You can kind of make this work for set top players but it will not work as desired when played on a computer.
Your buttons would have to be “AUTO ACTIVATE” button that actually jump to another menu/video track that plays the sound and then has the button in the selected state to jump to the actual video when pressed. The problems are, this is slow navigation with pausing, if you have background audio on the menu it will stop and restart after the voice over jump occurs. The biggest issue is auto activate does not work on computers so it will just sit there until the user hits enter but it get the the actual video the user has to hit enter twice, first jumps to the VO menu and then the 2nd to the video, can be very confusing.
So in the end the idea probably should be abandoned.