digital coax or toslink

Discussion in 'Hi-Fi Stereo Systems & Separates' started by kurtholz, Mar 22, 2006.

Tags:
  1. kurtholz

    kurtholz
    Guest

    Products Owned:
    0
    Products Wanted:
    0
    Ratings:
    +0
    Hi all

    So, just curious what the general consensus is on using the digital coax or toslink for a digital connection, which is better, my opinion is the toslink is a bit clearer connection on my arcmdv29, but the toslink cable i am using is better quality than my digital coax, though it is a kimber hero so not a junk cable, maybe going crazy on one cable would offer a better end result?

    thanks

    Kurt
     
  2. chrisor

    chrisor
    Standard Member

    Joined:
    Nov 5, 2005
    Messages:
    41
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    8
    Location:
    Melbourne
    Ratings:
    +3
    I too, would be very interested in finding an answer to this good question. I have used the optical path for both my HD television set-top box as well as my DV78 connectivity to my AVR300. The logic that I am using is that the information conveyed is in the digital domain and that altering the physical layer in the transport model should not modify the signal provided.

    The answer to this question IMHO will ideally come from ARCAM themselves rather than the subjective positions of listeners (but again this is only my opinion).

    The other possible advantage is that there is no electrical connection between these disparate boxes possibly masking the AVR300 sensitivity to earth-ing differences.
     
  3. mk

    mk
    Active Member

    Joined:
    Jan 1, 2001
    Messages:
    187
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    18
    Location:
    Lake District
    Ratings:
    +6
    I was speaking to ARCAM today with regards to a similar issue where I can't get my DV89 to speak to my AVR200 over digital coax.

    They suggested that I use toslink optical instead (as this already works for SKY+). When I asked about the quality issue he said that they should be identical and you won't hear any difference.

    This gave me confidence to try it at least as I was always under the impression that coax was better quality signal.
     
  4. chrisor

    chrisor
    Standard Member

    Joined:
    Nov 5, 2005
    Messages:
    41
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    8
    Location:
    Melbourne
    Ratings:
    +3
    From my reading on this topic the above is the generally stated position without any supplied technical justification. Getting the manufacturers position on this I think is the key including the technical reasoning will put this to bed!
     
  5. kurtholz

    kurtholz
    Guest

    Products Owned:
    0
    Products Wanted:
    0
    Ratings:
    +0
    actually, i did buy a very nice toslink, my little glitch has gone away to, i would concur, the sound is just as good, but i think the background is even clearer

    i like it, think i will stick with it

    regards'

    Kurt
     
  6. Silverstrand

    Silverstrand
    Guest

    Products Owned:
    0
    Products Wanted:
    0
    Ratings:
    +0
    I have tried with blind testing (using diff. inputs and let the listener switch), coax and TOS of medium standard (60Eur).
    Coax always wins in my setup when listening to detailed new recordings, the only explanation I can guess is that two optical sensors with necessary amp. stages will mess up the bitflow more than a straight wire can. I am sure this will differ between DVD players, and that the AVR does a good job in both cases.
     
  7. Fidelio

    Fidelio
    Active Member

    Joined:
    Jul 26, 2004
    Messages:
    143
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    21
    Location:
    Bristol
    Ratings:
    +5
    Digital Coax will always be superior to a Toslink connection. This is because of the amount of data that can be transfered by Coax is nuch greater than Toslink which may be compressed.

    In the HiFi world there are two optical connections. Toslink and AT & T where the latter is superior because of the greater volumes of data it can transfer. If users are finding Toslink superior to Coax then this may indeed be down to the quality of cable in use. I can recommend Kimber D60 or if you are relly flush Nordost Valhalla digital cable.

    Fidelio
     
  8. chrisor

    chrisor
    Standard Member

    Joined:
    Nov 5, 2005
    Messages:
    41
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    8
    Location:
    Melbourne
    Ratings:
    +3
    Fidelio, Is this your opinion only or can you support this position with technical data?

    If your statement is correct regarding the superiority of Coax, I'm not sure why the larger Storage Area Networks (SAN's) in mainframe computers (that handle terrabytes of data) use optical cabling rather than coax??

    Also, given that the signal is in the Digital domain and that the signal is there or not there, are you suggesting that cheaper cables loose some of the signal?

    Also, why are you introducing products into this thread?
     
  9. kurtholz

    kurtholz
    Guest

    Products Owned:
    0
    Products Wanted:
    0
    Ratings:
    +0
    I would be highly speculative about the NEED for digital coax over toslink,i will never be convinced that digital toslink can't carry enough information thru it's cable, i know the old debate about these two different cables,

    but when i can hear an improvement over coax in my system, all debates go out the window,plus

    with a quick internet search, there is plenty of other reviews sharing the same opinion,

    btw, i was using a kimber hero cable, the toslink is made by a small boutique maker here in usa, very nice quality for the money, also using his silver cables set for analog, a much improved upgrade over the kimber and less money, how often does taht happen

    my humble opinion, but so much cable voodoo out there, i think you should go with what sounds best in your own system

    good luck

    Kurt
     
  10. Fidelio

    Fidelio
    Active Member

    Joined:
    Jul 26, 2004
    Messages:
    143
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    21
    Location:
    Bristol
    Ratings:
    +5
    My opinion is founded on advice from my dealer who is well knowledged in the audio field.

    When I first went for a separate transport & dac I asked what is the best way to couple then. He said At & t because it has greater bandwidth than toslink. Then comes coax then toslink.

    Don't underestimate what cables can do for a system. I have done cable upgrades that have contributed more performance gains than replacing actual components.

    Fidelio
     
  11. Nic Rhodes

    Nic Rhodes
    Well-known Member

    Joined:
    Mar 23, 2001
    Messages:
    17,152
    Products Owned:
    0
    Products Wanted:
    1
    Trophy Points:
    133
    Location:
    Cumbria
    Ratings:
    +1,276

    I am not I buy into this and I use AT and T leads, more correctly called ST all the time. It is not down to the data rate at all, they are all conveying what is in effect tiny amounts of data. The digital audio signal are minute, just look at SD interlaced video in comparison and the rates are something like 50 x the data bandwidth!! and we use the same cables. In fact the coax I use has a bandwidth 10x what is needed for SD video and I use it for digital audio as well. What needs to be addressed is the cables doing what they need to, ie basic engineering principles. This is the usual jitter, screening etc etc. When you do these right (and both are cheap to do) then there really should be little difference between the two methods. For interest optical is translated to I2S and electrical is also translated into I2S. Neither are the native data formats the player uses internally, the SPDIF is just a transport medium, one that is easily screwed up.
     
  12. hedrick@rutgers

    [email protected]
    Standard Member

    Joined:
    May 9, 2005
    Messages:
    57
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    8
    Ratings:
    +1
    It shouldn't matter. In the early days some of the fiber cables were very lossy. But that shouldn't be true now. If I were going to do an unusually long run I would probably use coax, but otherwise not. One advantage of fiber is that because there is no electrical connection in situations where there are ground loops (i.e. hum) it could help. But in a typical situation there's no difference.
     
  13. Ian_S

    Ian_S
    Distinguished Member

    Joined:
    Dec 29, 2005
    Messages:
    7,163
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    163
    Ratings:
    +2,198
    Don't forget that SAN devices often use electronics to handle the signal at either end of the cable that cost more than our entire systems in most cases... :) Most of these devices can generate a signal that will travel several kilometres and deliver bit-perfect data. If you think the back of your system rack is bad, lift up a floor tile in your average data centre and you'll wonder how any of it ever works! :eek: So lack of cable interference is a very useful property in this environment.

    I'm going to ask a really dumb question now I'm sure, but what's an "AT and T" lead??
     
  14. chrisor

    chrisor
    Standard Member

    Joined:
    Nov 5, 2005
    Messages:
    41
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    8
    Location:
    Melbourne
    Ratings:
    +3
    In regard to having equipment that is very heavy to drive optical fibre in SAN's is not correct. These can br driven from PCI cards which if you know about computers would weigh a fraction of a kilogram or a pound.

    Also what you say in regard to the cost of electronics to drive SAN signals is also not correct. PCI optical cards can be obtained for less than $200.

    Hence your inference that it costs a lot of money to attain bit-perfect data transfer is also incorrect.

    See http://h18006.www1.hp.com/storage/saninfrastructure/hba.html for an example.
     
  15. Ian_S

    Ian_S
    Distinguished Member

    Joined:
    Dec 29, 2005
    Messages:
    7,163
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    163
    Ratings:
    +2,198
    Not sure where the 'heavy' thing comes from, don't see any mention of weight in what I posted anywhere... :confused:

    As for the link you gave, well the cards are all in excess of $1000 US, which was kind of my point. There's no way that that proportion of money will be used in your average DVD player, receiver, sky+ box etc. Even at £200 which would be a very cheap PCI HBA, that's £400 end to end, still nowhere near the amount available in home cinema kit design budgets.

    If you look at the cards for larger servers at the link you provide (i.e. UNIX machines that typically get used for large processing) the list price shoots up to around $3800 US, which is exactly my point, referring to the original comment by someone else that why do 'mainframes' use fiber and not cable... Now a mainframe fiber card will cost even more still.

    So with respect, my point about SAN devices costing a fortune (esp. in the mainframe world) was entirely correct even using your link where the prices are still very expensive. :mad:
     
  16. ANDY_DUTTON

    ANDY_DUTTON
    Active Member

    Joined:
    Aug 27, 2002
    Messages:
    498
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    28
    Location:
    Cambridge
    Ratings:
    +87
    Hi,

    There are pros and cons with each of the solutions.

    Coax has a very high bandwidth and can have very low jitter. Its performance is often limited by the implementation of the driver and receiver circuits. Potentially it has very low jitter but if the circuits impedance is not correctly matched (75R drive and 75R receiver) the signal can be distorted. This might sound a very easy thing to do but when you take into account parasitic capacitance on the output and inputs and the impedance of the output driver and the input receiver as well, it is not simply a case of putting a 75R resistor in the circuit. The bandwidth (speed) of the driver and receiver buffers is also a limiting factor.

    Coax circuits are much more susceptible to external noise particularly if the cable used is not a proper well screened 75R characteristic impedance cable. Some so called SPDIF(IEC958) cables are in fact screened twin core which normally has an impedance of 110R, this solution is good for analogue audio but a disaster for SPDIF digital. (it also the standard used for AES EBU the professional version of SPDIF so studios could use existing audio cable runs for digital) SPDIF Digital cables should be coaxial in construction and should be 75R impedance.

    Coax circuits can conduct high frequency noise (from the internal computers) from the source to the receiver. This can cause inter-modulation problems in the receiver.

    Some coax circuits do not provide galvanic isolation. That is DC current can flow from the source to the receiver and visa versa. This can cause hum in some systems. However most systems are isolated via a pulse transformer or capacitive coupling to alleviate this problem. This in itself can cause problems as it is difficult to make transformers or capacitive coupling circuits that maintain the 75R impedance that is required (see above).

    Toslink optical connections.

    Optical connections solve the noise and termination problems as the transmission is via light.

    Optical systems used to be very bandwidth limited and were only just good enough to transfer the signal without it failing. In these early systems the levels of Jitter were quite high and made recovery of the clock quite difficult for the phase locked loop.

    Now the optical systems are almost all capable of 96KHz transmission which means at 44.1 or 48KHz the are inside their operating limit. This is much better but still means that the rise time of the signal is not as fast as that possible on a coaxial cable, so potentially more jitter can be induced by changes in the decision threshold in the receiver circuit or movement in the power rails of the output device.

    Optical systems do have another down side which is that they are quite power hungry and can cause modulation of the power rails with the incoming SPDIF signal. This in turn can cause jitter in the rest of the system as any perturbation in the power rails causes a change in the decision threshold for the internal digital circuits introducing jitter.

    AT&T or ST links use a much higher speed version of the optical solution. This improves the problems with jitter in the link. However the high speed parts use even more power and so the power supply problem becomes even more complex. These parts are also not a standard for audio use so are only available on a very few, generally expensive, products. Where if the power supply problems are carefully considered I am sure they work well.

    So the bottom line is it depends on the specific circumstances your system is in. All systems are now very good and the differences are extremely small. It will depend on the quality of the cable / fiber you use (This quality is not always directly related to the price of the cable. See my comments about impedance). How noisy your environment is and as the changes in sound are to a certain extent a matter of preference, what sound you like.

    If anyone doesn't understand jitter and its effect on the reconstruction of a signal in a DAC. Please take a look at the many articles that describe this on the web. Just because a signal is made of 1 and 0's its not enough to get them to the other end of the cable without error to reconstruct the signal without error.

    For reconstruction without error you require a clock with no timing error on it (no jitter). This clock is recovered from the SPDIF data stream by analogue techniques (in most cases) using a Phase Locked Loop. This process never has no error on it however in very good PLL designs it is possible to get the error effect to be very low. (any system that claims no jitter is not true, I would like to think that they mean jitter distortion is below measurable levels) Once it has been recovered it is still easy to mess it up again if the power supply has noise on it, also every logic gate the signal passes through introduces some noise jitter and if there are impedance miss matches internally in the design this causes more jitter.

    Sorry this is so long but this actually a very simplified explanation of what is going on. (So for those in the know please consider this as a summary not an exhaustive description of the issues).

    Hope this helps,

    Regards, Andrew
     
  17. Ian_S

    Ian_S
    Distinguished Member

    Joined:
    Dec 29, 2005
    Messages:
    7,163
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    163
    Ratings:
    +2,198
    Thanks Andrew, that was very interesting.

    One question as a consequence of your explanation. You state that using 110R cables for SPDIF instead of 75R is not advisable, what about using 75R non-coaxial cable for normal analogue audio? Are there any downsides?

    Regards,

    Ian.
     
  18. Nic Rhodes

    Nic Rhodes
    Well-known Member

    Joined:
    Mar 23, 2001
    Messages:
    17,152
    Products Owned:
    0
    Products Wanted:
    1
    Trophy Points:
    133
    Location:
    Cumbria
    Ratings:
    +1,276
    Andy, an excellent explanation for everyone here :thumbsup: It is nice to see it covering the old and the new and where are now with our better knowledge and more modern chips. One of the best posts we have had here in a long while. :smashin:

    Ian for coax SPDIF stick to 75 ohm CI, avoid 50 Ohm and 110 ohm (unless you are using 110 ohm AES / EBU XLR connections). Ultra high qulity 75 ohm coax is some of the cheapest cable on the planet and generally <&#163;1/m for the best availalable, really no difference in price to the worst available. There is no justification to use anything else.
     
  19. Ian_S

    Ian_S
    Distinguished Member

    Joined:
    Dec 29, 2005
    Messages:
    7,163
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    163
    Ratings:
    +2,198
    Hi Nic,

    What I was asking was whether there are any downsides to using 75 ohm rated normal audio cables given that Andy states that normal audio leads were 110 ohm. The reason I ask, was because I bought some normal phono leads recently and was surprised to see the cable rated at 75 ohm. Given that Andy states a pretty good case as to why using anything other than 75 ohm coax for digital is a bad idea, I just wondered if there were other good reasons why using non-110 ohm cables for analogue audio signals might also not be recommended... :)
     
  20. Nic Rhodes

    Nic Rhodes
    Well-known Member

    Joined:
    Mar 23, 2001
    Messages:
    17,152
    Products Owned:
    0
    Products Wanted:
    1
    Trophy Points:
    133
    Location:
    Cumbria
    Ratings:
    +1,276
    The CI spec is because of things like transmission line effects, reflection etc. These things are basically at high frequencies of digital audio and video signals. At 'audio freq', say 5 Hz to 20KHz then no CI is not relevant and we in the realms of good L, C and R performance from cables. Get stuff that has good CSA, well screened, and good LCR properties etc. This can be done with several types of cables including the coax from digital transmissions but people generally use other cables for things like 'flexibility' but good quality 75 ohm coax will work well for audio as will 50 and 110 ohm stuff. I often use 75 ohm coax for audio. The world is your oyser for audio freq...
     
  21. hedrick@rutgers

    [email protected]
    Standard Member

    Joined:
    May 9, 2005
    Messages:
    57
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    8
    Ratings:
    +1
    There was a question about AT&T. That's American Telephone and Telegraph, which used to be The Telephone Company in the US. (Now it's one of several long distance carriers).

    The ST connector is used with the kind of fiber that carries high speed communications and networking. The cables and drivers typically used for hi-fi applications have a much lower bandwidth than the stuff that carries multi-gigabit networks. I have a feeling that optical connections for hi-fi have unjustly benefitted from the reputation of the much more expensive optical technology used in high-speed networking.
     

Share This Page

Loading...