I don't have a solution for the problem, but I also have been annoyed by it ever since I bought them - and have spent much more time thinking about why it might be happening than I would like. This is all I've come up with:
It can't solely be due to the fact that they're wireless because it doesn't happen when you answer a phone call. It also can't solely be due to the ANC (automatic noise canceling) because it doesn't happen when you plug in the wire and the ANC is still enabled. So my guess is that it has to do with decoding the high-bitrate wireless audio stream that is in use when you play music or a video. I assume that when you answer a phone call, the audio is not using that same high bitrate, so that's why the problem goes away in that case.
I have various other bluetooth earbuds/headphones that don't exhibit this problem at all, but none of those have ANC so it's not a direct comparison.