Well, NwAvGuys mention of double-blind testing busting the cable myth (ie. it turns out that cheap cables do the job just as well as expensive ones) is telling. However, even with that the fact is that I have some defective cables that sound completely awful. Even NwAvGuy mentions poor quality connectors potentially causing issues. I remember at school soldering a new plug on to my cheap headphones only for the audio quality to be catastrophic; I threw them away. (The reason was probably that I didn't know how to solder properly; doing the same job now I have no issues). So it's quite possible that there are sonic improvements under certain conditions: but that shouldn't be happening but for a flaw.
But my own argument leans more towards the fact of non-ideal amps masquerading as hifi mixed in to the fact that music/audio is such a subjective experience; even more so than photos/pictures. People are hearing real differences but not for the right reasons. For example, another factor I didn't mention is people not listening to the same tunes when auditioning equipment. I have tracks that are mastered with quite a bit of 3D DSP work. Others lack it. I can get the same effect by using the 'stereo width' setting in rockbox (set 'channel configuration' to 'custom' to enable that). I prefer it; I personally dislike in-the-audience recordings and prefer to be immersed in the middle of the band. Phase issues can do that with amps: give a 3D effect. The reviewer will be exulting the soundstage. Another reviwer will prefer in-the-audience, in which case poor channel separation will be an advantage. A proper amp should have no effect on soundstage. Impedance interactions can also 'improve' audio by their negative effect: for example a bass weak headhone gets a boomy lift at about 100Hz; some will like that since tastes/genetics vary so much. Some phones will have their characteristics helpfully moderated, but only by virtue of two flaws improving on one another is this happening (a hot treble being attenuated by high output impedace, for example, or vice versa, I can't remember; as well as the taming of an overly bassy headphone).
It's a heady cocktail of factors and a constant moving target. Magazines and profit motives don't help to dispel it.
Let's go through your list:
- improved soundstage: phase issues faking 3D
- tighter bass: corrected high output impedance
- greater extension: on the lows, again corrected output impedance
- more clarity: phase issues, as 3D is sometimes experienced as greater clarity (thinking of recent review I read here)
- more clarity: higher treble due to high out-put impedance
- wider dynamic range: noisy source corrected by whacking up the volume to max (in the quite common case where the noise itself doesn't increase with increased volume)
- Greater detail: as with 6, Windows XP bit stripping for volume control; where only max volume gives you all 16 bits. An amp helps here: max vol on XP, vol control on the amp
- (and one I am adding: sweeter audio: correction of subtle distortion when using weak amp with high current phones)
- (more bass: boomy 100Hz bass boost from high output impedance)
- (less bass: high output impedance attenuating an overly bassy phone)
- (in-the-audience effect, that some prefer, on tracks mastered with strong spatialisation: poor channel separation)
These are all real effects. Some caused by poor amps, some caused by decent amps helping to correct source flaws (ie. older ipods/iphones and galaxy sII high output impedance). Some caused by poor amps attenuating or boosting exagerrated or inadequate phones.
A good amp with a good source should be adding nothing but volume. Ie. clip+ and FIIO 5.
Here's how deep the rabbit hole goes: when I connected my ODAC to my laptop I had expected to be able to use digital volume control on the laptop as since it has 24bit it is immune to XP's bit stripping issue. However, I discovered some time later that Windows selected 16bit/44Khz by default!!!!!! Obviously I changed the setting immediately. Had a reviewer had max volume when testing one DAC and much less volume on testing the ODAC: the ODAC would have been objectively producing poorer audio, ***!!!!!!!!