It’s hard to say for me. It sounds like an echo. If the song would be played twice at nearly the same time it probably could lead to this effect.
I guess chatty wouldn’t emit feedback if the phone is locked? Would that be the explanation why it only happens with the screen unlocked?
I could do a recording of a few seconds if that’d help to judge the situation. Or is there a way to see that two processes are playing a sound if the screen is unlocked?