Actually, using a crossfade of just a few seconds doesn’t feel quite as lifelike to our ears. It’s clear we’ll need a slower, more seamless, lifelike transition between these two sounds. If you’re adding background sounds to existing video footage, you’ll have a visual cue for when one sound starts to die away and another starts to appear, but when we’re doing this ‘blind’, we have to do what feels right. 

In the real world, you’d expect the busy sounds of the metro station to die away more slowly as the hustle and bustle of the street comes in and eventually takes over. For a more cinematic experience, we might create a quicker, more jarring transition between two background ambiences to emphasize the change. In this example, we’ll make things as lifelike as possible…

We extend the crossfade region to be far longer – up to about 15 seconds – and listen back. It’s more realistic, but we can also change the crossover curve between the two, as we do below with the exponential curve. It sounds more realistic to our ears.