How about something with a voice recorder in it and an easy to get to replay button? You can hold the 'bear' in one hand with a finger able to get to the button and when you squeeze the bear's paw with one hand (stage magic misdirection, they watch the hand that moves to apparent purpose) you activate the playback with the other. What else... batteries and a speaker if the voice recorder only has a headphone jack.
Then you don't have to do mpeg or any of that and save months of frustration and a lot of money.