Ensuring the waveform remains intact even if the file is corrupted
Digital storage media is extremely vulnerable to the environment. Just look at SSDs: if left unplugged for a year or so, most files will be corrupted and unreadable, or if they're read, you'll just get meaningless noise. Even CDs that are supposed to last 100 years will pop at the slightest scratch.
To prevent this from happening, Fourier Analogue-in-Digital applies ECC very tightly and uses a discrete cosine transform instead of the fragile Fourier transform, which can ignore a small amount of error and still recover the original information.
Preserving the input from the microphone as much as possible
Avoiding PCM as it is means that 'perfect losslessness is impossible'. But we still aim to create the ultimate container for the sounds that touch people's hearts without using PCM, and with dedicated analogue domain DCT hardware, we aim for perfect archiving, allowing us to capture everything from the tremble of a single air molecule to the sound of crickets, clean hi-hats and heavy kicks, ear-ripping gunfire, the sound of a rocket being launched, and even the first sounds of the universe and store everything in this single format.
Making all data frames work independently
It would be terrible if one of the data frames became corrupted and none of the data that preceded or followed it could be read. To avoid this, all frames are designed to be independent, so that even if one of them becomes corrupted beyond any repair, it cannot affect the next frame.
Compressing
Compression makes data extremely vulnerable to damage. This makes it difficult to achieve FrAD's goal of being an archive, so FrAD has decided to give it up. There are plenty of good formats like FLAC where capacity is important...but most of them don't support floating point, which is sad.
DRM?
Aww man