Hi !
I see that spdif coaxial and toslink are the more common ports for digital transfer.
Some professional units have instead aes/ebu ports.
Are these 3 options completely equivalent or one is better than the other ?
I understand for instance that toslink is more prone to introduce jitter ?
Thanks a lot indeed, gino
I see that spdif coaxial and toslink are the more common ports for digital transfer.
Some professional units have instead aes/ebu ports.
Are these 3 options completely equivalent or one is better than the other ?
I understand for instance that toslink is more prone to introduce jitter ?
Thanks a lot indeed, gino
Depends on what you mean by better. S/PDIF-TOSLINK is optical so better for isolation, sounded ok to me but maybe a bit sterile, just not sure jitter was responsible. Actually i cant see how TOSLINK could introduce jitter . Coaxial is great for long runs, sounded fine too but one time i had PC noise issues, optical got around it until i fixed the actual problem. Then i went back to coaxial. AES is the pro version of S/PDIF, not sure if any difference sound wise but it can carry more info (2 channels). You into studio work?
Hi and thanks a lot indeed for the very helpful reply.
Sound quality is important. In particular I have doubts about toslink.
I heard that glassfibers sound better than plastic, that a little of optic coupling gel is needed on the cable terminations ... in a word a mess.
I will go with aes/ebu then.
No no ... I have just a dac with all three inputs available that I would like to use with the pc.
I am selecting a usb to spdif interface, and I need to know which kind of output to look for.
Thanks a lot again, gino
I go optical nowadays. That wire sure is tiny. And you're guaranteed electrical isolation unlike some coaxial sources which cut cost by not having output transformer.
Hi and thanks for the interesting advice.
Problem is the build quality of this port usually plastic and very poor.
Cables terminations can be ok
but the ports on the dacs are usually very bad
My insufficient understanding is that:
there is "consumer grade" TOSLINK devices and then there is professional quality devices for HiFi enthusiasts.
I think you need a theory in terms of how many conversions are in your chain*. For example, I think USB is "native" to a laptop and therefore outputting using a USB port does not introduce a conversion.
My USB signal goes to a Breeze box that outputs AES (one conversion albeit all digital) that goes to a DSP (Behringer, like many other people) and that signal skips right into the middle of things remaining digital.
The Behringer DSP converts the digital signal to analog and outputs to my several amps.
I can't say if my understanding is correct, and I hope others will comment on the theory as well as specific errors.
BTW, often very helpful to have ground isolation between boxes, such as optical cable provides. But the isolation and low-level transformers I've tested are amazingly flat and good, even really cheap ones.
Ben
*and how many times your signal gets re-clocked (or rather, struggles to make it down the chain using the initial time signal since it is better to have asynchronous boxes that re-clock cleanly)
there is "consumer grade" TOSLINK devices and then there is professional quality devices for HiFi enthusiasts.
I think you need a theory in terms of how many conversions are in your chain*. For example, I think USB is "native" to a laptop and therefore outputting using a USB port does not introduce a conversion.
My USB signal goes to a Breeze box that outputs AES (one conversion albeit all digital) that goes to a DSP (Behringer, like many other people) and that signal skips right into the middle of things remaining digital.
The Behringer DSP converts the digital signal to analog and outputs to my several amps.
I can't say if my understanding is correct, and I hope others will comment on the theory as well as specific errors.
BTW, often very helpful to have ground isolation between boxes, such as optical cable provides. But the isolation and low-level transformers I've tested are amazingly flat and good, even really cheap ones.
Ben
*and how many times your signal gets re-clocked (or rather, struggles to make it down the chain using the initial time signal since it is better to have asynchronous boxes that re-clock cleanly)
Last edited:
Hi thanks for the valuable advice.
I guess we can also say that the fewer conversion stages the better ?
Which transformers have you tested ? are on the digital signal or analog ?
Thanks again, gino
I'd rather not reply in detail because I'm no digital expert.
I tested some Pyramid analog isolation transformers that are mostly sold for medium-fidelity car audio systems. No more than about $12 for a stereo box (two transformers) with RCA-jack hardware on eBay.
Using REW test software, the two transformers were identical. They were flat within a dB or two 15-18kHz and except for the lowest frequencies (where your woofer has 5% distortion), the total harmonic distortion was below .1%. I was surprised. Good enough for me.
No need to take my word for it, see graph.
I had a little bit of audible hum in the modest audio system at my vacation home. Transformers fixed hum completely.
Ben
I tested some Pyramid analog isolation transformers that are mostly sold for medium-fidelity car audio systems. No more than about $12 for a stereo box (two transformers) with RCA-jack hardware on eBay.
Using REW test software, the two transformers were identical. They were flat within a dB or two 15-18kHz and except for the lowest frequencies (where your woofer has 5% distortion), the total harmonic distortion was below .1%. I was surprised. Good enough for me.
No need to take my word for it, see graph.
I had a little bit of audible hum in the modest audio system at my vacation home. Transformers fixed hum completely.
Ben
Attachments
Last edited:
In practical terms, these 3 signal often end up at the same receiver chip. The Toslink and SPDIF signals are the same, just transmission is different. AES is almost the same, but usually at a slightly higher voltage and balanced.
So many receiver chips will handle AES or SPDIF straight in, or the electrical signal coming from the Toslink optical pickup.
Consumer level optical isn't always the best, you can tell by looking at the eye pattern of the signal.
So many receiver chips will handle AES or SPDIF straight in, or the electrical signal coming from the Toslink optical pickup.
Consumer level optical isn't always the best, you can tell by looking at the eye pattern of the signal.
Thanks a lot indeed for the very interesting and helpful advice.
I have decided to use balanced connections to get rid of some noise, i do not know really if this is a very sane solution or not.
So i would go balanced also for digital even if the situation could be very different than with analog signals.
Thanks again, gino
For a balanced impedance connection:
a.) the output of the Source should have an accurately balanced impedance and
b.) the input of the Receiver should have an accurately balanced impedance.
The cable is usually assumed to already have sufficiently accurate balance of impedance.
Read W.Jung, Rane, D.Self.
a.) the output of the Source should have an accurately balanced impedance and
b.) the input of the Receiver should have an accurately balanced impedance.
The cable is usually assumed to already have sufficiently accurate balance of impedance.
Read W.Jung, Rane, D.Self.
For a differential digital signal you need to control the impedance of the complete path for optimum signal integrity...
Yes. but does a slight mismatch of impedances "colour" the sound, as is so often proclaimed by the cable community??
Nope, what a load of rubbish, its a digital signal so slight mismatches wont colour the sound. There is a thread where I did many sims. of SPDIF signals showing various impedance mismatches etc.
Maybe they should invest in this:
https://books.google.co.uk/books/about/High_speed_Digital_Design.html?id=H5SsQgAACAAJ
Read it and then talk facts not myths....
Even with long cables you don't get to many problems, its not high speed digital and to be honest most digital is not controlled impedance near enough will do. I put a couple of sim. shots up yesterday done with a 500mm 50R cable and a 1500mm 75R cable no problem what so ever. If it was a DDR3 interface or similar then I may worry....
Last edited:
It gets funnier when people use BNC cables and connectors "because RCA isn't 75 ohms" and then not realize their China SPDIF interface used 50-ohm.
The reflection occurs at the connectors, so 75 ohm coax connectors should be theoretically the best. Professional uses a separate coax clock cable in addition to the balanced AES EBU signal cable for this reason.
Reflections occur at any impedance mismatch
The reflection is minuscule, probably hard to tell the difference between a 50 and 75R connector on a scope, doing a simulation I could pin point a very small reflection and I had to zooooom in. Consumer with short distances it isn't a problem, for pro gear the distances are often greater so it can be more of a problem.
Attached is a close up of the difference between a 75R and 50R connector, the PDF is the schematic for the simulation.
The reflection is minuscule, probably hard to tell the difference between a 50 and 75R connector on a scope, doing a simulation I could pin point a very small reflection and I had to zooooom in. Consumer with short distances it isn't a problem, for pro gear the distances are often greater so it can be more of a problem.
Attached is a close up of the difference between a 75R and 50R connector, the PDF is the schematic for the simulation.
Attachments
- Status
- Not open for further replies.