In a recording, people try to pronounce things "properly", with better articulation than in a regular conversation with other native speakers. So in practice you'll often hear that "j" almost drop...