all fields are 32-bit words. Count is at offset 4. Table starts from 0x1C Table fields: hash number of file header offset (multiply by 0x10) data offset (multiply by 0x80) header size data size zero (or unknown) You don't actually need headers, because sample freq and channel number already containe...

But if you dont know what PathMusic or PathMusicTrack or PathMusicMap or .mpf or .mus files are, you probably dont know how to unpack this data. Don't underestimate me ;) If you check my profile, you can see I'm a reverse engineering specialist. That was very easy. This file is a sound bank contain...

Good that you got it. You see, in "mp3" mode ealayer3 doesn't even decode anything, it just puts the parts of data to standard mp3. But if you know how to decode the "path music" large blobs from RA3/Uprising/C&C4, I would love to know how... Can you send me example of such f...

I doubt you have seen every EALayer3 encoded file that has ever been produced... Yes, but statistically, after I've seen 1000s of files, chances I've seen all types are closing to 99%. Ok, you can find MPEG2 in low quality SFXs like explosions or gunshots, usually in sound banks, for example in Dea...

jfwfreo wrote:I have already identified at least one case where there is data being read by the official code but not by the ealayer3

Can you send me that file? You know, I just can't believe that can be. I converted dozens of games, and it never happened.

so it changed

whats the size of TextureFileCacheManifest.bin ?

Since all sounds from BF3 can be decoded properly without that flag, then I'd say this flag is not needed.

p.s. I compiled his code in .net, which required some changes. His original source was for usual c++. It also required some tricks to make it compile, but unfortunately I don't remember now what it was. I remember I found some clues googling about his tools. It was somewhere on someones blog or some...

I don't think you need debug symbols to figure out the header. I mean you don't need a debugger to tell that 1 is mono ans 2 is stereo. So if I remember correctly, changes I made were: 1. stop after end of the stream. useful for some games, but not for others 2. continue after sync lost. May be usef...

Either you're doing something wrong, or they changed file format.

Why do you even need that? Ealayer3 decoder is working fine and had no problems for many years. All changes needed were only about some header flags, and don't have any connection to the decoding process itself.

Since this kind of storing quaternions was described in some old 3d-modelling book, its highly possible that they do it like everybody else:

axis == 3 ----> q = [x, y, z, w]
axis == 2 ----> q = [x, y, w, z]
axis == 1 ----> q = [x, w, y, z]
axis == 0 ----> q = [w, x, y, z]

good that you made it!

I have no idea what kind of script is this, it would be better if you just tell me what data you getting is wrong. Hierarchy? Positions/rotations?