I think natevl and doctor_i are on the right track with this post.
As a hobbyist who works with electronics, I have a theory that kind of lines up with what everyone is saying here, since I have the same problem.
On a side note, apple decided to be their elitist selves and they changed the layout of TRRS.
Pretty much every other manufacturer uses TRRS as Left, Right, Mic/Control, GND
Apple uses TRRS as Left, Right, GND, Mic/Control
Back on track
The control feature of apple headphones makes a direct connection between the GND and the Mic/control to perform "xyz" feature.
iPhones and other apple products seemed to have a way of telling if regular headphones are plugged in (ground and mic ports being shorted out, or at equal electrical potential), this one does not.
When the volume hits a certain threshold, the ipod mistakes the higher return to ground as a button press. This could be easily fixed by a firmware update that checks to see if the gnd and mic connections are shorted. In the meantime, you can't use a third party cord at 100% volume, because as soon as any sound goes above the threshold of 100% mic, it's read as a button push.
or : as doctor_i suggested:
What apple could do at a firmware level:
1. Decrease sensitivity to removal of the connection
2. Add a toggle in settings to turn this feature off
3. Increase the amount of time after sensing removal before the device auto-pauses, to determine if that was an erroneaous detection.
4. Detect whether or not mic and gnd are shorted and disable control through headphones if they are, since this is the primary source of the problem
Hey apple, fix this problem please!