No speaker is ever better than a decent amplifier. Frequency response of even the best speakers is at best within +/- 2dB for the central part of the spectrum. Good amplifiers stay within 0.2 dB, or even far less. Distortion figures are a similar story.
The explanation is very simple: speakers have mass that can resonate, is slow, behaves in a non linear way etc. Good amplifiers have been sonically perfect for decades, and cannot be distinguished from each other in a double blind test.
See here for the classic and legendary test:
http://www.keith-snook.info/wireless-world-magazine/Wireless-World-1978/Valves%20versus%20Transistor...And here for a hilarious one:
http://matrixhifi.com/ENG_contenedor_ppec.htmIf you google you will find quite a few more.
I personally attended a private demonstration by Quad's designer Peter Walker. He had this set up where you could listen blind to his three famous amplifier designs (using studio master tapes). I thought I could hear a difference, but he just grinned: I had been no better than random. That was a lesson learned.
There are a few proviso's and I have listed those in my earlier post. If those apply, there will be sonic differences because one of the amplifiers does not meet the criterium of 'properly designed and used within its specification'. This is why many tube amplifiers fail the test.
If that is not the case, the perceived sonic differences are the product of level differences (the brain interprets louder as better). For a listening test levels should be kept within 0.2 dB, and that is such a small difference that you can only achieve this with a decent Volt meter (and not with an SPL meter, let alone with a smartphone app). Did you ever see a dealer equalizing levels in the demo room using a Volt meter?