To accomplish what you’re looking for we would need to have channel virtualization, that is to say processing of the sound even when it isn’t audible. Our current audio system relies on the audio hardware to keep the sounds playing and as such, when the hardware isn’t processing it, it stops and when it picks up again, it starts at the beginning.
We certainly hope to support channel virtualization in the new audio system, but we don’t have any timelines on when that will be available, even to preview.
In the meantime, your only option is to do something with that sound such that when you’re out of range of it is actually at 0.00001 volume and still playing, but that can become an issue if you have many such sounds as the hardware can only support vocalizing so many at a time.
Apologies for the inconvenience.