[LAD] Audio problem for numerical optimization
fons at linuxaudio.org
Sat Mar 1 12:33:42 UTC 2014
On Sat, Mar 01, 2014 at 11:17:50AM +0100, Ralf Mardorf wrote:
> "JAMin uses FFT, so it's a quite CPU and memory hungry beast." -
That is a very misleading statement. Jamin is a 'CPU and memory
hungry beast' not because it uses FFTs, but because it uses the
wrong algorithm in the first place (which in turn requires the
use of a high overlap factor to hide the resulting problems).
It's not a matter of inefficent coding, but of failed design.
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
More information about the Linux-audio-dev