Spdif: glass fiber optical cable vs acrylic optical cable vs coaxial cable

[danvu]

I'm shopping for a couple more of digital cables. Have read that they both have pros and cons but audiophiles usually prefer coaxial type. It is true? The whole thing's kinda confusing and unclear to me. So can some members tell me what you prefer/ use and why? Have anyone tried all types cable with good quality and hear any difference? Your comments are appreciated.

 

[Spike]

Theoretically, coax is preferred since the signal coming off the transport mechanism is electrical and gets "transmitted" via copper cable to the receiver. With optical connection, this electrical signal needs to be converted to optical pulses for transmission. On the receiving end, this optical pulses need to be converted back to electrical signal to be fed into the DAC chip. Coax, being electrical in nature, acts like an antenna and is susceptible to picking up noise. Optical cable does not have this problem. This is an over-simplistic view which rarely translates directly to the sound quality you can hear at the system level. The devil is in the implementation details, and whether your system has the resolution to allow you to hear the difference. That said, for coax digital cable, you should look for those with true 75-ohms specification and with good shielding. For optical, I usually look for the glass fiber optical cables. You should audition for yourself in your system.

Spike

 

[jmschnur]

The RCA cable should also be 2 meters or so but not 1 meter.

 

[amey01]

As Spike said - listen and see. It all depends on how each is implemented on your particular components. It is worth noting though - some people say optical is better for "noisy" components (eg a Squeezebox) as it isolates it from your system. I have a Squeezebox and I use coax without any issue. So, it all depends. Quite frankly, you're probably better off spending your time and money worrying about something else.

 

[bzr]

The RCA cable should also be 2 meters or so but not 1 meter.

Really? Why?

 

[danvu]

Thanks spike and amey for clearing things out. So I actually planned to get a good quality/shielding 75ohm coaxial and a glass fiber one to try out but wasn't so sure. Your comments give me more confident to do that.
And yeah, jmschnur, why should I get a 2m coaxial cable?

 

[Speedskater]

The RCA cable should also be 2 meters or so but not 1 meter.

Really? Why?
I second that.
It's the question of the day.

 

[jmschnur]

Really? Why?

Reduced jitter caused by impedance issues. The Logitech squeezebox audiophile forum has multiple threads on this. 3 feet turns out to be have some issues with most DACs. 1.5 meter or longer is a good choice. Too long leads to more possibility of rf interference. Easy way to check get both lengths and try it. See also:

.http://www.positive-feedback.com/Issue14/spdif.htm

 

[tsv_1]

Reduced jitter caused by impedance issues. The Logitech squeezebox audiophile forum has multiple threads on this. 3 feet turns out to be have some issues with most DACs. Easy way to check get both lengths and try it. See also:

.http://www.positive-feedback.com/Issue14/spdif.htm

Good article, thanks for sharing. This is one reason you see BNC connectors on higher build-quality equipment... helps to better maintain the ideal 75 Ohm characteristic impedance of the transmission line.

 

[Speedskater]

That article is sort of like writing:
If one plus one equals two then blue is a good color.

 

[tsv_1]

That article is sort of like writing:
If one plus one equals two then blue is a good color.

Huh? Nugent hit all the right points. It's about time domain reflections that are entirely dependent on transmission path length and signal propagation speeds. Can you comment beyond your non-sequitur analogy?

 

[FoxMulder]

I have always read and been told that a 1.5 meter digital cable was optimum.

 

[Audioseduction]

I use a Nordost Silver Shadow BNC Digital Cable and experiance excellent SQ performance. BNC on both ends is the best way to go.

 

[Speedskater]

Huh? Nugent hit all the right points. It's about time domain reflections that are entirely dependent on transmission path length and signal propagation speeds. Can you comment beyond your non-sequitur analogy?

While Steve Nugent is an interesting person, his writings are somewhere between a non-sequitur and a oxymoron. On his web-page http://www.empiricalaudio.com/ he has several interesting papers. In his writings, he will have a good technical discussion in one paragraph. The next paragraph will be into a fantasy world (like measuring speaker cables at 100 MHz). The third paragraph will contradict the first paragraph.

As for the S/PDIF cable paper, the old archived pro audio forum has a almost 200 post thread on the technical errors in the paper. Steve joined the thread for one day then left when his mistakes were not accepted.

http://recforums.prosoundweb.com/index.php/mv/msg/11678/0/0/21629/

 

[jmschnur]

Still that does not mean the comments on the length of t.he coax were wrong. Just that some of the reasons supporting a long length might be in error.

 

[Speedskater]

It sure doesn't make the comments correct!
No, really the thread covers why the co-ax length idea is incorrect.

 

[jmschnur]

Best way to check is to try three different lengths and see if you can tell if there is a difference .

 

[tsv_1]

While Steve Nugent is an interesting person, his writings are somewhere between a non-sequitur and a oxymoron. On his web-page http://www.empiricalaudio.com/ he has several interesting papers. In his writings, he will have a good technical discussion in one paragraph. The next paragraph will be into a fantasy world (like measuring speaker cables at 100 MHz). The third paragraph will contradict the first paragraph.

As for the S/PDIF cable paper, the old archived pro audio forum has a almost 200 post thread on the technical errors in the paper. Steve joined the thread for one day then left when his mistakes were not accepted.

http://recforums.prosoundweb.com/index.php/mv/msg/11678/0/0/21629/

Huh... I read the first few posts and I still agree with the basic point Nugent was making. Time domain reflections can and do have a very real impact on comparator slicing points based on cable propogation delay and path length (assuming non ideal impedance matching). That was his main point and it is accurate. I'm not defending his credentials or his behavior wrt bailing from the forum thread... but the basic tenet is accurate. And unlike Mr. Nugent, I don't plan to bail