Is AES/EBU better sounding ??
Jay Rose
|: We had a debate today about AES/EBU , SPDIF and TOS optical whether the
|: first one had better sound quality when transferring digital data or
|: that all three areof the same quality ...
Assuming they're all 16 bit words and properly wired (lightpiped?), the
audio data stream is the same. Header info may vary.
--
Jay Rose's Digital Playroom
industry journalist / Clio+Emmy winning Sound Designer
Want to learn audio for video? http://www.dplay.com/book
correct return address is jay at DPLAY and then dot com
danceon
first one had better sound quality when transferring digital data or
that all three areof the same quality ...
Thanks
Dimitrios
Scott Dorsey
If your D/A converter is clean and reclocks the data, then bits are bits
and it doesn't matter how those bits get there.
There are installations where you can hear a quality difference between
different interfaces. Those are installations which should invest in
better converters. Yes, I have been to places where there were audible
differences, but this is a sign that something else is wrong.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
David Satz
Scott Dorsey <klu...@netcom.com> wrote:
> If your D/A converter is clean and reclocks the data, then bits are bits
> and it doesn't matter how those bits get there.
>
> There are installations where you can hear a quality difference between
> different interfaces. Those are installations which should invest in
> better converters. Yes, I have been to places where there were audible
> differences, but this is a sign that something else is wrong.
and I would simply like to say, HOORAY for the common sense
and accuracy of this reply. So many people form (and publish)
their opinions about the "sound" of equipment without considering
the system in which it is being used, and often these opinions are
in fact a reaction to poor engineering elsewhere in the makeup of
the system.
--Best regards
Mike Rivers
> We had a debate today about AES/EBU , SPDIF and TOS optical whether the
> first one had better sound quality when transferring digital data or
> that all three areof the same quality ...
In theory, there should be no difference. In practice, AES/EBU tends
to have less jitter introduced by the cable than the other two means
of transport. It's really nothing to worry about, though.
--
I'm really Mike Rivers (mri...@d-and-d.com)
ScotFraser
In article <38753C06...@otenet.gr>, danceon wrote:
<<We had a debate today about AES/EBU , SPDIF and TOS optical whether the
first one had better sound quality when transferring digital data or
that all three areof the same quality ...>>
There is no difference in sound quality. All the same bits are transferred via
either protocol. The real difference between SPDIF & AES/EBU beyond the
connectors & voltage level is that SPDIF transfers DAT ID numbers & AES/EBU
does not.
Scott Fraser
Jay - Atlanta Digital
scotf...@aol.com (ScotFraser) wrote:
But there are subtle differences depending on if you are using glass or
plastic fiber for toslink (it's pretty much always plastic due to cost and
fraility, but many swear glass is better), the quality of the jacks, the
correct cable (mic cable is not AES cable). There are jitter issues
involved and some other variables. Doesn't make one superior to the other
always - there are pros and cons to each, and the well designed and
implimented of all flavors (toslink, SPDIF on RCA or BNC, AES on XLR, and
some other lesser know flavors like Sony SDIF) can be very good, while the
cheapo examples will be worse. It's not specifically the format that
makes the most difference.
-Jay Frigoletto
(Atlanta Digital)
We've moved to Hollywood!
http://www.promastering.com
rAdiO bAsTaRd
> >first one had better sound quality when transferring digital data or
> >that all three areof the same quality ...
Anyone's OPINION would be of no use here.
The FACT (non-debateable) is that there is no difference in sound.
You are in fact asking: "which sounds better, digital or digital?"
Scott Dorsey
>Anyone's OPINION would be of no use here.
>
>The FACT (non-debateable) is that there is no difference in sound.
>
>You are in fact asking: "which sounds better, digital or digital?"
Oh, I can make up plenty of situations on the test bench where sonic
differences get created. Take a Panasonic SV3700 and pull the outputs
into an early Yamaha D/A, and you'll hear substantial differences in
sound between the two, because the timing on the Panasonic data is
screwy and the Yamaha doesn't reclock it properly. Bits are bits,
but if the bits aren't arriving at the right time, the analogue waveform
won't be reconstructed properly.
The solution to this problem, of course, is to stop using badly designed
equipment with unstable PLLs, not to worry about which interface sounds
better. But a lot of people go looking in the wrong place when they
notice this kind of thing, and don't realize it's an artifact of the
converters and not a sign of poor data transfer.
Mike Rivers
> Anyone's OPINION would be of no use here.
> The FACT (non-debateable) is that there is no difference in sound.
>
> You are in fact asking: "which sounds better, digital or digital?"
Oh, if it was only that simple - and it would be, if those bits always
stayed in the same machine, on all projects, forever. The problems
come when we have to ship them from one unit to another, and that's
where we use AES/EBU and S/PDIF.
While the bits may remain the same, there are real, physical aspects
of cabling that affect the clocking of those bits, and that changes
the sound in usually subtle but detectable ways.
While the same problems can (and do) occur with either physical
interface format, the balanced nature of AES/EBU tends to be less
susceptible, all other things being equal, to things that cause timing
distortion. This doesn't mean that you should always use AES/EBU, it
just means that you may sometimes be able to get away with being a
little less careful and not cause noticable damage in the final
product.
David Satz
Mike Rivers <mri...@d-and-d.com> wrote:
> Oh, if it was only that simple - and it would be, if those bits always
> stayed in the same machine, on all projects, forever. The problems
> come when we have to ship them from one unit to another, and that's
> where we use AES/EBU and S/PDIF.
>
> While the bits may remain the same, there are real, physical aspects
> of cabling that affect the clocking of those bits, and that changes
> the sound in usually subtle but detectable ways.
And the next question is, are those bits then stored (on tape
or disk or ...) or are they being played back immediately?
Because if they're being stored, we're back to "bits is bits."
It wouldn't matter whether there is a ton of jitter in the
line or not, as long as the receiving device can still track it
and recover the correct data. And that is the usual case
in making any kind of transfers or dubs, barring improper
operation or equipment malfunction.
The case where your cautions would apply is the one that
mainly affects audiophiles who have more money than sense:
listening to digital recordings through a separate, outboard
DAC that derives its word clock signal from the data stream
it's being fed. That (stupid) approach _is_ vulnerable to the
effects you're talking about.
Or maybe I shouldn't be so snide toward audiophiles; they're
just doing what their pretentious, stupid magazines tell them
to do ("Separates are always better, so be sure to look down your
nose at CD players--instead buy a CD 'transport' and an outboard
DAC, and focus more on the quality of the cable between them
than on the fact that the timing information in that cable is
going in the wrong direction ...").
Scott Dorsey
>
>And the next question is, are those bits then stored (on tape
>or disk or ...) or are they being played back immediately?
>Because if they're being stored, we're back to "bits is bits."
>It wouldn't matter whether there is a ton of jitter in the
>line or not, as long as the receiving device can still track it
>and recover the correct data. And that is the usual case
>in making any kind of transfers or dubs, barring improper
>operation or equipment malfunction.
>
>The case where your cautions would apply is the one that
>mainly affects audiophiles who have more money than sense:
>listening to digital recordings through a separate, outboard
>DAC that derives its word clock signal from the data stream
>it's being fed. That (stupid) approach _is_ vulnerable to the
>effects you're talking about.
Nope. If you pop the cover of your typical CD player, you will see that
there is a recovered data clock generated from the eye-pattern data coming
off the CD, and that the internal DAC is synchronized with that clock.
If that clock has jitter (phase noise) on it, there will be audible sidebands
on pure tones fed through the machine. You can hear it, and you can see
it on a spectrum analyzer.
Having an outboard DAC is actually a good thing because it gives you more
points at which you can reclock the data. Some CD players out there actually
do a decent job with an internal PLL that produces a clean clock for the
internal DAC, even if the pit spacing on the CD is bad. Some don't.
>Or maybe I shouldn't be so snide toward audiophiles; they're
>just doing what their pretentious, stupid magazines tell them
>to do ("Separates are always better, so be sure to look down your
>nose at CD players--instead buy a CD 'transport' and an outboard
>DAC, and focus more on the quality of the cable between them
>than on the fact that the timing information in that cable is
>going in the wrong direction ...").
You should listen to some of this stuff. I can fault the audiophile
community for always thinking that "different" is "better," but when
it all comes down to it, I think anything that encourages people to
think more seriously about sound is inherently good.
Mike Rivers
In article <##nuwN4W$GA.255@cpmsnbbsa02> DS...@msn.com writes:
> The case where your cautions would apply is the one that
> mainly affects audiophiles who have more money than sense:
> listening to digital recordings through a separate, outboard
> DAC that derives its word clock signal from the data stream
> it's being fed. That (stupid) approach _is_ vulnerable to the
> effects you're talking about.
I don't follow audiophiles. Are they into external master word
clocks yet?
David Satz
> Nope. If you pop the cover of your typical CD player, you will see that
> there is a recovered data clock generated from the eye-pattern data coming
> off the CD, and that the internal DAC is synchronized with that clock.
Sorry, but you are stating it exactly backwards. The clock signal to
which you refer is seen only on the input side of the data buffer.
After error correction, the samples are de-interleaved and clocked
out of the buffer according to the DAC's wordclock.
The spindle servo is driven by a logic circuit whose job is to keep the
buffer from getting too full or too empty. If that requires adjusting
the spindle to faster or slower than real time, then that is what will
occur--but that has no effect on the rate or regularity with which
the recovered samples exit the buffer on their way to the DAC,
since the two clocks are not otherwise linked in any way.
Unfortunately, in some players (at all price ranges, not just the
cheap ones) the action of speeding up or slowing down the spindle
motor will inject noise into the power supply feeding the analog
circuitry. But that's a rather different problem.
> Having an outboard DAC is actually a good thing because it gives
> you more points at which you can reclock the data.
If we're talking about a data stream feeding a DAC, it _truly_ doesn't
matter how many times the stream is reclocked--what matters is
whether the DAC is linear, and is driven by a stable clock. And in
single-piece CD players, the DAC's clock is in no way dependent
on anything "upstream" from itself--it is the master clock for all
circuits that precede it, starting at the output of the sample buffer.
(I was trained by Sony to repair and maintain their studio digital
equipment back in the PCM-1600/PCM-1610 era, and have the
manuals here to back up what I'm saying. In general I find your
posts to be extremely well informed, and I suspect that this one
time you are simply making the type of mistake we all make from
time to time, perhaps typing one thing and meaning another?)
--best regards
Gordon Waters
Yeah, I think so, at least in a few instances. A lot of them seem to be into
that "MUST....ELIMINATE...CLOCK...JITTER..." thing in a very hyperactive
way... so it'd seem like it would be like taking candy from a baby, to
sell something like this...
Regards,
Gordon.
who just hopes none of those word clocks have the name "Tice" on them... -_-
--
GALAXY convention --------- Anime Weekend Atlanta 5- October 8-10,1999
/| || //| // /| ,, //~// //~// //~// ----- Marriott Gwinnett Hotel
//|| ||//||// //|| ./ //_// //_// //_// --- http://www.anime.net/~awa
//~~|| |/ |/ //~~|| / ,,_// ,,_// ,,_// Gordon Waters-...@crl.com
Scott Dorsey
>Scott Dorsey <klu...@netcom.com> wrote:
>
>> Nope. If you pop the cover of your typical CD player, you will see that
>> there is a recovered data clock generated from the eye-pattern data coming
>> off the CD, and that the internal DAC is synchronized with that clock.
>
>buffer from getting too full or too empty. If that requires adjusting
>the spindle to faster or slower than real time, then that is what will
>occur--but that has no effect on the rate or regularity with which
>the recovered samples exit the buffer on their way to the DAC,
>since the two clocks are not otherwise linked in any way.
This is certainly a much smarter way of doing things. Admittedly I have
only looked inside the older players, which have very limited buffering.
The older Philips chipsets definitely did work the way I describe although
I certainly am glad to see that has gone away.
>Unfortunately, in some players (at all price ranges, not just the
>cheap ones) the action of speeding up or slowing down the spindle
>motor will inject noise into the power supply feeding the analog
>circuitry. But that's a rather different problem.
Yes, and the servo itself invariably injects noise into the power supply
feeding the analogue circuitry as well. And then we have the matter of
noise on the clock line caused by supply variations (which really should
not be a big deal with a fixed clock).
>> Having an outboard DAC is actually a good thing because it gives
>> you more points at which you can reclock the data.
>
>If we're talking about a data stream feeding a DAC, it _truly_ doesn't
>matter how many times the stream is reclocked--what matters is
>whether the DAC is linear, and is driven by a stable clock. And in
>single-piece CD players, the DAC's clock is in no way dependent
>on anything "upstream" from itself--it is the master clock for all
>circuits that precede it, starting at the output of the sample buffer.
If this it the case, and the master clock itself is a fixed frequency
oscillator and not a PLL that has to be pulled to match an incoming data
rate, then it would seem that high clock stability should be pretty
easy to achieve. If this is the case than how come so many of the
single-piece CD players measure so poorly? (Assuming that they do
still measure so poorly; it has been almost ten years since I have
done serious bench tests on these things.)
>(I was trained by Sony to repair and maintain their studio digital
>equipment back in the PCM-1600/PCM-1610 era, and have the
>manuals here to back up what I'm saying. In general I find your
>posts to be extremely well informed, and I suspect that this one
>time you are simply making the type of mistake we all make from
>time to time, perhaps typing one thing and meaning another?)
No, I suspect that what has happened is that things have changed pretty
substantially since the time I seriously looked into this stuff back
in the early eighties. Any suggestions for references that I could
look into which would bring me a bit closer to the modern era?
I really do feel out of touch with the consumer stuff.