Double blind testing certainly has its place. I have found it useful sometimes when I'm asking my wife to tell me which of two variations she prefers, for example when I'm testing the speakers with and without filters or two drivers compared with four, etc.
For double blind testing to be effective I have to ensure as far as possible that only one 'variable' is being changed at a time so the volume level must always be equalized (for example).
I can think of three main limitations to this sort of testing in the field of hi-fi:
Firstly, when switching from one speaker to another you can't switch off the working of the brain (although clearly on this forum many try very hard indeed

). Let's say we have two speakers, one neutral and flat, the other coloured in its frequency response. If we listen to the coloured speaker for some time our 'ears' get used to its colourations and it sounds normal to us. If we then switch to the neutral speaker it will sound coloured with a 'reverse impression' of the colourations of the first speaker. Just like staring at a black dot on a sheet of paper for a while then seeing a white dot when we stop. So, the compensation of the human brain is the first thing that makes double blind testing difficult.
The second difficulty is that many pieces of equipment need to be warmed-up, run-in or whatever before they give of their best.
Thirdly, and perhaps most importantly, we can't always come to an instant decision in this field. Personally, I have to live with a piece of equipemnt for some time before I get a feel for its strengths and weaknesses and know whether I want it in my system.
So double blind testing has its place but, in the final analysis, normal listening to music rules.