On Tue, May 19, 2009 at 11:10:00AM -1000, Joel Roth wrote:
> I understand that the formal way of going about that to
Studying the way some software emulates a system does not
necessarily provide correct information on what is being
emulated, unless it is a complete and accurate physical
model, which is unlikely.
You may be able to find out some interesting parameter
values but there it ends.
For example the way Aeolus generates an organ pipe sound
(additive synthesis and some postprocessing) has no
relation at all to what happens in a real organ pipe
(which is a complex non-linear and sometimes partially
chaotic combination of an oscillatir and filter that
are tightly coupled and influence each other).
As an author of such software you make choices on how
to do certain things, for all sort of reasons ranging
from conceptual to very pragmatic. The latter may
obfuscate things considerably, and provide the author's
signature in a sense.
If someone would write an organ synth by re-implementing
the processing in Aeolus in entirely new code I would
probably have no problem in detecting that quite soon,
given the new code.
Io lo dico sempre: l'Italia è troppo stretta e lunga.
Linux-audio-user mailing list