[flecom]

Quote:

Originally Posted by eric343 Well, guess it's time to try using a single-mode fiber...

wont work, single mode is designed for lasers, not LED's

i doubt you will get enough light on the other end for the signal to actually get there

also i doubt how there can be a difference really between the glass and plastic, the speed of light is the speed of light...

the only thing i could see would be a poorly polished end connector... (do they polish audio fibers?)

anyway i wont comment much as the only fiber i work with is single/multimode fiber for data networks

 

[dshea_32665]

I mean no disrespect to those who know a lot about all the technical computer science and physics information describing that there can be no difference between the two. I really don't. I am enjoying your discussions very much and learning a lot.

I am just curious. For those who say there can be no difference, have you actually compared the two and listened for yourself? This isn't meant to be an antagonistic question and I am not intending to devalidate your discussions, perhaps there are other elements which affect the sound. I am just genuinely curious.

After listening for a couple of days, there is a clear difference between the two on my particular setup. It is exactly as Tomek and others have said previously. The sound is not identical and is definitely not a placebo affect. I get a slight digital headache and fatigue listening to the glass toslink after awhile. This always happens to me when I change a variable in my setup to make it more detailed/digital. It is why I can't listen to the AD8620 opamp in my DAC.

dshea

 

[JeffS]

I doubt too many would take your suggestion/question with disrespect.
I've used both expensive glass, and cheap plastic for interconnects years ago when recording to MD. Results were identical. Some may argue that this is because it's being converted to a lossy format even if it is a high bitrates...

What I would like to see, and I don't know if anyone on here has the access to the right equipment, is an actual measured test of both kinds of fiber.
Since I'm not working at my old job I no longer have access to fiber splicing and testing hardware. I'm curious if there is signal degredation between the two choices.

I'd also be interested to see the wave form output from the DAC with both glass and plastic.

Between these two tests, I think a difference could scientifically be established. Any thoughts?

Does anyone have the means to do any of the above testing?


dshea, I'm not saying that you aren't hearing a difference. I am questioning what actually is the difference. As I've said before, if you've got a sound setup you're happy with, congrats. After all, when all is said and done, it really is about the enjoyment of the music.

-Jeff

 

[gaboo]

Anyone with equipment to do that kind of study, would probably publish it an AES journal or conference, not on head-fi...

FYI gigabit ethernet had to take DMD induced jitter into account. But toslink uses LEDs so it shouldn't suffer from it...

 

[dshea_32665]

Thanks for the clarification Jeff. I wrote that message after reading the blood bath on the amps forum, so I wanted to ask my question in the way it was meant, with no chance of being misinterpreted. I am fascinated about what all the variables are for different sounds we are hearing, so I have been enjoying reading what people have to say wrestling with all this stuff.

dshea

 

[eric343]

Quote:

Originally Posted by flecom wont work, single mode is designed for lasers, not LED's i doubt you will get enough light on the other end for the signal to actually get there

I will have to do some experimentation, but I think that it's doable.

 

[flecom]

Quote:

Originally Posted by eric343 I will have to do some experimentation, but I think that it's doable.

its doable, yes... but it would involve lasers, LED light will simply not go more than maybe a foot through a single mode fiber if even that

 

[flecom]

Quote:

Originally Posted by dshea_32665 I mean no disrespect to those who know a lot about all the technical computer science and physics information describing that there can be no difference between the two. I really don't. I am enjoying your discussions very much and learning a lot. I am just curious. For those who say there can be no difference, have you actually compared the two and listened for yourself? This isn't meant to be an antagonistic question and I am not intending to devalidate your discussions, perhaps there are other elements which affect the sound. I am just genuinely curious. After listening for a couple of days, there is a clear difference between the two on my particular setup. It is exactly as Tomek and others have said previously. The sound is not identical and is definitely not a placebo affect. I get a slight digital headache and fatigue listening to the glass toslink after awhile. This always happens to me when I change a variable in my setup to make it more detailed/digital. It is why I can't listen to the AD8620 opamp in my DAC. dshea

honestly i have never used optical toslink until tonight actually

either way, no, i personally have not compared plastic/glass

is it possible that it makes a difference, of course, quite literally anything can make a difference, who knows maybe your glass fiber has a better connector and it closer to the LED/sensor?

but really i doubt that glass/plastic would make a differece...

but like i said, i base that only on my knowledge of fiber from data networks

is glass fiber better than plastic? of course, there is no plastic fiber used for data communications for a reason, but its mostly distance limitations

 

[roadbuster]

The notion that glass vs. plastic interconnects for the
transmission of digitally-encoded linear PCM audio signals
has a deterministic, filtering effect is preposterous. It's
simply not a believable idea that the distortion effects
between the two materials would lead to such an effect.

Consider three distortion effects on signals driven over
an optical cable: jitter, bit errors, and material-dependent
frequency response. Let's start with the last concept first.
The frequency response of any optical transmission medium
is going to be non-linear after some point. What is that
point for plastic and glass cables? Somewhere around
400MHz/km for shoddy plastics, and 2GHz/km for half-decent
glass. Given that your S/PDIF signal might be at 192KHz at
the most, and that the maximum length of cable that any of
you are using is, maybe, 10 feet, high-frequency roll-off is
not a meaningful distortion mechanism for digital audio signals.

Now, let's look at the source of error mentioned first: jitter.
To avoid confusion, let's define what jitter really is. Jitter
refers to the variation in a signal's phase relative to their
ideal position in time. For example, if an edge of a square
wave was supposed to arrive at time T, but actually showed-
up at time (T+t), then the jitter is 't.'

How much jitter is tolerable? Jitter is tolerable to the point
where the data is out-of-sync with the sampling window at
the receiver. How big is this sampling window? Ideally, it's
[ 1 / 2*sample frequency ]. Given a 192KHz signal, the
sampling window is 2.6 microseconds.

Ok, so we can shift any sample by up to 2.6 microseconds
and still be safe. How much jitter are the two mediums
introducing? For high-frequency (>1GHz) signals, I've seen
figures around 5-25ps for multimode glass fibre transmission.
Is anyone here convinced that the jitter introduced by plastic
would be 520 000x worse? I'm not.

Assuming arguendo that jitter and high-frequency roll-off are
actually distorting the signal to the point that the signals are
being interpreted incorrectly, is it possible that the errors in
transmission would lead to a change in the tone of the audio?
Given that the data being transmitted is the signal in the time
domain (linear PCM) and not in the frequency domain, how
would the bit errors have to occur in order to, say, increase
the bass of the signal? The errors would have to occur
periodically, and affect the same bits in each sample to
prevent violent modulation in the volume of the audio. Is this
conceivable? I hope not.

 

[gaboo]

Toshiba engineers must be perfect idiots when they publish the jitter of their own transmitters and receivers at 15-25ns.

IMO most of the trouble comes from the interfacing, not from the fiber itself.

 

[gaboo]

Quote:

Originally Posted by roadbuster How much jitter is tolerable? Jitter is tolerable to the point where the data is out-of-sync with the sampling window at the receiver. How big is this sampling window? Ideally, it's [ 1 / 2*sample frequency ]. Given a 192KHz signal, the sampling window is 2.6 microseconds.

Jitter is considered audible around 10ns (but the threshold varies with frequency). If the receiver lock its clock to the sender, then that's the threshold for the transmission line too. Shameless plug.

 

[roadbuster]

Quote:

Originally Posted by gaboo Toshiba engineers must be perfect idiots when they publish the jitter of their own transmitters and receivers at 15-25ns. IMO most of the trouble comes from the interfacing, not from the fiber itself.

Why would they be idiots? It's necessary for a manufacturer to
publish relevant operating specifications so the engineers can
design a functional product.

Just to be clear, those jitter numbers are for the drivers and
receivers, not just optical adaptor. For those devices, there's
jitter introduced in the electrical->optical->electrical conversion
by the transmitter and receiver, and that's where the "large"
jitter numbers are coming from.

You are right when you say that most of the trouble comes from
the interfacing. I spent a summer working on opto-electrical
converters for intra and inter-silicon communication, and the
delay & jitter added by the conversion is a real pain.

Quote:

Jitter is considered audible around 10ns

It might very well be the case, but any relevant digital audio
player has a buffer to decouple playback from the input bus. Thus,
small-scale variations which don't violate timing constraints do
not affect the audio playback.

Quote:

Shameless plug.

Great collection of links. Thanks.

 

[gaboo]

Quote:

Originally Posted by roadbuster It might very well be the case, but any relevant digital audio player has a buffer to decouple playback from the input bus. Thus, small-scale variations which don't violate timing constraints do not affect the audio playback.

Being a computer geek myself, that's what I assumed at first.
Boy, was I surpized...

 

[tunes]

I think the sound difference is due to the different refraction coefficient between glass and plastic. This results in different jitter, despite the almost nil bit error rate.


Remeber: adio (from the DAC) = bits + clock. Plenty of info here: http://www6.head-fi.org/forums/showthread.php?t=90069

See posts below. At least with Chord DACs there should be no change in SQ between glass or plastic as long as they are high quality.

Originally Posted by tunes View Post

Can Rob or anyone else with experience tell me if there is an audible difference between a high quality plastic versus glass TOSLINK to mini cable. Seems there are many different opinions with some saying it is not possible to get differences since any signal degradation would just prevent any modulation to get accurate sound. I don't want to waste money on glass if they are more fragile and offer no sound quality enhancement over plastic for a short cable run.

So Lifatec Glass Toslink Cable versus Sys Concept
SEE the explanation below as to why plastic may be better.

Concerpthttp://www.sysconcept.ca/product_info.php?products_id=364&osCsid=2h2pduirapjbckjecublgmo250

I would love to know what actual Toslink Rob uses or prefers to use in his spare time!
@Rob Watts

I'm guessing the answer will be 'Chord's optical cables of course'
Bog standard plastic - except for 192 kHz, when I use very short bog standard plastic!

My DAC's are intolerant of jitter, and also intolerant of optical's asymettric rise and fall times - the SPDIF rx in the FPGA automatically calibrates for this. So long as it is bit perfect, I have not been able to hear any difference.

I have bought Kabel Direkt - excellent service.