On Wed, Nov 11, 2009 at 10:59:39PM +0000, Folderol wrote: > An entire cycle of 48kHz is about 20 uS so jitter would have to be > significantly less than that to avoid 'cogging'. It's not the amplitude of the jitter that matters but the spectrum. Even the best and most expensive audio cards have lots of very low frequency jitter. And if you lock them to an external reference the local VLF jitter is replaced by that of the reference (within the PLL bandwidth). Jitter creates phase modulation, and the amplitude of the sidebands increases with signal frequency. To have a clean top end any non-VLF jitter should be below a nanosecond or so. It's not so difficult at all to achieve this with a PLL provided the reference is stable. Syncing two or more cards (so they will produce time-aligned samples if given the same analog input) is another matter. But if the 'soundcard' is dedicated hardware there are relatively simple solutions for that. Ciao, -- FA Io lo dico sempre: l'Italia è troppo stretta e lunga. _______________________________________________ Linux-audio-user mailing list Linux-audio-user@xxxxxxxxxxxxxxxxxxxx http://lists.linuxaudio.org/mailman/listinfo/linux-audio-user
