Tosklink - Optical - Role in hi-end?

[aps]

It’s always struck me as odd that Toslink / Optical hasn’t had a role in high-end computer audio given the potential for electrical isolation from the computer. The conventional wisdom is that this is because optical has higher jitter than RCA / Coax when delivering S/PDIF signals. Is this, in fact, correct? And, if so, is it inherent to optical or are there exceptional optical implmentations? And what is the promise, or otherwise, of optical Thunderbolt cables?

 

APS

[mayhem13]

It’s always struck me as odd that Toslink / Optical hasn’t had a role in high-end computer audio given the potential for electrical isolation from the computer. The conventional wisdom is that this is because optical has higher jitter than RCA / Coax when delivering S/PDIF signals. Is this, in fact, correct? And, if so, is it inherent to optical or are there exceptional optical implmentations? And what is the promise, or otherwise, of optical Thunderbolt cables?

 

APS

 

All things being equal, yes....Optical transmission is more prone to jitter BUT as you mentions isolates all forms of EMI which when enter the DAC chip can promote jitter. Most of today's competent DACs are designed to handle jitter effectively and reduce it's effects well below the threshold of audibility.

 

Subjectively speaking and in my opinion, a well built Glass cable used between my devices that use a MiniTos connection did outperform entry level plastic Toslink cables. But we're not talking boutique stuff, but less than $60 as far as I can recall. It may not have been the glass at all, but the MiniTos connector which was a much tighter fit to the optical output and at the DAC side as well.

 

Incidentally, there was no Coax connector for direct comparison on the Apple devices used.

[One and a half]

One other reason is that commercial chips limit transmission rate to 96 kHz in optocouplers.

Proper (not toslink ) optical interfaces do exist for pro audio and industrial use but the cost is often more than the pc/mac itself.

[wgscott]

My sense is Toslink is better than people usually give it credit for. A priori, it would seem to be better than anything but a (perfect) wireless connection. I think one of the problems is the quality of mass-produced cables. I had a $0.31 optical cable that couldn't handle 96kHz, and a $10 one that seems nearly as good as my $80 siflex glass cable. I had just assumed all were created equal. If the quality is so variable that they are not all equally able to transmit 0s and 1s, it is easy to see how they can get to have a bad reputation compared to coax, all of which should be at least approximately the same, at a given length.

 

The Schitt bitfrost is made primarily for optical. They wouldn't do that if it inherently sucked.

[sandyk]

Another problem with the cheaper plastic fibre Toslink cables, is that the mating ends are easily scurffed and damaged by repeated insertions. The very nature of the socket design and it's colour , plus the fact that the sockets are normally at the rear of the equipment that may be located inside a cabinet does not assist in that regard.

[spdif-usb]

The measured jitter (at the analog output of a DAC) using Toslink or optical S/PDIF is usually higher than when using coaxial (RCA or BNC) S/PDIF. Even though I agree that transport jitter nowadays has become much less of a problem in pure terms of audibility, it still exists nonetheless. Think of it this way. If a DAC chip uses an I²S implementation, the jitter caused by this implementation will be affected by the transport jitter caused by the S/PDIF interface that connects to it. From a manufacturer's standpoint, this poses all kinds of restrictions on the design and choice of components, so it is only logical the fact one simply doesn't ignore. There are tons of other variables literally at play.

Optocouplers tend to degrade over time. Cable distance and sharp cable bends can matter to the end user also. A delta sigma DAC often uses an ASRC the cost effectiveness of which depends on jitter. It isn't an easy task of always knowing in advance how or if one type of error (distortion) in the signal will audibly manifest itself in the presence of other errors. How is the designer of a DAC product supposed to know exactly which kind of amplification and speakers will be preferred by the potential customer? How does one define adequate headroom? You've got Moore's law at your side and Murphy's law breathing down your neck so if transport jitter can be avoided more easily (read: more commercially convincingly) than some other forms of improvement can be achieved then what do you think the law of economics will whisper to you do when it finally comes down to decisionmaking? Just my 2 cents.

[Kiwi2]

Personally I don't think it's really about the reasons already mentioned.

 

I think the reason toslink hasn't become popular with the high-end crowd, is that you can't pimp out a toslink cable with gold connectors or "superior shielding" and sell it for $2k. By design that $20 toslink cable down at Best Buy will perform just the same. I mean, who would take any such ad for a goldplated shielded optical cable in a high-end magazine seriously?

 

And we all know audiophiles don't like using cheap stuff because anything cheap sounds like crap. Kind of a catch 22 for toslink and audiophiles.

[Jud]

The Schitt bitfrost is made primarily for optical. They wouldn't do that if it inherently sucked.

 

Wasn't aware of that. Where did the information come from?

[sandyk]

My Oppo 981 DVD-A /SACD player has both Toslink and Coax SPDIF outputs. My DAC has 2 x Toslink and 1 Coax SPDIF input. I am readily able to remotely switch between both , and demonstrate to others, as I have done recently, that using the Coax connection results in better HF detail with a further improvement in Soundstage. Perhaps if I was to use the glass fibre Toslink cable that wgscott prefers, the difference may be far less obvious? Coming from a Telecommunications background, I am well aware of the much lower attenuation and vastly improved bandwidth capabilities of glass fibre over plastic fibre.

I shouldn't need to go there about the resultant higher Jitter that usually results from the additional steps of conversion to optical at the transmission end, and the conversion from optical at the receiving end may result in. Limited bandwidth in Optical receivers with their very relaxed quoted specifications can play a part here too, as does the quality of their power supply.

[spdif-usb]

Personally I don't think it's really about the reasons already mentioned.

 

I think the reason toslink hasn't become popular with the high-end crowd, is that you can't pimp out a toslink cable with gold connectors or "superior shielding" and sell it for $2k. By design that $20 toslink cable down at Best Buy will perform just the same. I mean, who would take any such ad for a goldplated shielded optical cable in a high-end magazine seriously?

 

And we all know audiophiles don't like using cheap stuff because anything cheap sounds like crap. Kind of a catch 22 for toslink and audiophiles.

Isn't that exactly what I was saying?

(read: more commercially convincingly) than some other forms of improvement

[wgscott]

Wasn't aware of that. Where did the information come from?

 

You.

 

I guess that would mean I am wrong.

 

(They do sell USB as an add-on, but say that optical is at least as good, right?)

[jhwalker]

You.

 

I guess that would mean I am wrong.

 

(They do sell USB as an add-on, but say that optical is at least as good, right?)

 

Their actual statement is that they believe SPDIF is the best interface - they include USB because they know people want it, but basically say it would not be their choice.

 

From their website:

 

"Well then, how about your USB? Is it fully buzzword compliant? 24/192? Async?It is absolutely buzzword compliant! Not only is it USB async, but it’s USB 2.0 async that’ll do up to 24 bits/192 kHz sampling rates. Yes. USB 2.0. Not antique 1.1. It works without drivers on Mac and we provide drivers for Windows 7, Vista, and XP. It’s a good-sounding, reliable, solid implementation of USB. But that’s like saying, “Well, its a very nice meal, given that the chef could only work with McDonald’s hamburgers.”*

 

Wait. Are you saying USB is crap?

We’re saying we put a ton of time into our USB implementation, but, to our ears, USB still doesn’t quite offer the performance of SPDIF. And we can even get into shades of gray on SPDIF too: consider Mike Moffat’s AT&T ST-optical interfaces and Sumo’s Axiom/Theorem transport and D/A, which had a separate low-jitter master clock connection from the transport."

[spdif-usb]

The USB implementation they (Schiit) are using is still far from the best USB can offer IMO.

[Jud]

The USB implementation they (Schiit) are using is still far from the best USB can offer IMO.

 

But not bad for a $100 retail board (may have gone to $150 since I bought my Bifrost).

[Jud]

Originally Posted by Jud

Wasn't aware of that. Where did the information come from?

 

 

You.

 

I guess that would mean I am wrong.

 

(They do sell USB as an add-on, but say that optical is at least as good, right?)

 

I think it means not so much that you are wrong, but that I was confusing (like the old radio serial - "the power to cloud men's minds").

 

So let me go through a sort of "survey" of what I understand to be the primary advantages and disadvantages of three popular interfaces. Those who know more (and you are legion), please feel free to correct anything below that's inaccurate.

 

Toslink - One of the popular forms of S/PDIF (Sony/Phillips Digital Interconnect Format). Advantage: Won't transmit electrical interference. Disadvantages: Limited bandwidth, at least in the most common implementations. (I don't know for certain if that's why the toslink connections I'm familiar with appear to be limited to 24/96 resolution. Information on the web seems to be all over the place about this.) At least some popular sources appear to have a lot of jitter, and toslink is less capable of dealing with this potential problem than async USB.

 

Coax - Another popular form of S/PDIF. Advantages: Higher bandwidth than toslink. Was the primary "high end" digital audio interface for a while, so sources may have less jitter, and some means of dealing with jitter have been developed (e.g., locking the receiver's clock to that of the source through a PLL). Disadvantages: Less capable of remedying potential jitter problems than async USB. Subject to transmitting electrical interference.

 

Asynchronous USB - Advantages: Capable of ameliorating jitter to a greater extent than S/PDIF, at least in current popular implementations. With USB2, plenty of bandwidth. Virtually all computers have this interface, so no special transport, sound card, or converter is necessary. Disadvantages: Subject to transmitting electrical interference. ("Galvanic isolation" has become a marketing term, and doesn't assure complete isolation from electrical interference, especially through ground.) USB2 requires drivers on the world's most common OS (Windows), while USB 1.1 does not require drivers but is bandwidth-limited and thus restricted to 24/96 resolution.

 

So looking at the above, to me it's "horses for courses." If you have a very low-jitter source and your primary problem is electrical interference, toslink's your ticket, especially if almost all your material is 24/96 or lower resolution, and your player can down-convert anything above that without sounding bad. If your source is relatively high in jitter and electrical interference is not a big problem, async USB is great, particularly with Mac or Linux, where the OS has drivers for USB2. With a low-jitter, low-noise source, coax might sound best.

 

Which is, I guess, an especially long-winded way of saying "it depends."