Ralph Smeets wrote:
>
[snip]
>
> Sorry "Rat", but you're wrong.
>
> I agree that bits are bits. But coax and TosLink don't do error-correction since
> the S/PDIF protocol doesn't do error-correction.
>
Maybe not, but it does do jitter correction. The S/PDIF lower level
protocol is what the computer people call "self clocking". That is,
the data stream not only encodes the data stream it also encodes
the clock that was used to generate it. The receiver runs a clock
generator synchronized to the encoded clock and uses that to
decode the data part. This bit stream is then assembled into words,
one for each sample, and the words are clock out to the DAC using
another clock synchronized to the first clock generator.
The draft AES/EBU standard specifies the max clock jitter _at
the DAC_ as 1 nanosecond. Most pro equipment will beat that and
most consumer equipment won't be too far behind.
And none of this depends in the least on what medium is carrying
the bit stream.
> I agree that for recording, it doesn't matter that there is jitter.
>
> But before we go any further, I think it's time to explain 'what is jitter'.
>
> Jitter is an effect that is caused by the fact that the interval between the
> samples varies a little. This variation isn't constant. Ie, consider a series of
> samples: T1, T2, T3 and T4. Let's say they where recorded at 0, 1, 2 and 3
> seconds. The value of T1, T2, T3 and T4 equals 1, -1, 1, -1. Ie:
> 1 @ 0.0
> -1 @ 1.0
> 1 @ 2.0
> -1 @ 3.0
>
> So lets play them back. But with some jitter introduced of -0.1 +0.1 +0.1 -0.1.
> Thus,
> 1 @ -0.1
> -1 @ 1.1
> 1 @ 2.1
> -1 @ 2.9
>
> I think you agree that this series of bits represents another waveform. That's
> what we call jitter!
>
> Cheers,
> Ralph
You're proposing a 10% jitter. The usual 44.1 kHz sample rate is
~22700 nanoseconds/sample. The bit time is roughly 333 ns/bit.
The receiving clock will lose sync with the data stream if the
bit jitter is 1/2 bit time or 160 ns. Actually, it would be real
hard to build a clock generator that would sync up with a jitter
of 1/4 or 80 ns. So lets assume an 80ns/bit jitter as a worst
case.
The samples are transmitted as 32 bit words. The time jitter on a
word is (bit jitter)/sqrt(n) or roughly 14 ns. Compared to the the
sample time this is .06% absolute worst case.
I'll concur with the person who suggested applying green magic
marker to the cable.
++PLS
-----------------------------------------------------------------
To stop getting this list send a message containing just the word
"unsubscribe" to [EMAIL PROTECTED]
