Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Viitalahde]

crna59 wrote on Tue, 05 August 2008 03:12

No "waffle irons" here... just straight cable!

What, no cable lifters? And that cable loop forms an inductor!

(cricket chirping)

[crna59]

Viitalahde wrote on Tue, 05 August 2008 10:32

What, no cable lifters? And that cable loop forms an inductor! (cricket chirping)

LOL....   cable lifters are only used on carpet where static can build up.

What's amazing about those cables, the shielding is so great, you can't detect voltage anywhere with a meter!

Regards,
Bruce

[Andy Peters]

Barry Hufker wrote on Tue, 05 August 2008 08:01

For analog, my understanding is "the fewer the connectors the better".  For digital, jitter only matters at the A/D and D/A stages.  Anything in between gets re-clocked at the next device.

For digital, the fewer connectors, the better. Connectors are impedance discontinuities (unless the connector was explicitly designed for high-speed digital transmission, which none of the audio standards were), and as such lead to reflections and all sorts of bad things. The longer the cable, the worse the result. If the receiving device can't sync on the incoming digital audio stream due to the reflections, you're hosed.

Analog audio connectors? Keep 'em clean and they are indistinguishable from the cable. Seriously.

-a

[Barry Hufker]

While I don't know how connectors figure into this,

At Mogami's site they say no more than 150 meters (492 feet) for AES.
Canare says 1,180 feet (350 meters), but they don't say whether that is equalized against high frequency (sampling rate) loss.
Gotham Cable doesn't specify.
The book, "Modern Recording Techniques," says 100 meters for AES (398 feet).
The book, "Fundamentals of Digital Audio," says no more than 350 feet with AES.
All assume of course a true AES cable with a Z of 110 ohms.

[Andrew Hamilton]

Barry Hufker wrote on Tue, 05 August 2008 11:01

For analog, my understanding is "the fewer the connectors the better".  For digital, jitter only matters at the A/D and D/A stages.  Anything in between gets re-clocked at the next device.

OK, however, isn't it true that all jitter-attenuating DACs are only that - jitter attenuating?  There is still some transmission jitter, depending on the length of cable, its properties, any reflections, # of PLLs, etc, no?...  Most of the time, it's so small that it can't be perceived.  That's why it's sometimes called "jitter elimination" (e.g., a significant amount gets reduced)...    But the accumulation of transmission jitter, I think, can be heard when the signal must go through several PLLs (two DAWs, L2, SFC2, etc...) in live audition mode - even with a Lavry Blue D/A using CrystalLock.  

So, if you have a long run and find you are faced between choosing to connect or to splice, I'm wondering if going the splice route is, however subtly so, the mo' better way, at least theoretically.


Thanks,
    Andrew

[Barry Hufker]

The only time jitter matters is A/D and D/A.  And the amount of jitter depends upon the quality of your converter.  Jitter at any point in between doesn't matter at all.

[tom eaton]

Andy Peters wrote on Tue, 05 August 2008 15:08

Analog audio connectors? Keep 'em clean and they are indistinguishable from the cable. Seriously. -a

I've always used the gold plated Neutriks, but have had some recent discussions about the merits of silver plated connectors (I am told that critical military and hospital connectors are usually silver).

I have used brass, nickel, and gold plated connectors in my life... and have a fondness for the brass for purely tactile reasons...and perhaps a lack of plating.

My instinct tells me that connectors are an obvious weak link, but I have no scientific evidence to back that up!  Keeping them clean is always a good idea, though.

tom

[cass anawaty]

Barry Hufker wrote on Tue, 05 August 2008 23:02

The only time jitter matters is A/D and D/A.  And the amount of jitter depends upon the quality of your converter.  Jitter at any point in between doesn't matter at all.

Not sure on that, or I might have misunderstood you--what do you mean by "in between"?  In between devices, or internal jitter of the converters themselves?

My understanding is that jitter is at it's worst during transmission between digital devices, esp when the clock is traveling w/ the data stream.

Anyone want to clarify?

[tom eaton]

I would prefer to have any processing that happens to a signal while in the digital domain be processing as accurate (true to the original) a signal as possible.  I understand Barry's basic point, but if you're mixing digitally, you want your mixer to be subject to as little jitter as possible...regardless of the converter on the other end.  You don't have to be monitoring a d/a for the daw to be working...

A good clock is beneficial at every stage, not just conversion.

tom

[Barry Hufker]

Individual devices in a digital signal chain re-clock the data.  Each device may produce more or less jitter than the device before it.  There are only two times when jitter matters. The amount of jitter matters is when you convert from A to D.  Then the jitter is a kind of distortion which is permanently a part of the recording.  The other time is D to A, when the amount of jitter in the D to A affects signal quality while you listen.

Here is an excellent paper by Bob Katz on the subject:
http://www.digido.com/bob-katz/jitter.html

[Andy Peters]

tom eaton wrote on Tue, 05 August 2008 19:07

I would prefer to have any processing that happens to a signal while in the digital domain be processing as accurate (true to the original) a signal as possible.  I understand Barry's basic point, but if you're mixing digitally, you want your mixer to be subject to as little jitter as possible...regardless of the converter on the other end.  You don't have to be monitoring a d/a for the daw to be working... A good clock is beneficial at every stage, not just conversion.

Once the data are in the device -- like, say, a digital mixer -- then the clocking in that domain is irrelevant.  By "clocking" we mean the word clock that tells the converters when a sample begins.

The calculations inside the DSP are done at whatever clock frequency the chip supports.

-a

[dcollins]

crna59 wrote on Tue, 05 August 2008 10:33

What's amazing about those cables, the shielding is so great, you can't detect voltage anywhere with a meter!

Anywhere where?


"For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled."

I should make it my .sig!

DC

[tom eaton]

Andy Peters wrote on Wed, 06 August 2008 01:07

tom eaton wrote on Tue, 05 August 2008 19:07 I would prefer to have any processing that happens to a signal while in the digital domain be processing as accurate (true to the original) a signal as possible.  I understand Barry's basic point, but if you're mixing digitally, you want your mixer to be subject to as little jitter as possible...regardless of the converter on the other end.  You don't have to be monitoring a d/a for the daw to be working... A good clock is beneficial at every stage, not just conversion. Once the data are in the device -- like, say, a digital mixer -- then the clocking in that domain is irrelevant.  By "clocking" we mean the word clock that tells the converters when a sample begins. The calculations inside the DSP are done at whatever clock frequency the chip supports. -a

Okay... educate me.

If I can varispeed my daw with my external clock, which I can, I assume (perhaps incorrectly) that the processing (delays, reverbs, etc.) that are happening in the DAW are changing with the clock, and therefore that the accuracy of the mix engine DOES rely on the clock.

You are saying that the mix engine is NOT subject to clocking?

tom

[Andrew Hamilton]

Barry Hufker wrote on Tue, 05 August 2008 18:02

The only time jitter matters is A/D and D/A.  And the amount of jitter depends upon the quality of your converter.  Jitter at any point in between doesn't matter at all.

Unfortunately, one must listen to the monitor D/A converter that may or may not be fed a signal the jitter of which is already low enough in level for the converter to attenuate beyond perception...  Having several digital patches in series will accumulate more of that ephemeral, transmission (clocking) jitter than if you were to patch directly from a DAW to a DAC (with a short, correct cable).  

Wouldn't it be best, in the interest of jitter management, therefore, to splice your S/P DIF and AES/EBU interconnects which are too short for their runs (rather than to connect them, using "noisy" connectors to make the daisy chain)?  



Andrew