The question is as the question is. It won't be this isolated sound though, it'll be a sound on top of background music. I think I can get a hold of the sound file itself from the data, but at worst I'd just make my own recording and ask for a less accurate match - if variation is possible in whatever method may be proposed.
Theoretically you could separate your desired sound by ICA method or other blind source separation methods from the background music, then design a matched filter by your separated sound. The matched filter have high response when the desired pattern (which is your desired sound) exist in the input even when it buried in noise (background music).
The matched filter has acceptable performance but i'm not sure about the ICA and other source separation methods and highly depend on your input signals (desired and background sound).