Digital Coax vs Toslink(optical) - need answers
Aug 15, 2008 at 1:05 AM Post #31 of 56
Quote:

Originally Posted by Pars /img/forum/go_quote.gif
How long is the Zu? The Oxy (fuel) is supposed to be an analog interconnect. Try something 2m or longer and see if you can tell any difference.


The Zu is 1 meter and is a very stiff, solid core, shielded cable. Even though it is supposed to be an analog interconnect I thought it might also be used as a digital cable.

Why try a longer cable? All my cables are Dayton 18".

USG
 
Aug 15, 2008 at 1:34 AM Post #32 of 56
Quote:

Originally Posted by bigshot /img/forum/go_quote.gif
If it does perform differently using a different source or DAC, my guess is that the source or DAC is responsible, not the cable itself. Therefore, there is no inherent difference between toslink and coax. Either one will work. Does that make sense?


Again, I'm not saying there is a difference between the 'sound' of toslink and coax cables. Obviously they are different, one transmits light pulses and the other electrical pulses. What I am saying is that depending on the transport and/or dac they can and will sound different. Usually, given hi-end equipment you can bank on coax sounding better, but if you have a lot of interference, long cable runs, or a computer as source, toslink may be superior.

I'm not sure I really follow everything on that page either, but he seems to be saying that toslink has some inherent technical issues. Perhaps these are what cause the sonic differences?

upstateguy - Get some 75 ohm coax cable and a pair of Canare RCAP connectors and whip up your own IC that will smoke those two. [assuming here, have no experience with the Dayton]

Or just order one from here: Digital Audio Cables at Blue Jeans Cable
 
Aug 15, 2008 at 2:15 AM Post #33 of 56
Quote:

Originally Posted by philodox /img/forum/go_quote.gif

upstateguy - Get some 75 ohm coax cable and a pair of Canare RCAP connectors and whip up your own IC that will smoke those two. [assuming here, have no experience with the Dayton]

Or just order one from here: Digital Audio Cables at Blue Jeans Cable



Thanks for the tip, Philodox, I'll check it out...

here's where I got the Dayton Cables they're not bad at all.....

USG
 
Aug 15, 2008 at 3:35 AM Post #34 of 56
Quote:

Originally Posted by Slaughter /img/forum/go_quote.gif
It's a perfect analogy and you prove it by saying the audio will get to the other end without audible errors, thus proving that there is no difference between coax and optical. Thanks for clearing that up. And in case you didn't know, a transferred file doesn't always recover from transmission errors, thus the need for MD5 when we download large files. So again, they are exactly the same.


Hey fighting is not fun, but here goes. I sell file transfer services for a living, and the company I founded in my living room (now with over 100 employess and offices in 5 cities) is involved in moving 45% of the digital press releases in North America and 10% of feature news to newspapers every day over the internet.

I designed it all with W. Richard (Rich) Stevens, the author of Unix Network Programming (now unfortunately no longer with us), and he thanks me in the preface of Vol 2 and others of his books. I give speeches at news transmission meetings all over the world, and my company is considered the technology leader in the space by everyone.

Be careful if you try to take me on re file transfers. I have been studing the statistical properties of error propagation in communications for 40 years. I was an Member of Techincal Staff at Bell Labs. I have a BA from Princeton, summa cum laude, and a PhD from Yale, both in statistics. I won the Theory and Methods award from the Journal of the American Statistical Association, and have been selling file transfer software since 1981.

Standard file transfer protocols do have re-transmission and error correcting built in, and transfer perfectly if the transmission line stays up.

You have no idea what MD5 really is. A simple CRC-32 checksum will work for file integrity; MD5 is a cryptographically secure message digest used to stop tampering. It can be used as a simple check-sum on a file but that is overkill. There is typically no need in most environments to compute and verify a checksum; sometimes there is, it depends. Not relevant here.

Most file transfers have mild real-time requirements, or none. Streaming the bits in time to the music is not involved in file transfer. It is in digital audio.

It is as simple as that. Think about it. Streaming over the internet is the proper analogy to sending digital audio over a cable, not file transfers. As I said, some ethernet or WiFi digital audio protocols use buffers and can support re-transmits, up to a limit where the timing of the music takes over. That's a little like file transfer, but not much.

I only wanted to stop the "just bits" argument. It's: "bits AND timing". That's all. I don't want to fight. But I know my stuff, thank you, and your reply post used sarcasm, which is a no-no for me.
 
Aug 15, 2008 at 4:09 AM Post #35 of 56
Quote:

Originally Posted by upstateguy /img/forum/go_quote.gif
The Zu is 1 meter and is a very stiff, solid core, shielded cable. Even though it is supposed to be an analog interconnect I thought it might also be used as a digital cable.

Why try a longer cable? All my cables are Dayton 18".

USG



Reflections and timing. The longer cable allows any reflections to occur after the receiver is clocking an edge transition.

S/PDIF Circuits
 
Aug 15, 2008 at 4:27 AM Post #36 of 56
Thanks for posting your resume, it was not helpful. You took my example way to far, most likely due to your background. Of course you can break it down further as Ethernet is not Toslink or Coax. Okay I concede, it is bits and timing, but the fact is that 99.999% of the time the bits get there on time, so the timing piece is almost irrelevant as it is a given that the bits will get there, just as it is with file transfers. Lighten up.
 
Aug 15, 2008 at 4:28 AM Post #37 of 56
Quote:

Originally Posted by Pars /img/forum/go_quote.gif
Reflections and timing. The longer cable allows any reflections to occur after the receiver is clocking an edge transition.

S/PDIF Circuits



Great reading, thanks for the url. They use 6m cables. I have read elsewhere that 1m is fine. (18" would not be). Do you agree?
 
Aug 15, 2008 at 4:36 AM Post #38 of 56
Quote:

Originally Posted by Slaughter /img/forum/go_quote.gif
Thanks for posting your resume, it was not helpful. You took my example way to far ... the fact is that 99.999% of the time the bits get there on time, so the timing piece is almost irrelevant as it is a given that the bits will get there, just as it is with file transfers. Lighten up.


Didn't mean not to be "light" ... sorry. We are all here to enjoy our hobby, nothing more. I just want things to be said right. And without sarcasm.

BTW I posted (parts of) my resume only to establish credentials, which I find helpful when I read the works of others. Nothing more.

I agree, and have often posted, that most bitsteam errors are in fact inaudible, just as you say. But then people go past that and claim "bit perfect" etc. And sometimes they say "any cable will do", and that's not really true either.

We part as friends (I hope).
 
Aug 15, 2008 at 4:45 AM Post #39 of 56
Quote:

Originally Posted by wavoman /img/forum/go_quote.gif
Great reading, thanks for the url. They use 6m cables. I have read elsewhere that 1m is fine. (18" would not be). Do you agree?


I would say 1.5m min., but I haven't done the math. Clocking in SPDIF is at ~2.8MHz IIRC.
 
Aug 15, 2008 at 5:03 PM Post #40 of 56
If there is no difference in sound, and if the bitstream errors are inaudible, then any cable really will do.

See ya
Steve
 
Aug 15, 2008 at 7:09 PM Post #41 of 56
I'm not rephrasing what I've been saying yet again... This is getting ridiculous.

'Believe' what you will.
 
Aug 15, 2008 at 8:26 PM Post #42 of 56
This is another one of those
deadhorse.gif
topics. Such fun!

Basically, just get both from BJC and see for yourself. Go with what you think sounds better. It's not that expensive... you can sell the loser if you can't take the $ hit of having both.
 
Aug 15, 2008 at 9:06 PM Post #43 of 56
Quote:

Originally Posted by floydenheimer /img/forum/go_quote.gif
This is another one of those
deadhorse.gif
topics. Such fun!

Basically, just get both from BJC and see for yourself. Go with what you think sounds better. It's not that expensive... you can sell the loser if you can't take the $ hit of having both.



Too funny.
 
Aug 16, 2008 at 1:06 AM Post #44 of 56
Quote:

Originally Posted by bigshot /img/forum/go_quote.gif
If there is no difference in sound, and if the bitstream errors are inaudible, then any cable really will do.

See ya
Steve



It's all subjective. Some claim to hear differences, others don't. For me personally, I'll use the less expensive digital coax for runs less than 10feet, anything more than that and I think it would be wise to spring for the optical.
 
Aug 16, 2008 at 1:13 AM Post #45 of 56
Quote:

Originally Posted by bigshot /img/forum/go_quote.gif
If there is no difference in sound, and if the bitstream errors are inaudible, then any cable really will do.


True, but IMO not all bitstream errors are inaudible. Take a cheap no-name USB 1.1 cable with big-ass ferrite beads and use it to stream 24/96 digital audio from your PC to your DAC, then compare with a decent cable. I did this and I hear mush from the cheap-y.

Could be placebo effect, no doubt. So we are going to test this again with a blind testing protocol at a NJ mini-meet around XMAS time. Will report.
 

Users who are viewing this thread

Back
Top