01-11-2016, 05:57 AM
(This post was last modified: 01-11-2016, 04:52 PM by Raccoon Sam.)
Puggsoy is right. There are a few essentials I can give though:
![[Image: 9iArkjG.jpg]](http://i.imgur.com/9iArkjG.jpg)
This is Ange Albertini's visual guide on how WAVs work. It's a simple representation on how bytes, essentially the ones and zeros, are read by a typical audio editor. Just like something as meaningless as #FF0000 can mean "Red" in the right context, those bits in that specific order can mean a certain amplitude modulation. If the above image doesn't make sense, you should familiarise yourself on how files in general work. Open a text file in a Hex editor and figure out why exactly "48 45 4C 4C 4F" means "HELLO" and how does a typical text editor know that.
Then, open a BMP file in a hex editor and figure out how exactly those seemingly random bytes transform into an image when opened in Paint.
![[Image: S9SLLMb.jpg]](http://i.imgur.com/S9SLLMb.jpg)
^ Ange made a guide on that too.
Once you've thoroughly understood the concept behind files and how they're all simply nothing but ones and zeros, just interpreted differently, you can begin reading file format specifications. The Internet is full of them and once you understand a file format specification, you have the power to read files with your mind simply by looking at the file's bytes.
Even so, most files that have a specification available have viewers/editors made for them.
I strongly suggest reading the simplest guides by Ange. They help visualise the files very well.
VGMStream is a fantastic tool that lets you convert many many proprietary Video Game audio formats to simpler formats. If you download its source code, you can, technically, read their specifications. This will take time and effort, but in the end you might understand how a .xyz works.
Now rewind back to the WAV guide. What are the core components of a sound? Sample rate, amplitude, all that jazz? Your job is to figure out how the unknown file format's seemingly random bytes translate to those attributes. It's not easy, but it's the only way. You can try importing a file to Audacity as Raw data and understanding how it sounds somewhat reasonable but still a bit glitchy if you set the sample rate to 44100... and make a connection on how all your crazy files begin with "AC 44" (which is 44100 in decimal).
You could also notice how your files begin with the bytes 50 4B, which just so happens to be the ZIP Magic number. Therefore, the files are ZIP-compressed.
You could also view the file through a spectrogram and figure something out. Or, by deduction, realise that the song is not waveform data at all, but sequenced music (think MIDI).
There's no one-size-fits-all solution, but "figuring out how formats work" is the same job for all - Audio, Video, Images... you have to dwell inside the bytes and understand their language. You can help this dwelling with helper tools, like Tile Molester for graphics, Audacity for Audio, binwalk for random craziness... Whatever you do, the feeling you get when you crack a format is amazing.
EDIT: obviously emulators with audio debuggers can help a lot too. Some save state formats might dump the uncompressed audio session (or a part of it) to the file too. If a game has compressed graphics, usually you can make a save state and open that in a graphics editor to find the pure decompressed sprites so I'm led to believe the save states might have snippets of true audio in there too. If you can extract a piece of this uncompressed audio, you can compare it with the compressed audio, make notes, figure something out and maybe write a specification of the format even though a player/editor wouldn't exist yet!
This is Ange Albertini's visual guide on how WAVs work. It's a simple representation on how bytes, essentially the ones and zeros, are read by a typical audio editor. Just like something as meaningless as #FF0000 can mean "Red" in the right context, those bits in that specific order can mean a certain amplitude modulation. If the above image doesn't make sense, you should familiarise yourself on how files in general work. Open a text file in a Hex editor and figure out why exactly "48 45 4C 4C 4F" means "HELLO" and how does a typical text editor know that.
Then, open a BMP file in a hex editor and figure out how exactly those seemingly random bytes transform into an image when opened in Paint.
^ Ange made a guide on that too.
Once you've thoroughly understood the concept behind files and how they're all simply nothing but ones and zeros, just interpreted differently, you can begin reading file format specifications. The Internet is full of them and once you understand a file format specification, you have the power to read files with your mind simply by looking at the file's bytes.
Even so, most files that have a specification available have viewers/editors made for them.
I strongly suggest reading the simplest guides by Ange. They help visualise the files very well.
VGMStream is a fantastic tool that lets you convert many many proprietary Video Game audio formats to simpler formats. If you download its source code, you can, technically, read their specifications. This will take time and effort, but in the end you might understand how a .xyz works.
Now rewind back to the WAV guide. What are the core components of a sound? Sample rate, amplitude, all that jazz? Your job is to figure out how the unknown file format's seemingly random bytes translate to those attributes. It's not easy, but it's the only way. You can try importing a file to Audacity as Raw data and understanding how it sounds somewhat reasonable but still a bit glitchy if you set the sample rate to 44100... and make a connection on how all your crazy files begin with "AC 44" (which is 44100 in decimal).
You could also notice how your files begin with the bytes 50 4B, which just so happens to be the ZIP Magic number. Therefore, the files are ZIP-compressed.
You could also view the file through a spectrogram and figure something out. Or, by deduction, realise that the song is not waveform data at all, but sequenced music (think MIDI).
There's no one-size-fits-all solution, but "figuring out how formats work" is the same job for all - Audio, Video, Images... you have to dwell inside the bytes and understand their language. You can help this dwelling with helper tools, like Tile Molester for graphics, Audacity for Audio, binwalk for random craziness... Whatever you do, the feeling you get when you crack a format is amazing.
EDIT: obviously emulators with audio debuggers can help a lot too. Some save state formats might dump the uncompressed audio session (or a part of it) to the file too. If a game has compressed graphics, usually you can make a save state and open that in a graphics editor to find the pure decompressed sprites so I'm led to believe the save states might have snippets of true audio in there too. If you can extract a piece of this uncompressed audio, you can compare it with the compressed audio, make notes, figure something out and maybe write a specification of the format even though a player/editor wouldn't exist yet!