♣ "So you find that for final externally cabled data transmission to a DAC, fibre optics and the necessary conversion from electrical to optical and back is the best solution even for short home-based distances of a few meters?"
♦ "At first I thought that this was like shooting sparrows with cannons. But the more I thought about it, the more I liked this idea. Of course it's better. It's the most effective galvanic isolation."
♣ "I understand the tech arguments for optical isolation but always assumed that as in any other data conversion, there must be some losses inside the process?"
♦"These arguments have been around for 20 years but I disagree. If the complexity of data transmission really played such a vital role, why would we have digital audio, streamers and servers with their networks in the first place? If modulating, manipulating, transporting data and especially the error corrections in computers, hard disks, optical disks and other storage media (Reed Solomon algorithms) were such a problem, why does any of it exist? Most importantly, why does it work correctly? But now I must mention a few more things which are mostly overlooked. Very few people ever talk about EMC. Most people think electromagnetic compatibility is written by lawyers just to bother us engineers. But in fact everything is EMC as soon as we transmit electrical signals. Eric Bogatin calls it signal integrity. Signal integrity depends on EMC effects. All of this involves jitter, stability, noise, self interference, skewing, delay, signal speed, ground bouncing and more. Clocking and reclocking enter as well. It's a big topic. Only a few people have any idea how much interference there is and exactly how it enters the signal. All of it is an EMC problem. Few people realize how truly important the ground system of any device is. That's a huge factor. It all starts with the circuit diagram and ends with the PCB design. I know very few engineers who understand that electrical energy doesn't transfer in the conductor but dielectric media.
"One might now advocate that it's best to keep things simple and avoid all complex forms of signal transmission. I'm simply not convinced. No matter which form of signal transmission I choose—whether Manchester encoded, modulated or straight—everything always depends on the care I take in constructing my transmission path. And that's only possible when we know what EMC actually means: signal integrity. That's why I think whether you use S/PDIF, HDMI, Ethernet, USB, Firewire, Token-ring, FDDI or other data transfer concepts, it's all about execution so knowledge and care. In base digital we either recognize a bit or not. There's nothing else. But digital signals also transmit interference. This interference has an effect when we convert digital to analog. It even causes issues farther down the analog chain. But it has no influence at all on digital computing as long as the interference doesn't exceed the acceptable level of digital circuits. Again, it's a clear case of electromagnetic compatibility."
To paraphrase, the bit-perfect argument in generic IT doesn't account for interference because with its built-in error protocols, it doesn't impact computing at all. It's in high-performance audio where the same interference can create audible downstream issues in the analog realm. Those issues are not about corrupted bits.
As we're beginning to see with networking DACs, perhaps fibre-optic transmission will eventually become the top choice for ultra-performance audio to supplant USB, Ethernet copper, AES/EBU and S/PDIF coax? Ed Meitner and John Stronczer have long thought so. Ditto Wadax with their optical Akasa interface. Now we're back on topic.
Just then fs.com/uk sold a 20m OS2 single-mode fibre-optic cable for £10 whilst sundry Cisco or Juniper SFP modules wanted ~£25/ea. Let's hope audiophilia doesn't invade this space to apply its usual gold plating. Imagine needing 20km of fibre for a broadband installation and being billed Nordost Odin coin. We'd all be back to Internet-over-telephone bandwidth.
Finally, should we only have a single SFP port at our disposal to make running actual fibre impossible, we can still exploit an RJ45-to-optical cartridge like the one at left. With now 3½ pages expended on foreplay, onto aural consummation. Did I get any?
In a recent review intro, engineer Mario CanEver said the following: "I have in production the ZeroUno SSD-HPA and ZeroUno Plus, the first based on transistors, the second on tubes. Both can work as DAC or preamplifier because they implement an analog board to accept analog inputs. Alternately they can be ordered with just digital inputs. We have fans of the solid-state ZeroUno SSD and fans of the triode Plus. The sound is different but a matter of feeling only..." I italicized the relevant bit for emphasis. Though a proper engineer, Mario freely acknowledges that whilst circuits may measure identical—his qualifier 'only' suggests as much—our experience of them can still differ on a feeling level. That transcends the famously rigid attitude of, if it measures the same it sounds the same. From today's intro we already know that we're not playing in the realm of corrupted bits or signal dropouts. On that score, nothing changed. Yet something did. How to describe it depends on our vocabulary. Some writers invoke 'more humanity' to distinguish a sound that seems less mechanical. Others like 'more organic' as though to suggest fewer pesticides, emulsivers, homogenizers and other equivalents of modern food processing for longer shelf life. With my vocabulary and trigger points, I'll point at more flow from less subliminal reluctance; more ease from less micro grit; and a more relaxed than tense delivery. All of it suggests a small gestalt shift that's not found by dissecting each musical bar but rather, inquiring into how music moves temporally. I like the visual of two 100m Olympian sprinters who cross the finishing line at exactly the same time. On that score it's a perfect draw. The measurement brigade exits the picture content that there's no more to it. Yet watching both athletes in motion, we might like one better. Perhaps their body moved more elegantly. Perhaps rather than tense and competitively concerned, they seemed celebratory as expressed in Usain Bolt's famous smile when looking back at the rest of the pack. Now the runner who moved more fluidly so less choppily and enjoyed herself more is my winner. If you preferred the other, no argument on my part.
EMM Labs' Optilink | Bel Canto Black's ST-fibre links
Such distinctions are entirely entrusted to our subjective perception. We must inquire into how we feel; or rather, how a particular playback presentation makes us feel. If our listening MO doesn't include those inner sensors and observational sensitivity, we needn't consider the fibre/copper axis at least not in my rural conditions of minimal external interference. I can't possibly know to what extent the delta could magnify in a dense urban environment of stacked WiFi networks and hustling Bluetooth roamers whose totality creates far more UHF noise than we see on our tractor-heavy estuary road of a few sporadic residential neighbours and farms. Now the difference might require far less discernment. My big-city days are simply long past. I've joined the Country bumpkin Club. Yet even here I favoured the short fibre link over equivalent CAT7a.
Feeling into my body/mind state rather than looking outward at the inanimate hifi hardware, I sensed that with the optical spur in play, I was more relaxed and involved. Like the Egyptian after-death belief of having one's soul weighed against a feather, it's simply the antithesis of any coarse business. How much we care shall depend entirely on just how tuned into the whole playback process we are; and what we expect from it. I suspect that the few current sightings of fibre optics in networked audio will not only grow more numerous but expand to server/transport ⇒ DAC feeds even for just local files as is my majority application. [Wadax Akasa version left and right.]
Cen.Grand's pending launch of their GLS1.0 server in fact could well introduce just such an example. We're promised a world's first Intel-based audio interface that even supports high-rate multi-channel DSD and includes a discrete clock line like I²S. I also appreciate just how thin and cheap fibre-optical wire is whilst lending itself to very long perfectly non-lossy runs to transmit digital signal from A to B. Whenever our household's nomadic alarm goes off again to trigger another move, I'll be looking to replace our existing copper looms between router/modem and computers with more invisible fibre optics whilst harnessing their superior noise isolation. That seems like a complete no-brainer. So, kudos to COS Engineering for investing R&D into applying this IT tech to fine home hifi.
