Arny Krueger wrote:
Carey Carlan wrote:
(Scott Dorsey) wrote in
:
We have successfully shot S-PDIF from one building to the
next in downtown Baltimore just by pointing the things out
the windows. Channel reliability wasn't the best and I'd
worry a lot about rain fade, but it worked better than I'd
expected and it allowed us to get good quality audio around.
How did you convert spdif to a video signal?
SP/DIF and video are very similar signals. Both are 1 volt
p-p @ 75 ohms. In theory 24/96 SP/DIF has somewhat greater
bandwidth requirements than NTSC video, but 24/44 is far
closer at around 6-8 MHz.
Indeed, if do a little hostorical research, you'll discover that
the choice of 44.1 kHz sample rate was dictated by the fact that
the only large-capacity storage devices for digital audio at the
time (late '70s) were video recorders. The sample rate was chosen
such that an integral number of frames could be stored on each
scan line.
For example, take NTSC video with its 525 lines and 60 Hz field
rate. Assume 35 blanked lines, that leaves 490 lines per frame,
or 245 lines pere field. There's plenty of bandwidth to fit 3
samples per scan line, so:
60 fld/s * 245 lin/fls * 3 sample/lin = 44,100 samples/s
For 625 lines @50 Hz, you have 37 blanked lines, levaing 588
lines/frame or 294 per field, 3 samples per line, and you get
50 fld/s * 294 lin/fls * 3 sample/lin = 44,100 samples/s
Clever, eh?
The point being that, yes, while it can be said that digital audio
signal are a lot like video signals, at one time, they WERE video
signals (and still are, it's just that there aren't a lot of
video recorders being used for this purpose any more).
The issues mentioned of jitter and the like could be remedied by
a properly design DAC where the D/A clock is not tied intimately
to theincoming bit rate. The longer-term average, on the order
of several millieseconds, is quite stable and accurate over this
sort of a link, while the bit-level timing could well be off by a
bit.