A very large scale DBT was done at two separate DIY events about 2 years ago, comparing electrolytic to high-grade poly caps. No statistically meaningful differences were observed under the blinded conditions. Additionally, no pre-existing credible perceptual research would suggest such a difference to exist.
I know all about those test. The testing method was flawed. Here is what happened.
They got a group of people together in an environment that they were not familiar with. They put together a sound system that none of the participants had heard before. They used questionably crappy gear. They played music that many were not familiar with at all. They played a complete track all the way through, made a cap switch then played it all the way through again.
No one had any reference for what they heard and could not recall any difference because too much time had passed from start to start and they could not recall the differences.
Prior to that Dennis Murphy got the whole thing started using two identical speakers. One was using cheap poly caps. I think they might have been Solens. The other used Sonicaps. His testing method was a little different, but I don't recall the whole thing. In the end every person all claimed one to be much better then the other, but I think he might have had one undecided. Seems like there were nine guys there. The speaker that they all picked was full of the Sonicaps.
I later did a similar comparison with output coupling caps on a D/A converter. I had a switch on the back of the unit that allowed me to select between caps. One was a Sonicap, and one was an Axon poly cap. All cap values were matched.
I was able to replicate the results from both Dennis's test and the others. First I played a complete song all the way through, made the switch and played it all the way through again. Most had no clue. It asked too much of them to remember that much. They didn't know the system. I didn't even know the system. It was at someone else's house. The music was known to some, but not to others... They looked at me like there was no way.
Then I played about a 15 second intro to a song. I played it four times straight and gave everyone a little bit of time to memorize how it sounded. Then I made a switch and many nailed the differences right away. Without me saying a word the whole group was spouting off differences. We tried it again with a different song. A few seconds of it repeatedly played, then switched. After a few times through there would be a consensus and they felt like that they knew which cap was which.
After that I could repeat the process on different pieces of music playing a few seconds at a time and the whole group could then identify if it was cap "A" or cap "B" that they were listening to. They had learn which cap sound which way. Only I knew which was which. Sometimes I tried to fake them out and I really didn't switch anything, but they figured that one out pretty quick. They were learning the differences and if they did not hear them then they knew that was no change. They were nailing it every time.
After a while they noticed what the audible cues were and without even hearing them both they could identify if it was cap "A" or cap "B".
You see one was cleaner, had a quieter noise floor to it and lacked smearing compared to the other. Pretty soon all they had to do was listen for that aspect and they nailed it. A single acoustic instrument playing an intro made it easy.
Then as soon as it gets too easy they are all bored with it and don't want to play anymore.
It is all in how you conduct the test. You can claim double blind this and that all you want, but when the group gets it right 100% of the time and in just a few seconds then it is pretty conclusive for me.
