If under controlled blind conditions I would wager that you would not recognise "what you enjoy" about one dac over another you supposedly didnt.
As for the new, multibit no name DAC, by using musical samples of my selection then doing A-B, I'm 100% positive I could tell them apart in a blind test.
I would like if I could set valid, proper blind tests more easily, and preferrably alone myself (impossible, right?), because I don't like to steal else's time. However it takes some effort, like measuring volume from the output to set up identical loudness levels. Having some handy console for switching gear for the A-B would be helpful as well - I don't have such.
But the thing is...none of it is really important to me. I do trust that my own non-blind hearing is good enough for the purpose. After all, it's the way how I will be listening to my system in the end. If I don't enjoy the sound I won't be able to fool myself for a longer period of time. Also, if I call myself an audiophile, I should be able to identify and discern the points of dissatisfaction in the sound of my system, as well as specific difference between gear, when listened in my system. I really don't think I need a blind test for this as I really can describe the difference, unless it's so subtle that it's pretty much irrelevant whether it actually exists or not.
One more thing: my impressions are created through days and weeks of listening. It's not easy at all to set up the blind test that should last days and weeks. How do I set up such? I can be fooled by other things different to bias eliminated by a blind test, whether they are of subjective nature (how I feel at the moment of listening and the dynamics of it), or of objective nature (any possible fluctuation of a sound at the system as a whole, whether because of temperature change/instability or because of AC power quality fluctuation (voltage and noise from the AC network). So, only a long term listening experience should be able to minimize the effect of those. Blind test alone as being done in a realistic environment is simply not enough (not to mention that some blind test setups might be designed with a flaw - I can give an example in my next post).