dcollins wrote on Sun, 30 December 2007 05:12 |
Larrchild wrote on Sat, 29 December 2007 16:10 |
Which begs the question: "If they are now locking to GPS reference, why can't we?"
All the broadcast stations do it.
|
Because they need the great long-term accuracy (like loses a second every billion years) and audio wants great short-term accuracy (like 10^-12 jitter).
|
Spot on. Of all the audio myths I've seen promoted lately the one that surprises me most is that sound quality should improve with absolute accuracy. It's like claiming the positional stability of a puppy on a leash is improved by tying the leash to a lamp post located exactly on the equator instead of another one down the road.
Guido Tent has knocked together a reference oscillator that we use to calibrate house syncs for people who need grade 1 accuracy. The reference oscillator puts out 44.1k and 48kHz within 0.01ppm. The output is phenomenally jittery, you can see it rattle on a scope. Point is, the actual oscillator is an oven oscillator at 10.000MHz. To get audio rates, a PLL is added. Since jitter doesn't matter when all you need is an exact frequency over the long run, a 74HC7046 (!!!) is used for that.
I daresay rubidium oscillators also don't come at audio rates so you'll always need to slap on a PLL or a digital synthesizer. Whether the result is any good for audio (ie. jitter performance) is entirely determined by *that* circuit.