Dithering Vs digital domain crossovers?

[s3tup]

Hi all!
I've got some interesting (at least for me) question about CD mastering + digital domain crossovers.

Some CDs are mastered using dithering technology, which is used to burn on the limited-dynamics-range media a low-level information by modulating it with kind of pulse-width modulation.

In regular system, when the signal is being converted in the DAC and passes thru DAC's output filter (low pass), the dithered high-frequency "noise" goes back to the frenquencies it was intended to be (say from 44khz pulses to 1khz wave).

When we use digital domain crossover, the high-pass section takes the noise intended for midwoofer -> filters it thru low-pass filter of the tweeter's DAC, and then... tweeter plays low freqencies. It surely can't produce them with required level, as it has rolling-off frequency response = we loose the depth of soundstage, which cosnists of these low-level sounds.

Or have i missed something? << that's the interesting question

[SY]

First, ALL digital recordings use dither; it's fundamental to getting high resolution with a small sacrifice of noise floor. Second, dither is not really PWM, it's noise. Plain old noise. The LF component of the noise is part of the signal passed by the LF part of your crossover (the crossover is dumb and doesn't know the difference between the encoded music and the encoded noise) and then on to the woofer. Likewise, the HF component of the noise is passed through the HF crossover and then on to the tweeter. The only LF noise that the tweeter sees is the intrinsic noise of the crossover- for any vaguely competent unit, that noise will be minuscule, 100dB or more down.