DVD-HD, Blu-Ray, and pro audio is 24 bits @ 192 ks/s. This is 4.608
Mbits/s.
So, this could be converted to a 4.608 MHz pulse density modulated
signal (like a SACD) with no loss of information. Or you could use half
that frequency and a two level pulse train (-1, 0, +1) -- which is
supposed to be better.
This is pulse density modulation, not pulse width modulation.
This is basically an RF signal so it isn't practical to run the output
stage at that frequency. The Sharp 1 Bit stuff does but it is very
expensive.
You can multiply this pulse train by (1+Z^-1) and greatly reduce the
high frequency content but this would loose information.
It might be possible that analog feedback of the digital signal could
make this work.
IIUC, pulse width modulation is another question. PWM with that clock
frequency has much less information. The generally accepted method
seems to be to try to compensate for this by dithering and applying
noise shaping to the dithering.
With clocked PWM there seems to be a somewhat strange problem here. The
limit is on the clock frequency so there is a direct relationship
between the frequency of the PWM output signal and the number of
different pulse widths.
Most of the designs that I see are based on using a PWM infrequency of
more than 48 KHz. I haven't seen anything showing an analysis to show
that using a higher PWM frequency with the same clock will improve the
output.
--
JRT
_______________________________________________
Open-hardware mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-hardware
