaker. Different goal, different approach, happy listener.
Albert, you posted a list of features and elevated them to the level of standards for an ideal speaker. To quote you:

"My "impossible" speaker posting was exactly on topic because it listed ideal standards for that single driver system, including the crossover design and ability to be driven (impedance load and efficiency specs). If you do not believe that these are important to a speakers design and performance, you have much to learn."

For many of the features it is irrelevant whether you are referring to a single- or multi-driver speaker, and I’m still uncertain as to the need for such absolutes. For example, why do you specify an 8 ohm impedance as ideal? Are 4 ohm speakers inherently worse? Or why must speakers have a sensitivity in the specific range you mention? There are many audiophile-grade amplifiers that can drive the 4-ohm, inefficient Thiels to very high volumes without working up a sweat. And the size constraint baffles me completely.

There is more than one approach that leads to state of the art, especially in speaker design. Some violate one or more of your standards by a wide margin - that does not make it a design flaw.
Two responses for Onhwy61. It is true, the Quad 63 has alot of wire that effectively is part of the transducer. Sort of like the dynamic speakers have lots of wire in the voice coil. These are simply a part of the transducer design that one picks. You could say the perfect transducer is none, but that leaves us without music. So, I'm not saying reduce to zero, it is already effectively a one way system - with disadvantages, I'll grant you, but also great strengths. I am a fan of the Quads and other electrostats for their legendary strengths. Also, I've never designed or built electrostats, but have done many dynamic types and have wrestled with the multi way issues many times. I have concluded that single-way provides the most direct and pure interface currently possible. I will settle for the one way limitations for my system. Regarding digital crossovers having great potential, I'll believe that when I hear it. Thanx for adding to the post.
Within my own post is the following: "it listed ideal standards for that single driver system."

The topic as posted by Twl is on the subject of single driver systems. My post addresses that subject much the same as an opinion as to the best CD player, preamp or turntable.

In Twl's world of single driver speakers, most if not all are powered by low output SET amps (as already mentioned in my post). Those designs REQUIRE the efficiently, impedance and other factors that I listed as ideal.

Simply put, a valid opinion and answer on this topic.