[bobbooo]

 
I think you are making mountains out of molehills here.  Why don't you try this.  Make a track of some of your music.  You can do this with free sound editor software.  Or if you have analog RCA plugs somewhere in the circuit path, just unplug one channel.
 
Play only one channel and do not play the other channel.  Put only the silent channel of the headphone to your ear.  See if you can hear anything.  You'll be listening directly to crosstalk.  If you can barely hear anything at all it simply isn't a problem.  Obviously if you hear nothing it is not a problem. 

 
Isn't the point of this sound science section (and all science) to hypothesise there's a mountain, and then either prove, using theoretical calculations or practical experimentation (or preferably both), that it is a mountain, or disprove it's a mountain and prove it is in fact a molehill? I have actually already done a little experiment similar to what you've suggested, and I could hear crosstalk from the 'silent' channel, even on my DAP's lowest volume setting, so I think it is audible in real-world listening. Of course this isn't a great test as some of what I heard could have been natural acoustic crosstalk (as opposed to the electrical crosstalk I'm testing for) through airborne sound leakage from the non-silent driver, but I tried to eliminate this by err...wrapping it in a duvet. It’s also not clear how this test translates to the audibility of crosstalk when actually listening to stereo music through both channels. An ABX test would probably be needed for that, but as I don’t have a ‘reference’ audio chain with lower crosstalk, I can’t do this. 

 

[spruce music]

   
Isn't the point of this sound science section (and all science) to hypothesise there's a mountain, and then either prove, using theoretical calculations or practical experimentation (or preferably both), that it is a mountain, or disprove it's a mountain and prove it is in fact a molehill? I have actually already done a little experiment similar to what you've suggested, and I could hear crosstalk from the 'silent' channel, even on my DAP's lowest volume setting, so I think it is audible in real-world listening. Of course this isn't a great test as some of what I heard could have been natural acoustic crosstalk (as opposed to the electrical crosstalk I'm testing for) through airborne sound leakage from the non-silent driver, but I tried to eliminate this by err...wrapping it in a duvet. It’s also not clear how this test translates to the audibility of crosstalk when actually listening to stereo music through both channels. An ABX test would probably be needed for that, but as I don’t have a ‘reference’ audio chain with lower crosstalk, I can’t do this. 

Well there is the bit about explanations should be as simple as possible.  And in this case it isn't science for science's sake though if you are just curious nothing wrong with that.  So a simple empirical test like I suggested is within the realm of science.  As are your questions, hypothesis.
 
Now here is another suggestion to work with your current limitations (no reference system without cross-talk).  It is what I did to find out somewhere around 40-45db is enough.  In Audacity I put a single channel of music.  I duplicated that.  Converted it to stereo.  So initially you hear this as mono in your headphones.  I split the stereo channels though they retained right and left status.  Then you can mute each channel as you wish.  So then I reduced level in one channel in -10 db steps.  I played it and during playback muted the reduced channel then unmuted.  Did I hear a change when I muted the reduced channel?  If yes I reduced it some more and tried again.  Eventually when the low level channel was muted I could hear no difference.
 
Now this is extremely one sided.  Not a normal signal in music.  In fact with music playing in both channels I can be pretty sure less than 45 db channel separation would actually not be heard.  I say this because if I listen only to the reduced level channel I can hear it alone down to about -60 db.  Yet when it is at -45 db and the other channel is playing it sounds the same whether I mute this channel or not.  We know masking is in effect and when more normal music with more loudness though good separation is playing in both ears the crosstalk can be higher and still not heard.
 
So even though you hear music in one channel without the other playing, it doesn't mean it will be enough to be heard in the context of stereo.  That is why in my earlier description I said silent or barely heard.  Now I am not trying to shut you down on doing measurements and experiments.  Was simply trying to see if crosstalk levels are enough to keep being bothered about on your part.