I'm basically in the opposite boat. I'm new to file formats and its hard for me to understand them. On the other hand, I'm really good at coding. If you are able to explain how the format works and how it's structured, I can turn that into a program for you. If there is any other coding help you need to get to a level where you understand the files, I can also help with that.
This is the best I was able to find about bink audio in general, I don't know if the same thing applies to .binka https://wiki.multimedia.cx/index.php/Bink_Audio
Other than that all the other information I have is in the first post.
So for the last weeks I looked into MSS and TF|2's binaries. It seems like...
(1) MSS10 does not provide an audio decoder API like previous MSS versions did. (?)
(2) TF|2 engine creates a MSS queue at startup then, when a Source script requests to play an audio, adds some sort of "play event" to the queue for MSS to handle it.
Fascinating on number 2. However number 1 confuses the hell out of me too, if there is no decoder how is it decoding the audio?