Taralabs cables

Thus far, we have fairly well established the power that we must have in order to avoid outright overload when reproducing original orchestral level through a speaker of known efficiency. But it is not all the power we should have on hand, because there’s more to fidelity than just reproducing sound at the proper volume.
Anyone who has perused an amplifier’s power-vs-distortion curve will have noticed that distortion rises gradually with output until just below the overload point, beyond which the distortion skyrockets. This is one reason why a high-powered amplifier is likely to sound better than a low-powered one even at every low power levels. They may both be operating at well below their overload point, but the fact that the high-powered one is running at 1/10 of full power when the other is at 8/10 of full power will mean that the former is contributing less distortion at all times  and this will generally show up as cleaner, more "comfortable" sound.

There's a second reason why high-powered amplifiers should outperform low-powered ones, even at low output levels. It is customary to equip an amplifier with an output transformer that is no larger than it has to be in order to yield full rated power in the middle range. The British are still making low-powered amplifiers with substantial output transformers, but the prevailing attitude in the US seems to be that the low-powered amplifier is sort of a stopgap component, to tide the buyer over until he can afford to purchase something good. There is rarely any attempt to design a really good low-powered amplifier. As a result, the typical 10-watter, even though it may well meet its rated power at 1kHz, is severely limited in power capability at both ends of the spectrum. The power loss is usually most severe at the low end, where there is often a great deal of energy in the audio signal, so the unit may only be able to deliver half, or less, of its rated power before the program material overloads it 

Even the biggest, costliest amplifiers exhibit this power loss at the frequency extremes, but in these, the losses don't usually start until well beyond the audible range.
Let's assume now that we have access to an amplifier's power response curve, and can see that it will deliver its full rated power to 20Hz. Is this any guarantee that it will sound the same, at low levels, as a high-powered unit? It is not.
Power response curves show the power levels at which different frequencies will generate the same 2% distortion at which the midband power is usually rated (fig.3). What they fail to show is distortion at less-than- maximum power levels. An amplifier that yields 2% distortion at full rated output may yield 0.2% at half power, or its distortion may never drop below 1% regardless of how little power we drive from it. And since we do most of our listening at power levels far below overload, the amplifier's minimum distortion, or "residual" distortion, is of considerable interest to us. Here, again, is where the typical low-powered job falls far short of its heftier ilk.

The light output transformer in most low-powered amplifiers is susceptible to core saturation at low frequencies, and even though this may be held low enough to meet overload limits down to, say, 20Hz, it nonetheless imposes a severe limit on the amplifier's low-frequency residual. Thus, typically, the low end will exhibit increasing distortion with decreasing frequency, even at the very lowest output power levels. At 1 watt, where the mid-band is contributing only 0.3% or so distortion, there may be 1% distortion at 30Hz.
Actually, it is a rare low-powered amplifier that will produce as little as 0.3% distortion at low levels, even through the midband. Most of them, sloppily designed as they are, have enough distortion in their earlier stages to hold their residual at about 0.75% no matter how good their output stage may be, so they can never sound as good as the more carefully designed high- powered units. The few exceptions to this rule are so costly that one might just as well buy a higher-powered unit and be done with it.
There are extenuating circumstances occasionally, though. Loudspeakers and amplifiers that ate designed specifically for one another should be used together regardless of the amplifier's power rating. Some speakers are fragile, and will burn out if hard-hit by a hefty amplifier. Fusing helps, but the series resistance in the line reduces the electrical damping applied to the speaker, inhibiting the amplifier's ability to prevent spurious cone vibrations. Consequently, if you must use such a speaker, it's advisable to bypass its fuse, and couple the speaker directly to an amplifier that won't be able to damage it.

High-power advocates have always claimed that one reason a high-powered amplifier sounds better than a top-notch low-powered job, even at low levels, is because the big one's reserve power gives it better control of the speaker's voice-coil. It was reasoned that a large reserve of power, operating through a tight negative-feedback system, could bring more power to bear more rapidly for suppressing spurious vibrations of the speaker cone. This sounded plausible, until the first of the all-transistor amplifiers came along and befogged the issue.
Transistors just do not behave like tubes. Transistor amplifiers whose measured distortion is higher than that of the cheapest "hi-fi" amplifiers somehow manage to sound much better than they should, and the absence of an output transformer from most transistor amplifiers (the low-impedance transistors connect directly to the speaker) eliminates most of the annoyance value of marginal overload on peak passages. As a result, a transistor amplifier seems to produce far more clean power than a tube amplifier of the same rated output.