Fiber optics in high end audio

If isolation in digital audio is so critical to sound quality, why are there so few optical cable systems?

[@InsideOfMyOwnMind]

That strive for compatibility is the exact opposite of what some manufacturers were doing on the 90s where you had components with all kinds of proprietary ribbon cables between them. Sony, Fisher and Sansui come to mind.

[@utube321piotr]

Bob Watts of Chord Audio, who is an absolute expert on digital signals and DACs is of the opinion that optical is always a better method of signal transfer than any USB standard even when seemingly the sampling rate is much higher.

[@Gym-t3l]

I saw an article about, I think it was a student that has developed a program that allows many magnitudes more data to be transferred optically with current lines and conditions.

[@Roosville1]

SFP optical devices with 10Gb bandwidth and a reach measured in miles are now low 10’s dollars each. There are many standard chipsets that will caretake the interface and provide an electrical signal in and out for the equipment. The housing for the device is also standard with application data, EMC gasket and PCB layout available from the supplier. I2S is a Philips standard from around forty years back. The optical connection is a dual “figure-8” fibre (separate TX/RX with a LC connector, a comms standard and available everywhere. Porting the data into packets to transport is the only challenge here, a FPGA Eng will do this in a few weeks. I can’t see the issue.

[@benwu7980]

From a networking perspective, you are right, though they tend to use ASIC's in switches and jitter is not so much of an issue. Doing it on audio with fpga's is going to add quite a cost. I'd like to see it done, but would fear the same thing as Paul said, that it would just be another standard used by one company.

[@Roosville1]

@@benwu7980 I think at this level, the FPGA cost is a one time, basically SPDIF packaged over a standard protocol, within a data wrapper with forward error correction, however the SFP is pretty much data agnostic, you could in theory cheapen the interface and literally shovel I2C or SPDIF over it. However this approach may be over simplified not having an overhead / control chanel. You could do this with an ARM based Raspberry Pi !

[@manitoublack]

Not sure where the issue is. -HDMI 48GB/s optical cables exist and are not expensive. -USB active optical cables exist and are not expensive. -RJ45 to fibre converters exist and are cheap... -WiFi is effectively free. So if you want to use fibre interconnects to eliminate the possibility of ground loops. It's pretty easy and affordable. Especially for long + 10m digital cable runs.

[@granttaylor3697]

For all my requirements TosLink works well, it is a common standard and keeps the noise down between analog digital circuitry.

[@groovypotato-cx7bq]

I appreciate your insight from a business perspective

[@ClassifiedBrief]

There is a trade off and to keep it simple - optical connection will reduce noise as it's true galvanic isolation. However, it Adds jitter because of the conversion to optical and then back out of optical. It depends on your noise and baseline jitter if optical is better. Optical REALLY likes good power, like more than anything I've ever tested. Not just linear power supplies but I'm talking running on pure ultracapitor power will sound substantially better than a linear power supply powering the optical converters and the better the power the lower the jitter in the conversion. Hope this helps others I've done a lot of testing.

[@larazss3254]

Does jitter represent a problem for audio playback when using optical connection for LAN duties (Ethernet data transport from server to a streamer)?

[@Dj-Jon-E-C]

Does not matter the signal sent is in digital it does not respond to analogue it's either on or off. analogue it does not read so any analogue interference does nothing unless it's to the point of blocking the information. Only the anologue side where audio picks up noise etc that was not within the original audio.

[@marcse7en]

I too noticed the flickering light! ... It will likely affect the sound quality of the gear in the listening room! (roll eyes!) Paul mentions his chief engineer ... He should get him to look at that light! It's funny that Paul talks about "light" while standing underneath a faulty one! 🤣

[@shanestephenson8423]

Hi Paul, Andreas Koch of Playback Designs, uses a glass fibre AT&T link between their Dac and their transport / streamer server. And as far as I'm aware, there are a few others that are doing this to, but I agree with you that it is not a broadly adopted medium yet.👍🙂🎧

[@Dawizbuffy]

Didnt use Meitner it first🤔😎

[@stephens2r338]

dCS were the first to use a seperate clock in home audio. They used SPDIF and could only be used by their products. Over the last 25 years a few companys have offered the same option. For exsample Accuphase offered it through an RJ45. However the I2r connection is still the most popular.

[@DaveJ6515]

My external clock is connected to my Rossini dac via a digital 50 ohm bnc cable (actually two: one for each baseline frequency).

[@stephens2r338]

@@DaveJ6515 You're right Dave, me bad. I ment to say BNC.

[@pebbleschan6085]

Blinky light at top right corner! 😂

[@zoltar808]

😰

[@johnnytoobad7785]

Those overhead lights need a "Power Plant" with galvanic isolation.

[@Bassotronics]

It's in need of a new fiber optic transiever.

[@andygrenn680]

Must be the room acoustics…

[@3Cr15w311]

@@johnnytoobad7785 Nah, they just need to be relamped. :)

[@Fastvoice]

I'm surprised that you still use flickering fluorescent tubes as your office lighting. Didn't you convert it to LED tubes yet?

[@tubefreeeasy]

It felt good when I switched my kitchen to LED. I had problems like headaches and the flickering light would affect my psyche in negative ways.

[@chungang7037]

Light transfer function! But that flickering light up top there is likely spazzing out because the room isn't galvanically isolated and noise is running rampant through your workspace, hope you aren't listening to music there, yikes!

[@5thlevelweb887]

I get Paul's point but it seems like a no-brainer to connect the ethernet ahead of a music streamer to an ethernet to optical adapter then an optical to ethernet adapter to take all the noise off the line. Being ahead of the streamer and DAC, their reclocking functions should eliminate any added jitter.

[@joz411no8]

There are no multi-billion dollar companies developing next-level/next-gen standard interfaces that provide superior sound. It's just more likely that people like us who really care about high fidelity audio, are left with choosing the manufacturers we like and matching the components the best we can.

[@TheDanEdwards]

Over the years there have been several attempts at a standard optical interface in the digital world. FDDI was a thing for awhile before it become OBE. Then just a few years ago ODI was proposed by a consortium (including Intel) but I think it has just languished. In the communications world the ITU has a standard (ODN) for high speed networks, though using that for a consumer audio device is overkill yet I am surprised some high-end manufacturer has not tried.

[@davidsicking7514]

To put it simply, fabricating fiber connections, splicing or similar is difficult. It requires polishing the end of the individual light duct. The commercial products in that field are available but are used for patient contact electronics. Analog Devices is one source. By comparison copper wire only takes a solder iron. The optical transfer can most easily be within the audio/digital package.

[@dell177]

They really should update the TosLink optical standard, the physical media is capable of a koy more than the spec allows. Same thing for the Redbook limitation of 441k, had they chosen 48k it would have been better for all involved if the brick wall filter had been further out.

[@s.mikechang4544]

In reality, The telecom or computer network massively use hi-end optical fiber, It's depends on cost. I own a Theta Data II & Theta Generation III connected by ST Fiber, In stereophile recommended class A is depend on ST.fiber. If you use coaxial or Tos-link fiber would be just Class B. And some Wadia Transport & DAC use ST.optical fiber.There are all gorgeous in the past time. So, in these time, naturally HDMI i square 2 is the better way. But dCS still not in the way with it

[@paulstubbs7678]

I have a fibre optic USB extender, seems like a good idea to use on my sound system, however with no way to easily measure jitter, I don't know whether it'll improve or worsen things, so for now it can stay on the shelf.

[@Chiefelgin]

I'm using fiberoptics hdmi to connect to my Yamaha a4a. 10meter perfection.

[@babubabu12345]

Paul Sir correctly answered for this topics.

[@jamesfarrow6752]

Shortly after adding a server to my system, I loaned an ADOT fibre kit and I could tell instantly that there was an improvement in sound quality so I went ahead and purchased one. The kit is connected to my network switch via Ethernet and a cable with SFP connections from the kit to the server.

[@o0Donuts0o]

Sorry but you wasted your money. To pull data from your server to your DAC or whatever you will be using Ethernet. And is transported using packets. Those packets are signed ensuring that the frame data is what was sent. If an error occurs during transport, CRC can be used to correct the errant data. If not, it is retried. You aren’t hearing anything except the sound of wasted money leaving your bank account. I don’t care what you think you are hearing but we have been transporting packets over Ethernet for decades over 1000’s of KM, reliably. Your run of fiber has introduced nothing that hasn’t been done over CAT cable. Packet transport has been solved. Audiophiles parting with money for foolery hasn’t.

[@jamesfarrow6752]

@@o0Donuts0o I disagree. I admit to not being able to argue from a technical perspective but I trust my ears. Also the manufacturer of my server has released a different type of cable that connects between the fibre kit and the server, whicg has improved things further.

[@o0Donuts0o]

@@jamesfarrow6752 Believe what you want to believe but Ethernet is Ethernet and standardized by the IEEE decades ago. Whether you use CAT5/e/6 or fiber, you are getting the same packet, at 1500 bytes each, same header, same data, same CRC. They cannot be different otherwise you will have nothing coming from your server because the network interfaces will drop the traffic. You have bought snake oil.

[@wisehippo3072]

Same here. ADOT has improved the sound quality.

[@wisehippo3072]

​@@o0Donuts0oWho the hell are you to tell him he wasted his money? You don't know his audio system. You sound like one of those guys who are bitter because they can't afford high-end audio gear and who deal with it by insulting every piece of equipment they can't afford and everyone who dares to praise it, let alone spend their own money on it.

[@digggerrjones7345]

While not the industry standard, there are a number of CD players/transports out there featuring I2S via HDMI!!

[@RotatingToroids]

Very honest Paul!

[@mrronenza]

Great video. Thanks. If I will add 1Meter Cyber Cable at the end of the present ethernet cable to my streamer. Will the last short cyber cable consider to be as a Galvenic Isolation ? thanks.

[@ThinkingBetter]

Data is data. Audio data is rather low bandwidth so transferring digital audio is trivial. Best is using asynchronous connection allowing the DAC to master the clock precisely without PLL or any sample rate conversion. USB C is currently the best most recent interconnect between computer and DAC, and Ethernet is best interconnect to your internet router while optical internet to your home is the best way to get no fuss high reliable bandwidth into your home. Problem with TOSLink is that it’s one-way synchronous link and adding the complexity of a dual clock of the data flow potentially yielding jitter. Galvanic decoupling is nice but it was more needed in the past with heavy power hungry DIY Windows PCs. A modern power efficient Mac Mini using Apple silicon or similar is excellent in avoiding noise into the analog part of your system.

[@chungang7037]

USB is great, but many audiophiles seem to harbor memories of noisy pcs from 30 years ago and manufacturers perpetuate the myth because the amount of money on the line would likely sink some many companies outright. But weirdly they don't worry about audible distortion, for example, they worry about what they can't hear but still believe is altering the data signal somehow and affecting the music. If it if feels like a medieval superstition thats becuase it is.

[@ThinkingBetter]

@@chungang7037 Well said. I couldn’t have said it better myself.

[@maidsandmuses]

Unfortunately, unlike Ethernet, the *USB* *audio* connection standard is not asynchronous, but *isochronous* . That means the amount of data transferred in each isochronous data packets _is_ varied to allow buffered synchronisation to the master clock in the DAC, but there is no facility for re-sending of lost or corrupted packets, (which there is for Ethernet protocols). That is why the USB cable quality debate for digital USB audio connections is not the complete nonsense it may appear to be, because it is not a true asynchronous connection with a corrective data integrity protocol.

[@chungang7037]

@@ThinkingBetter Thanks, also I wanted to say you are also right about the mac mini. I've seen it used as the center for an entire home studio for recording music as well.

[@VideoArchiveGuy]

@@chungang7037 It IS audible and is trivial to demonstrate.

[@ford1546]

Has all ps audio products fiber optic and coax for more compatible with things?

[@gotham61]

Bring back the AT&T ST glass fiber connection promoted by Wadia 35 years ago

[@kenp9073]

Still has to be converted back to analog at some point.......so what's the point? Signal is only as pristine as the lowest quality piece of equipment it goes through. Lets switch everything over to Dante!

[@Dj-Jon-E-C]

Digital information is either there or not there nothing from digital end can create any noise to it. It's only when it's on the analogue side that you can pick up noise. Digital only reads digital information. example if I print something on my printer does it matter on the quality of cable? No because as long as it can read the data it will print. If it has interference it wont print the interference. It's a bit like when HDMI cables came out some people would buy the expensive cable thinking it will provide a better picture but no it wont.

[@Pete.across.the.street]

Lol. You are just plain wrong. For the printer yes, for audio no.

[@Dj-Jon-E-C]

@@Pete.across.the.street If the audio going through it digital then what makes it any different? Apart from analogue side I get that can pick up unwanted things.

[@Pete.across.the.street]

@@Dj-Jon-E-C noise

[@Dj-Jon-E-C]

@@Pete.across.the.street That noise would be from the analogue end. Digital is numbers 1s and 0s be it music, video, data etc that's all digital reads as numbers. Analogue noise does not run through digital its not read by it. But if say the cable is really poor then you may get clicks and lost sound due to parts of the 1s and 0's lost. Like digital tv the picture breaks up if the signal is poor, you never see static or noise within the picture or anything else apart from seeing it break up same with the sound interference can cause signal loss. With fibre optic it's only light so earth noise etc it does not pick up. Any interference noise you hear would be before it's converted on transmit end or on the receiver end that is coming out as analogue. If I send analogue noise through digital it wont pick it up, if it does then it's because the analogue side is picking up that noise not the digital. I may be wrong and open for discussion as how we learn but that's from what I know and if I can be proven wrong then fair play.

[@Pete.across.the.street]

@@Dj-Jon-E-C there's digital noise as well other noise in the digital section that is not galvanicly isolated

[@tonep3168]

You can’t so easily get the Audiophools to so easily hand out their cash when fibre is fibre. But a couple of thousand bucks for a metre of copper is an easy target for them. I wonder who the first fool to buy a cable riser for an optical cable will be, and how much they would pay for it? Lol

[@HittingImage]

Audio certified fiber. You read it here first :-)

[@benwu7980]

When there are people buying digital cabling like an audioquest dragon for hdmi (eARC), you can be sure they'd make some fibre if were products that use it. Not slamming on Audioquest, digital is not quite as simple as just out=in=out.

[@Hondaguru1122]

Synthetic snake oil 😂

[@RoderikvanReekum]

I hope PSaudio did get royalties from that hdmi thing.

[@rainiergruber9022]

Before you continue developing that new product you mentioned, please fix that light on the ceiling!

[@ShahidiSabri]

i think the bit rate of the red light fiber optic , is still lagging , for a very high resolution DAC , particularly with a stereo format . maybe by utilizing a purple led light fiber optic , this problem could be overcome .

[@analoghardwaretops3976]

your DAC is the "tortoise"..in front of the fibre optic " hare ".....

[@DJS-pf5pf]

MSB has done it, why not get it from them and start a new standard?

[@anonamouse5917]

Fibre has a benefit if you have very long distance or massive amounts of electrical noise. For 99.99% of audiophiles you'll be fine with a well shielded cable that uses paired differential signaling.

[@BroSweets730]

Can I offer my “theory” on why they say toslink connections are usually a last choice. And why I think fiber optics may not be a good medium for sound as compared to copper wire? Because sound is ultimately vibration. A light wave is just that. Very “light” so to speak. It can’t carry the density of an lish audio signal full of vibration like a metal wire can. Hence why you get a crappier sound from fiber. When you read something like “a single thread of optical thread can transfer as much data as a thousand copper wires” you would probably instinctively think as I did, “wow that would have to make wire obsolete for just about everything!” And then I thought maybe that’s the issue. It’s good for transferring just that “data” and not so much audio signals

[@l0zerth]

It's a nice theory, but when the signal is sent over fiber optic, it's still digital, meaning as long as there is no data loss and you run [good] reclockers and buffers in the appropriate spots, that won't affect anything.

[@BroSweets730]

@@l0zerth yea I know, idk, I can’t put my finger on it yet but I just think there’s something about the medium of light. There has to be. Why do you think a lot of people complain about the sound quality of toslink? I can’t tell either way honestly

[@BroSweets730]

?