Two questions: I tend to find these by hex values or offsets, is the offset always the same or it is a different offset every time? and is it possible your script could be modified to find and split and or get these to convert? Hope i worded that right.
It didn't work quite well on other banks. I got a 27MB tail from another bank. Trimming off the first few blocks only yielded me 4 voice lines.
I can rewrite the script to slice the tail. The problem is simply slicing off each data chunk won't work.
So far I've found that the binka header is necessary for mss to recognize the file. However it would not work if I snap the header onto any individual data chunk - the conversion result is a 44-byte empty wav file.
Unless I know how the data chunks are grouped/organized I'm afraid slicing is pretty much useless.
E: for your first question - do you mean whether the data chunks have same size? No, their size apparently varies.
E2: According to this wav header struct https://www.isip.piconepress.com/projec ... 1_p05.html
, 44 bytes is exactly the header length. That means it's a valid wav file but surely no audio data is inside.
E3: I'll still write a script for doing that anyway...
E4: I'm also doing a rewrite of the BinkA2Wav converter. When that's done it'll be possible to convert sound data directly in memory without having to save it to disk. i promise this is my last edit