Usually the opposite. "Finesse," whatever that means, is irrelevant. Most crosstalk happens at source switches, volume controls, and coupling because of channel proximity.
Another review based on imagination. Unplug the source right input. Put some music on at a decent level. How much music do you hear coming out the right speaker? How much is that barely audible signal (if at all) going to change anything.
Because you won't hear a difference past about 18 db channel separation and 118 db of channel separation. That is how our hearing works. Has a lot to do with how much blending an off axis source has between one ear and the other. Our hearing makes use of timing at lower frequencies, and intensity at higher frequencies with a mix of the two in between. 25 db is plenty of channel separation for perception of stereo.
Sounds like a sloppy test I would expect to hear recommended by subjectivists. Even for THD+N the most lenient threshold considered acceptable by most people who evaluate audio gear based on science and measurements is -85 dB. It's completely ridiculous to me that there are still people on here expecting to be taken seriously claiming that bleeding of opposite-channel information should have no discernible effect even up to levels higher by several orders of magnitude.
Give some thought to how much separation is present in the source material in the first place; then it will not be as "ridicilous".
Well, here's your chance to make a real contribution to our understanding of stereo,
Since electronic music exists that is under no obligation to reproduce or simulate instruments as if playing in a physical space, captured with 2 microphones, leading to inherent crosstalk etc. etc., there is always an expectation that arbitrary numbers of samples per track could be 100% L-R separated, and a good reproduction system should maintain that to the very limit of human capabilities to detect.
I could, just out of curiosity, but notice also the limits of what that experiment can prove, according to the author themselves:
No better than doing my own test with my own song, with a real-world (not synthetic) cable-change based alteration. And I detected differences that tell me -20 dB is far from sufficient, for me. But it's true that I only compared vs. -40 dB, since I only had the cables I had, and the difference I was hearing could have been realized anywhere in between. So it's not impossible that I have the same true threshold as the above author, that -30(?) dB happens to be enough for me as well, and that I was actually getting no extra benefit going from there to -40.
Doesn't seem to be worth it, since all the gear I currently own is keeping me comfortably far away from -20 dB crosstalk, and if I ever think of getting any other ridiculously low-impedance headphones (or fall prey to the virus of IEMs) I will just take care to put them on a 4-wire SE cable at least (or better yet balanced), and that should have me covered. (Balanced is everywhere these days, so I don't expect a significant cost penalty for my insistence on having at most -40 dB crosstalk.)
sure, minus that pesky "human capabilities to detect".
So where in modern equipment is the crosstalk worst than -40db? This is a non issue.Since electronic music exists that is under no obligation to reproduce or simulate instruments as if playing in a physical space, captured with 2 microphones, leading to inherent crosstalk etc. etc., there is always an expectation that arbitrary numbers of samples per track could be 100% L-R separated, and a good reproduction system should maintain that to the very limit of human capabilities to detect.
I could, just out of curiosity, but notice also the limits of what that experiment can prove, according to the author themselves:
No better than doing my own test with my own song, with a real-world (not synthetic) cable-change based alteration. And I detected differences that tell me -20 dB is far from sufficient, for me. But it's true that I only compared vs. -40 dB, since I only had the cables I had, and the difference I was hearing could have been realized anywhere in between. So it's not impossible that I have the same true threshold as the above author, that -30(?) dB happens to be enough for me as well, and that I was actually getting no extra benefit going from there to -40.
I imagine it would take a lot of work to set up a test to really clarify this - I would need to see what the frequency envelope of real-world crosstalk tends to be, if by any chance it isn't flat, then I'd have to set up some way to apply crosstalk with that envelope to as many songs from my collection as I wanted, while also being able to tweak the amount of crosstalk from, say, -20 to -50 dB. And then listen and compare and listen and compare 'til I'm blue in the face.
...won't bother and continue to assert dubious stuff for which I have no evidence."
It's like I'm just talking to myself here. Didn't I just link you to the example of the Verum 1? With the stock cable those have crosstalk at -20 dB due to their very low impedance drivers, high impedance wires (1.8 ohm round-trip) and long common return wire. The voltage divider that forms there creates comical amounts of crosstalk.
Correct. My experiment, which is as subjectivity influenced and as conclusive as his, possibly more because my alteration wasn't synthetic. The work has been done and I have my conclusion, which I can oppose to anyone else's subjectivity influenced conclusion from their test of "that one arbitrary special song I happened to choose that day". Talking about that test like it's God's Standard for Separation, brought down to us on stone tablets from the mountain, is quite funny to see in a science-focused forum.
That's delusional, to put it mildly. As is your claim absent any evidence. Try using your ears and not peeking. It's easy to make two files of whatever music you like, add crosstalk to one, then ABX them. Do it. Or not, and remain the guy claiming he can levitate 20 feet in the air, but you'll just have to take my word for it.
