On Tue, Jun 08, 2010 at 08:53:43AM +0200, Philipp Überbacher wrote: > So the actual problem isn't the noise but its correlation with the signal?
Yes. > I'm a bit curious about the first graph. The actual signal is the ~1kHz > one, but what are all the other 'spikes'? The quantisation error appearing as distortion. This the problem that is solved by using dithering. 1 kHz at FS = 48 kHz means a period of exactly 48 samples, the same values are repeated for each period. So the quantisation error is periodic as well, with a period of 1 kHz. In other words, distortion. Adding noise randomizes the quantisation error. Ciao, -- FA O tu, che porte, correndo si ? E guerra e morte ! _______________________________________________ Linux-audio-dev mailing list [email protected] http://lists.linuxaudio.org/listinfo/linux-audio-dev
