UbiSoft Audio Stream Decoder

[pietastesgood]

S'okay. How would we go about to parse these headers?

[Zench]

S'okay. How would we go about to parse these headers?

There's a header file floating around that contains some of the structures used. That would give us a big clue.

[OrangeC]

I dont have the game so maybe pie guy can supply you with those.

[pietastesgood]

The contents of this post was deleted because of possible forum rules violation.

[Gocha]

Hi peaple,

Zench had updated Decoder version to 0.63, see the post #1

But I am very sad it could not encode back files. For example it looses the full price for me as I can't encode files back to make changes in game.

I hope one day that would be possible

[OrangeC]

Would be very cool to add xma stream support for the gui.

[Zench]

But I am very sad it could not encode back files. For example it looses the full price for me as I can't encode files back to make changes in game.

I hope one day that would be possible

When I said that, I spoke too soon.

While we can't encode the format of SC1 voice files, we can (very easily) encode a different format that works just as well! Modify the MAPS.LM0 file a little bit, and it works!

Sorry that I didn't think of it before.

Would be very cool to add xma stream support for the gui.

It would be nice. I'll think about it.

[OrangeC]

Any luck with those header xma parsings?

[Zench]

Any luck with those header xma parsings?

Oh, oops. Totally forgot.

[OrangeC]

haha no problem man. if you need any samples let me know i have creed 1 also.

But i think ubisoft xma streams are all the same with the 08 offset thing.

[hcs]

pie brought this to my attention, I've written something to handle the Prince of Persia streams. The files it produces have plain XMA after some header info.

The trouble I'm running into is that Maki + Elika is dumping 0x31 too few bytes of the Common_BAO_0x00440f00 file I'm using for testing. It looks like it should start dumping the file 0x31 bytes later. I'll give an example here of pointing my utility directly at the file within Data360_StreamedSounds.forge.

Code: Select all

PoP360XMA Data360_StreamedSounds.forge 0x1BAA9B8

The offset is optional, if the files were extracted properly you could exclude it, but here we're pointing it at a file starting at 0x1BAA9B8 in the container. PoP360XMA detects the end of stream properly.

It produces the following files:

Code: Select all

0: Data360_StreamedSounds.forge_000 3975784 bytes
1: Data360_StreamedSounds.forge_001 3977864 bytes
2: Data360_StreamedSounds.forge_002 3803684 bytes
3: Data360_StreamedSounds.forge_003 4201036 bytes
4: Data360_StreamedSounds.forge_004 3582464 bytes

Data360_StreamedSounds.forge_000 has the stream at 0x268, 001 at 0x288, 002 at 0x224, 003 at 0x24C, 004 at 0x200. xma_parse 0.6 can handle all these (0.5 was tripped up by 003) but you don't actually have to parse the stream to get it, it's just the rest of the file.
PoP360XMA 0.1 and xma_parse 0.6 on my page http://hcs64.com/vgm_ripping.html

[pietastesgood]

Nice HCS!! Great work!

[OrangeC]

Thanks HCS.

[Kataah]

Be sure that you look inside the the other sound container too - iam not sure if this game use the same sound system like Assassin's Creed 1 PC - had no time to look inside this game yet but if yes then you will find sound files and maybe headers for the main music inside the other container too.

great work hcs

[AlphaTwentyThree]

The contents of this post was deleted because of possible forum rules violation.