Just found this interesting thread. Keep up the exchange of ideas guys. Let me post some of mine.
(1) Amplifers ideally should work as a "straight wire with gain." I think this is the highest end in amplifier design. Among different brands that achieve this end with the fewest compromises, they SHOULD sound the same on an instantaneous A/B test. But very few do. There are simply too many variables and design compromises due to production cost contraints, not the least of which is speaker matching, making many consumer products sound less than ideal.
There are many amplifer circuit topologies and designs that by their operation, do impart some sonic character or coloration to the sound they produce, from negligible to considerable. I have come accross amplifiers that exhibit non-linear responses as you go up and down the db ladder. I have come accross amplifiers like the entry level Sherwoods that have severe roll-off characteristic at the extreme ends of the audible spectra. On the other hand, some amps like the Aragon, Krell and Sansui can really go down to DC as verified by my spectral sweep. And they exhibit the most musical and tightest bass notes I've ever heard. And there are some that can generate frequencies only dogs are interested to hear. I have come across amps like the NAD and Carver which retain their non-fatiguing character even at very loud volumes exceeding their clipping points. That's because they use special circuits to effect tube-like soft clipping at high volumes.
Then there are amps that sound anemic and even harsh at low volumes, coming to life only at high volumes. Many powerful professional gears like the Peavey, Hafler and QSC behave this way. I think such amps were really designed to be used at high volumes.
(2) Speakers do affect the final sound you would hear from an amp. I have once thought that speaker types and loads didn't matter. Unitl I came across difficult speaker loads which seemed to prefer one amplifier over another to sound best, while sounding awful or thin in another. Often, the muscle of an amp can dicate the outcome. I've heard some planar speakers like the Magnepan sound anemic on a 100 wpc Marantz receiver which otheriwse sounded gorgeous on a Wharfedale. But the Magnepan's breathtaking brillance came to life with a 200 wpc Bryston and an Aragon.
Then there is this issue of Damping factor in amps that I also didn't consider materially consequential to sonic delivery. But I must confess listening to a Rotel and an Acurus with high damping figures compared to a Onkyo and a Yamaha with much less on the same B&W load convinced me that adequate damping has an effect on the tightness and accuracy of bass notes.
(3) Reproducing adequate SPLs in a room to make convincing soundstage realism require some muscle. While it is possible to extract only a few watts of power to deliver ear-splitting decibels in a small room like in many SET gears, power or the lack of it can spell the difference between constrained and easy listening. An amp with little dynamic headroom can benefit from much continuous power. While an amp with great dynamic headroom can be moderately powered on a continuous basis. I really would have no way of testing this in a showroom except to trust the power ratings on its technical specsheet. I am not aware of any salesman who will allow the custumer to test the maximum abilites of an amp. And that is where my distress comes when specsheets on power rating are less than candid.
Here's where much of the misunderstandings come in the other threads. The quality of the sound form an amp may, on first inspection, have nothing to do with the power ratings on the specsheet. But it does when you start to drive the amp at volume levels you were misled to expect it can handle based on an overstated power specs. Since there is no regulation in this area and with so many standards like DIN, JEITA, FTC, EIA being bandied around as the bases for power rating, power specs are used by manufacturers to tell half-truths that make excellent marketing tools. An amp can be rated as a 30wpc or a 100wpc amp depending on the standards and the measuring condition used. But enough of that.
Often the design of receivers and amps can also be misleading. In many receivers and integrated amps, the volume control circuitry seems to be designed in a way as to bring the amp to full power already at the 10 o'clock position. The better to impress. "hey, if it can sound so loud at that point, imagine what it'd be at max position." All i can say is, it will sound with so much garbage THDs at the max position.
(4) There are amplifiers that behave like constant current sources and thus have difficulty driving low impedances. These amps usually require minimum 8-16 ohm loads, will not allow simultaneos A and B speakers to be driven, unless both are 16 ohms and generally exhibit poor dynamics at high volumes. They may sound gorgeous at moderate volumes, but exhibit severe harshness the louder it gets due to the clipped harmonics its inadeqaute headroom cannot reproduce. Some amps like the Yamahas use a current divding circuitry at the back panel to limit the current going into a 4-ohm load, thus, making the sound appear to be identical between a 4ohm speaker and an 8ohm speaker. They are specified to handle 4 and 8 ohm loads. These amps can really benefit from a high continuous power rating.
Then there are amps that behave as constant voltage sources that exhibit high current capabilities to drive difficult loads down to 1 ohm, but mostly 2 ohms. These amps can sound detailed and full-bodied even at high volumes courtesy of their dynamic headroom and generous power supplies. The two go together. An amp cannot have a good dynamic headroom without prodigious power reserves based on its continuous power rating. And the sound will be dictated accordingly as easy or constrained.
I've heard my old Sansui and Onkyo literally pant for breath on loud passages of most classical pieces if I wanted to bring the quiet passages to my listening levels. Maybe not too obvious with pop and jazz music. But the constraint is audibly evident by the lack of delineation of instruments in a complex orchestral passage and a seeming congestion thereof that makes listening strained and fatiguing. Power is not everything in amps. But it does determine much of the sonic quality when listening at high SPLs - something many audiophiles think are good qualities of an amp, especially when their aim is to reproduce realistic SPLs of the same magnitude one gets in the front seats, much less the podium.
Personally, I look at the following as good attributes of an amp, given a good set of speakers of begin with.
(a) Sufficient power reserves for dynamic headroom, at least 1.5 db from the continuous rated power. That means, a 100watt amp driving an 8-ohm speaker should be able to deliver instantaneous peaks of 150 watts or more at that load. I may not yearn for realistic SPLs, but I sure wouldn't want to miss the dynamics of a piano or a drum when initially struck which can exceed 1.5db in some cases. There is no way to test this at high volumes (see if you can with the salesman over your shoulders.) So only the specsheet will be my basis. The weight of the amp is generally indicative of this. The heftier the better for the same power rating. In addition, I go only for High-current amps. Aragon, Acurus, Bryston, Classe, NAD, Rotel, Harman Kardon, Parasound, Outlaw, Krell, Sunfire, Theta, easily come to mind.
(b) Linearity and extended response. Tranparency is the word used by some. I expect the amp to deliver everything that the source player can deliver as transparently as possible. So if my SACD or DVD-A player can deliver 2Hz up to 88khz, + or - 1db, I expect nothing less from my amp. I don't care if some of the frequencies are inaudible. For sure the harmonics are there and a lively debate is raging as to whether such inaudible harmonics can differentiate a good amp from an exceptional one. Maybe it does, maybe it doesn't. But I wouldn't want to miss it if it did. (Better to err on the side of plenty.) In fact, only an amp with a 2hz-100Khz +0, -0.5db would be a good start for me. And this should be the response at whatever volume setting up to the maximum THD levels of, say 0.09%.
(c) Robust construction simply indicates a responsible and thoughtful design. You can't expect aluminum heatsinks to prosper in a plastic casing. Amplifers anmd speakers are probably the least subject to technological changes. At least not with the same obsolescence as players, preamps and recievers. So I'd expect my investment on an amp to last longer. A robust construction would go a long way to doing that. And may still command some residual resale value after years of service. Talking about robust construction, a lot is said about using 5-way banana terminals over the spring-types for speaker connections. Personally, I prefer the former. I won't eschew the notion that spring-type speaker terminations are a joke, but if they can't accommodate 10 gauge wires, I'd shun away from amps that sport them.