Do you need a seperate volume level for each file? Won't one do? If the files have different volumes, you could normalize them.
How do you normalise the sound level in a collection of MPEG files which may have either AAC audio or MP2 audio in them?
Anyway, isn't it easier to create a playlist (essentially a little database) of media object names and their volume level. This way when the playlist is running un-attended, you avoid loud, soft, loud, soft, loud, loud, soft. And you dont need to re-process your audio files which will also serve to degrade their quality and faithfulness to the original further.
I actually cannot believe that no-one has created a playlist format that supports volume level. For some odd reason, they seem to have a fixation on duration parameters, which are almost (not entirely) pointless. A Volume level based on an index that the user can determine is a VERY useful feature.