S/PDIF, was Re: 240volts and ground hum problems.

[The Dark force of dance]

Y-ellow Ingo 'n' y'all.

At 03:06 PM 06/16/99 +0200, Ingo Debus wrote:

>I just added a coaxial S/PDIF out to my Sony DTC-790 DAT recorder (has
>optical out only, but optical and coax in). It was very simple: I just
>connected the signal from one pin of the optical transmitter via a 390
>ohms resistor to the output jack. I saw in an old Elrad project that
>they did it the same way from the output of a 74HC04. Works good. But
>I'm wondering if I should add anything? Protection zener diodes perhaps?
>I don't quite see the need for a transformer, since the ground of the
>coax input is connected to the DAT recorder's ground.
>Why did Sony leave out this feature if it costs just one resistor and
>one jack??

Yup that sound fairy nuff to me. I'm sure manufacturers do this kinda crap
simply to try and force people to use the more expensive optical hardware.
I imagine there are a lot of people either buying converter boxes or
replacing their other hardware, upping sales and keeping a sector of the
industry moving but for 2 cheap components.

I'm sure you don't need transformer isolation. don't know about protection.
The resistor is probably all the protection you need. (If I've imagined
it's orientation right?) In my case however I'm getting to the point where
isolation is starting to look necessary. I am facing a similar problem to
those people who would get pissed off about having to use optical and it
just amazes me the lack of forethought that goes into these things.

My O1V is a wonderful box. I am more than pleased with it. Except that
you'd think with all the wall to wall DSP power they could actually work
out how to mix S/PDIF signals. In video these days, in fact for the past 11
years that I know of, we've had the availability of things called "Digital
frame synchronizers." Or "Digital frame stores." not to be confused with
TBCs, (Time base correctors) which have been around for just about ever.
Video, though not intrinsically digital, has the same problems as digital
audio. It requires synchronization. In order to mix 2 or more live cameras
together, each camera, traditionally had to be locked to the same sync
source. It's called "Gen-locking" and amiga people should be familiar with
it. 

If you then wanted to mix in a pre-recorded video source as well as the
live cameras you have a problem. You either have to derive you master sync
from the video machine, which means it has to be running all the time and
is not convenient, or, you have to have a very special A/B roll video
machine that can synchronize the tape to the master sync. As it turns out,
this is not all that hard and the video industry have been doing it for
just about ever.

But along came the advent of cheap TBCs and frame stores which meant that
even though a master sync was necessary, this sync could be derived  from
one of the cameras or any other source. As long as the frame sync box was
synchronized it could align every other video signal. It does this simply
by having the ability to delay each video signal by one half frame. It then
re-spits the signal out in sync with the master. You don't even notice it. 

TBCs are so common these days they actually build them into betacams.
Especially betacam SP racks. It means you can actually press the pause or
even rewind button on the tape machine and the damn thing will stay locked.
You might see some glitchy frame lines but you'll still get a stable picture.

The thing is if they can do all that so damn cheaply for video, which is
much harder to do, why, with all the DSP power available to them, do they
not provide anything other than analogue inputs on all the channels of all
the digital mixers? Considering that it's actually cheaper to provide a
digital input than it is an analogue one. On the O1V, they even provide and
I/O for external word-clock sync. Derr...

The reason is that you can buy an "Options adaptor." Which I don't have a
price on but we're probably looking at another 500 bux at least. You can
only get an AES/EBU adaptor. (Or a single TOSSLINK adaptor)

The point is that  all this stuff can be done and one day may even become
as standard as these things are in the video world, but for now it's all to
do with marketing and nothing to do with how expensive it is.

People thing digital is expensive because the hardware is more expensive.
Because it's more complex. What they fail to realize is that Digital stuff
should be cheaper than analogue because it's about 80% cheaper to stuff a
couple of million transistors on a substrate than it is to design and build
a good analogue circuit. And coupled with the fact that it's so
reproducible and libratized that manufacturing costs are also way down as
compared. Companies like sony and all the synth companies don't make
digital synths because they're cleaner, but because they're so much cheaper
to make.  So in essence, we're all getting ripped off with this digital
gear but what can you do?

>I heard S/PDIF and AES/EBU can just be connected, but I never tried
>myself.

In theory, there is not a whole swag of difference between AES/EBU and
S/PDIF.  Apart from electrically speaking of course. All S/PDIF is in fact
is a the same protocol as  AES/EBU but the sub codes are different. But as
they say, the devil is in the details.

First of all they had to shift a whole pile of stuff in the AES/EBU sub
code arrangement in order to make room for the SCMS copy protection. (Why
did they even bother?) I can easily tell most of my digital gear to think
it's something other than what it is. I always master CDs which will
happily tell any digital system they are playing through, that they are in
fact a synthesizer. Not a problem. But I'm not sure how the subcodes for
SCMS and the subcodes shifted to accommodate them, would effect the
receivers of the opposite standards.

All the texts I've read and the  few experts I've spoken to say. "You just
don't do that." But these guys don't live in the real world. Where it's
either a case of finding a way to actually do it, or not doing it at all.
Like Ingo I have no qualms about hacking into some piece of hardware when
it only costs me a resistor and a jack to try it out. But when it costs 500
bux or more and it's less than certain, I really feel I should be more sure
of what I'm messing with. :)

So it looks as though in theory this should be entirely possible.
Electrically all you would do is hook the S/PDIF to the AES/EBU in the same
way that you would hook an unbalanced line into a balanced one. As long as
it didn't bawk at the subcodes, everything else should work. Knock on wood. :)

Be absolutely Icebox.

 _ __        _                              
| "_ \      | |         batzman at all-electric.com / aek at all-electric.com
| |_)/  __ _| |_ ____           ALL ELECTRIC KITCHEN               
|  _ \ / _` | __|___ |  The Elementary Urban sanity CD
| |_) | (_| | |_  / /   out now on Transmission Communications
|_,__/ \__,_|\__|/ /    
                / ,__   http://www.all-electric.com
Goodfortune    |_____|