2021-07-13, 01:01 PM
@davidferguson
I wrote some code to translate a "binary.bin" (exported from Logic 1.x via the binary export at 8 bit) into raw binary data.
The result is a file with the ending .ac3, but it doesn't play (yet?). For what it's worth, MPC-HC does open the file and shows 1 second as length, but doesn't really do any audio. ffmpeg says it recognizes some kind of ac3 but isn't able to decode it either. Mind you, I'm not 100% sure I did the file generation correctly, that might very well also be a source of error. For example it's possible that some fuckery with endianness must be done, but this should just be the bits straight as they were captured, into binary data.
Method: I took the raw data from the logic analyzer export. Whenever the block channel was active, I interpreted the data at that point as the beginning of a new byte. Then each clock cycle triggered a new reading. The value actually saved into the file was that of the data channel of course.
But what I did find:
![[Image: NjE03wN.png]](https://i.imgur.com/NjE03wN.png)
There is a repeating pattern every 512 bytes. The above image is the file opened in Photoshop as raw input. It's plausible to assume that the alignment of this picture isn't correct since I just wrote the binary data into the file exactly as it was captured.
I tried a similar thing with AC3 encoded via ffmpeg, but I wasn't able to get this nice of a repetition. Best I got was with 1950 pixels width:
![[Image: 1smHNqT.png]](https://i.imgur.com/1smHNqT.png)
I've attached the decoded "AC-3" file in a zip to this post. Maybe someone has another idea. The repeating pattern has me optimistic. I wonder if this is already the raw ac-3 data stream or if it is still capsuled in some kind of transport format for the purpose of error correction.
Edit: Here's the parsing code btw: 🔍https://github.com/TomArrow/DolbyDigitalSerialParser
I wrote some code to translate a "binary.bin" (exported from Logic 1.x via the binary export at 8 bit) into raw binary data.
The result is a file with the ending .ac3, but it doesn't play (yet?). For what it's worth, MPC-HC does open the file and shows 1 second as length, but doesn't really do any audio. ffmpeg says it recognizes some kind of ac3 but isn't able to decode it either. Mind you, I'm not 100% sure I did the file generation correctly, that might very well also be a source of error. For example it's possible that some fuckery with endianness must be done, but this should just be the bits straight as they were captured, into binary data.
Method: I took the raw data from the logic analyzer export. Whenever the block channel was active, I interpreted the data at that point as the beginning of a new byte. Then each clock cycle triggered a new reading. The value actually saved into the file was that of the data channel of course.
But what I did find:
There is a repeating pattern every 512 bytes. The above image is the file opened in Photoshop as raw input. It's plausible to assume that the alignment of this picture isn't correct since I just wrote the binary data into the file exactly as it was captured.
I tried a similar thing with AC3 encoded via ffmpeg, but I wasn't able to get this nice of a repetition. Best I got was with 1950 pixels width:
I've attached the decoded "AC-3" file in a zip to this post. Maybe someone has another idea. The repeating pattern has me optimistic. I wonder if this is already the raw ac-3 data stream or if it is still capsuled in some kind of transport format for the purpose of error correction.
Edit: Here's the parsing code btw: 🔍https://github.com/TomArrow/DolbyDigitalSerialParser