Read on Twitter
Something to think about: in software, we're used to this broad idea of "copyright" where anything related to the original code is liable to be a "derivative work". For example, the idea of clean-room reverse engineering comes from here.
So, we say, theoretically you can read some code, write some similar code, and commit copyright infringement. Hence the legal defense of clean-room RE.
But copyright isn't actually a magic viral concept like this. It covers very specific things.
But copyright isn't actually a magic viral concept like this. It covers very specific things.
One example from the music industry is Taylor Swift, who is currently reverse engineering her own old masters, to which she does not own the rights (just the composition copyright), and re-creating *as identical as possible* versions, to which she will.
At *best* all that stands between the original versions and her re-creations is a human ear trying to remake them as closely as possible. But, for all we know, it could be more technical: using spectrograms to match EQ curves and such.
(This is, incidentally, a hobby of mine)
(This is, incidentally, a hobby of mine)
So you can "hand-decompile" a song like that and be in the clear as long as you don't copy any audio directly. Cloning all the details of the original recording (mix, instrumentation, mastering, etc) is OK; it is not a derivative work of the recording, only of the composition.
How come we assume that reverse engineering some binary and writing some code that is somewhat similar would be copyright infringement? Obviously straight copy and paste from a decompiler would be a derivative work... but why do we obsess over a *human* accidentally "copying"?
(FWIW, in the above video, only the voices were ripped from the original, the rest is re-created from scratch. Also, all the synths are instances of Helm.)
