Trying to rip segmented .mib music from "Wallace and Gromit"

[jsboy]

I'm trying to rip the music from the PS2 disc of "Wallace and Gromit in Project Zoo".

I've got the original .mib-files from the disc on my HDD and tried to convert it to .wav .

However, it creates .wav-files that are segmented, which means that that it contains all 3 or 4 different versions of a track in one file.

The problem is that it plays the first 0.5s of version 1, then the first 0.5s of version 2, then the first 0.5s of version 3, then 0.5-1.0s of version 1 and so on.

How am I able to rip such a specific kind of video game music perfectly (maybe with HEX editing?)?
Any help is really appreciated!

Here you have some links.
To the original .MIB file
To the .wav file

[Raccoon Sam]

It appears that.... The wav is sliced at intervals of 28112 samples. First there's 28112 samples for the first part's left channel, then another 28112 chunk for the first part's right channel. Then the drum track(?) the same way, 28112 left, 28112 right. Then the second part, left and right. Drums again, left, right... Repeat this 1568 times (1568 * 28112 = 44079616, which is, according to vgmstream, the total amount of samples) and get re-ordering. If it helps to visualise, That WAV is a monotrack:

/ L: 1 2 3 4 5 6 7 8 ...
\ R: -

But you want it to be two stereo tracks:

/L: 1 5 ...
|R: 2 6 ...
|L: 3 7 ...
\R: 4 8 ...

I don't know if there's a method to automatise this. Maybe an audio editor that supports scripting?

EDIT: Apparently Adobe Audition can "record favorites."
Favorites let you record one or more operations in the Waveform Editor as a single command that you can access from the Favorites menu or run on multiple files as a part of a batch process.

I guess... that if you recorded a favorite that cuts a 28112 sample-long clip, pastes it below the current track and does it again three times with different tracks and saved THAT as a keyboard shortcut, you could press that hotkey 392 times and have your neatly tidied multi-track. Sounds ass-backwards but that's all I got sorry.

[jsboy]

Thanks for your reply!
How did you calculate the exact interval length?

[Raccoon Sam]

Well, it was a given that 44079616 was the total amount, because vgmstream said so. I zoomed in super super close to the waveform to the point where I thought I heard the first 'snap'.

I tested it out by calculating 44079616 / 28112. It was a nice round 1568. Had it been any more or less, I would've gotten 1568.11156172 or 1567.88845415 or whatever.

I tried selecting another range after that, from 28112 to 56224 and saw a similar snap. Again for the third time, from 56224 to 84336 and so on.

EDIT: After a stupidly painstaking session with Audition, TextMate, ffmpeg and Audacity I managed to do a workflow that let me get all the 1568 files and arrange them as I had envisioned.
MP3 link
WAV link

I loaded the original WAV to Audition, set up a marker at 0 and 28112 and 56224 and closed it.
Then I opened the Audition's .sesx file in TextMate and found the XML strings <marker startTime:0/length:28112> etc.
I then recorded a TextMate Macro that writes 1568 markers, each with the length of 28112 and with the start time of n+28112.
I then saved and re-opened Audition, and saw my markers. I exported them all as separate files.
I THEN made four folders, "A", "B", "C" and "D" and wrote a script in terminal that moved 0.wav, 4.wav, 8.wav etc. to "A" and 1.wav, 5.wav, 9.wav etc. to "B" and so on.
Then I used ffmpeg's command line interface to concatenate the files in each folder to one.
Then I had four WAVs. Which I imported to Audacity and set up the left/right channels.

But hey it works lol. There's just gotta be a better way to do this.