Originally Posted by chroot /img/forum/go_quote.gif
Hey guys,
I'm relatively new here, and do not have as much experience with hi-fi gear as many of you. However, I do have something most of you probably do not: a Master's of Electrical Engineering from Stanford. I work as an integrated circuit designer (yes, including amplifiers) for one of the largest high-performance analog companies in the world.
I would like to make a few points, some from an electrical engineering perspective, and some simply observations of human behavior.
1) The "difficulty" misconception.
Most audiophiles want to believe that designing and building hi-fi sound equipment is difficult. They are used to seeing very elaborate equipment full of exotic materials and components. Their experience (and the messages targeted to them by the manufacturers of such equipment) leads them to have a very skewed understanding of what's "difficult" in the field of electrical engineering.
To make it clear: shuffling bits across a noisy bus at 10 Gbit/s is a difficult problem. Building a 98% efficient 200 kW microwave amplifier is a difficult problem.
Sending a 2 V peak-to-peak, 20 kHz band-limited signal across a 12-inch piece of shielded wire is not a difficult problem. At all. It's a cookie-cutter problem, solved handily by the basic circuit topologies shown in dozens of electrical-engineering textbooks. You can build an audio line-driver with essentially any THD you want, for example, without having to use your brain at all.
It's hard to build excellent speakers. It's hard to build excellent headphones. It's hard to build excellent codecs. It's easy as pie to build excellent line-drivers that will work well over anything from cryogenically-treated super-engineered science-project cable to RadioShack's cheapest speaker wire.
2) The capacitance misconception.
Many of the people who try to have empirical discussions about cables will bring up figures like capacitance. They mean well, but they are misguided. They are almost universally using a linear no-threshold model without realizing it. A linear no-threshold model is embodied in the concepts "less capacitance always produces equivalently better sound" and "you can never have too little capacitance."
Linear no-threshold modelling has unfortunately become downright mainstream in modern society. Everything from exposure to carcinogens and radiation to climate change is now almost exclusively discussed in the context of a linear no-threshold model by the popular media.
The problem is that a linear no-threshold model is essentially never physically valid. In the context of audiophile gear, many people misunderstand that capacitance is necessary for your amplifier to operate properly. Almost all op-amps, for example, require some load capacitance to remain stable. If you lower the capacitance, the amplifier will begin to distort or, worse, spontaneously oscillate. Op-amps are designed for efficiency over a relatively broad range of acceptable load capacitances, and thus capacitance does not obey a linear no-threshold model. Furthermore, the engineers who built your sound equipment stuck some capacitors on the board specifically to make sure the amplifier always sees enough load capacitance.
Besides -- look at the specifications on your cables. Almost all of them, even the super high-grade cables, will have capacitances in the ballpark of 10-20 pF per foot. The lesson is that you should be more concerned about the length of your cables than the type of your cables, if you worry about capacitance at all. (And you probably shouldn't.)
Besides, if you really want zero-capacitance cables, you can do it with basic matching networks at both ends. All you need is a couple of small, cheap inductors and viola your cable presents a purely real impedance to your amplifier.
And don't forget those chintzy 3.5mm mini-phono plugs! They're terrible from an electrical engineering perspective. They have large capacitance, large contact resistance, etc. If you're really concerned about interconnect, why not use modern connectors like SMA or SMB that have vastly superior electrical characteristics? That'll certainly have a much larger effect on the overall transmission line than simply connecting two mini-phono plugs with a wire as big as a baby's arm.
3) The resistivity misconception.
The second figure everyone brings up is resistivity. Since you're not driving power over your interconnect, resistance should be the least of your concern. If you look at the telegrapher's equations, you'll see that real resistance only contributes to attenuation. In other words, you'll lose some signal amplitude over a length of cable if its resitance is high. Real resistance does not alter the transmitted waveform in any other way; it does not affect waveform shape or spectral content.
Oxygen-free copper and so on provide improvements in resistivity of at most about 2%. This means, well, essentially nothing. It means you'll have to turn the knob on your receiver a couple more microns clockwise if you use normal cables.
4) The size misconception
Many people intuitively believe that larger conductors are better. Some people buy luidcrously large conductors -- large enough to use as mains power cabling for a hospital -- in the hopes that it will improve sound quality.
It's already been discussed here many times, so I won't belabor the point, but there's no point in using an enormous cable to connect integrated circuits. Forget about the pins on the iPod dock. Forget about the 10 micron board traces. Realize that inside the black plastic package of your integrated circuits, the signals are being carried on tiny gold wires thinner than a human hair.
And, I just have to pick on TheMarchingMule a bit:
Anytime someone tries to personify electrons, or make weird analogies between fluid flow or feng shui and electron conduction, be wary: that person doesn't know anything about how signals actually propagate through wires.
5) The measurement misconception.
Anyone who tells you they can hear things that cannot be conclusively shown on a decent-quality oscilloscope is lying to you (and probably to themselves, as well). Measurement equipment is hundreds or thousands of times more sensitive than human senses. It's positively silly for someone to claim that they can hear something that a high-end spectrum analzyer cannot detect.
6) The tone-color misconception.
Audiophiles openly admit that they buy amplifiers and headphones because they like the way they color the sound. People love tube amplifiers, for example, specifically because they are such poor amplifiers, from an electrical engineering perspective. They color the sound quite strongly, changing its spectral content enormously. Yet people enjoy the sound, and pay loads of money for them. On the other hand, the same people will spend hundreds of dollars on cables that they believe to be superior because they do not color their sound.
6) The price misconception.
As has been mentioned (and demonstrated), people have a fascination with price, and tend to think that anything expensive must be good.
Anyone who's ever listened to music before can tell you that a $20 pair of headphones don't sound as good as a $100 pair of headphones (unless they're Bose). It's so easy to tell the difference between headphones that the market economy reliably drives their prices to a reasonable delta -- given equal market exposure, the good ones cost more than the bad ones, almost universally. (Marketing screws this up a bit by affecting market exposure, but I digress.)
On the other hand, almost no one can really tell the difference between cables, even experienced audiophiles. Perhaps a few people really can, but the majority of the market cannot. This means the market cannot reliably set prices, which is why you find cables priced at anything from $1 to $3000.
The stock market works the same way, by the way -- it's hard to make a windfall on large, well-known stocks where everyone has the same information, because everyone pretty much agrees on what the prices should be. When you get down to the small, relatively unknown companies, no one really knows which ones are better, and the prices between them can fluctuate wildly with no real rhyme or reason.
8) What to do?
Now for the warm and fuzzy part.
I personally feel the best hi-fi system is the one you enjoy the most. "Enjoyment" is a subjective term, of course. Personally, I enjoy musical variety more than anything else. I love having (and knowing) a large collection of music, because I enjoy always having the right music for every occassion. I like always being able to get people dancing at parties. I like having people come up to me and say "wow, I love this music, what is it?" I like taking road trips with people, and having them tell me later that the best part of the trip was the music.
Other people like the look of their sound systems. They want lights and meters and fancy looking interconnects. Maybe they want to impress people, or maybe lights and meters are just their thing. To them, it's incredibly important that their system look as good as it sounds. So be it.
Other people simply view cables as the "final touches" on their masterpiece sound system. They may admit that it doesn't really affect the sound perceptibly, but they enjoy the system more because it feels "complete" and finished to them. So be it.
Again, the best hi-fi is the one you enjoy and use the most.
- Warren
|
