1. Register NOW and become part of this fantastic knowledge base forum! This message will go away once you have registered.

A Word Clock Cannot serve two Masters, "Innie" or

Discussion in 'Recording' started by Hawkeye, Sep 9, 2005.

  1. Hawkeye

    Hawkeye Active Member

    I'm waiting for delivery of an A/D card for my Focusrite Trakmaster which I will output via SPDIF to my Edirol DA-2496. I got the card cheap and want to experiment to see which converters are better. Yes, yes, I know I should have better preamps but I'm focusing on converters for the moment as that is what my current two options are. On the preamp front, I have an ART Pro MPA on the way that I snagged on Evilbay.

    Question? Should I

    A: Make the Focusrite A/D (the digital sending source) the master clock, and make the Edirol DA-2496 (the receiving digital device) the slave to the Focusrite master clock. I would set the "Clock" to "External" on the front panel. or...

    B: Should I make the DA-2496 the master clock and connect a BNC word clock cable in addition to the SPDIF from the DA-2496 to the Trakmaster.

    My understanding is that I should make the Focusrite the master because it is the only other digital device besides the DA-2496. I think I also only need just the SPDIF cable conncected between the devices then and not also the BNC Word Clock cable.

    I will be tracking from:

    Trakmaster > SPDIF out > DA-2496 SPDIF in ch 7-8
    Pro MPA analog out > DA-2496 analog line in ch 5-6

    Any heard the differences or have advice one way or the other?

    Thanks,
     
  2. Opus2000

    Opus2000 Well-Known Member

    Typically word clock is the better way to go but it really depends on the clock source and how well it's buffered and designed.

    Yes, the Focusrite should be the master. Follow the path basically...

    Anyway you look at it it won't matter what you use as the clock source as you'll most likely get the same results whether you use SPDIF or Word Clock.

    Paying attention to converters is a very important aspect of recording so don't think that you are doing anything wrong!

    Opus :D
     
  3. Hawkeye

    Hawkeye Active Member

    Tick tock

    Thanks Opus, I appreciate your explanation. I am ordering a Canare 75 ohm SPDIF cable. I'm thinking I should order a second with the BNC connectors for the word clock and if I understand correctly, setting the Trakmaster A/D as the Master I can try connecting SPDIF only or SPDIF + Word Clock.

    Correct?
     
  4. Opus2000

    Opus2000 Well-Known Member

    Yes,
    You can connect both...set the Edirol to clock from word clock and then within the recording application select the SPDIF input as the source...

    This way the device is clocked via word clock and SPDIF just simply sends audio and the clock embedded within the SPDIF signal is ignored!

    Cheers

    Opus :D
     
  5. Hawkeye

    Hawkeye Active Member

    THanks again Opus2K. I've ordered both a SP/DIF RCA cable and a word clock BNC cable. I'll try both and see which sounds better (if there is indeed a difference that I can tell, given the quality of the converters I'm using). If jitter is lower with one method vs. the other it should be audible IF that difference is significant.
     
  6. Opus2000

    Opus2000 Well-Known Member

    One can not truly "hear" jitter. What you heard before was clock error since the Edirol was set to internal mode and so was the 428...

    If you do hear a difference I would be surprised since both clock sources are coming from the same box, the 428.

    Good luck

    Opus :D
     
  7. Hawkeye

    Hawkeye Active Member

    Agreed, you cannot hear jitter on its own. As jitter is the misalignment of sample data (measured in little teeny picoseconds) that occurs only in the digital domain, I guess I was referring to that oh so subjective perception of whatever sounds better.

    Thanks
     
  8. ghellquist

    ghellquist Member

    Some thoughts.

    1 - very sweeping statement.
    You always get less jitter if you run from the internal crystal oscillator. (the sellers try to make you think differently, but then that is their way of making you haul out the plastic). Broken or downright bad stuff might be an exception. If it sounds "better" one way or another is a completely different question which includes the listeners perception. I have yet to meet any engineer that even knows hos to measure jitter and would be even more surprised if he has the equipment. (Show me wrong there and I will be happy).

    So the short take is:

    2 - when tracking put your most important signals through your best AD. Clock the AD internally and clock everything else from there.

    3 - when you only listen (and use a soundcard with built-in DA), set that as master.

    4 - using lighpipe as clock source works quite as good as using word-clock cable on BNC-contacts.

    5 - don´t ever chain a word-clock signal by going "through" the equipment. Use a T connector instead.

    6 - use the right kind of cable for WC (it should be 75 Ohm).

    7 - preferrably use a T-connector and termination resistor on the last leg of the word-clock chain. (Or if the unit has switchable termination, enable it on the last unit ONLY). All specifications expects this termination, but calculations usually shows that it is not strictly necessary for short WC cable lengths. But why kill the gain of WC by taking chanse?

    And finally, do exactly what you are planning to do - listen. Your ears tell you what you find sounds best and that´s it. Everyone else in the world might have a different opinion, but what you hear you hear.

    Ooh, and by the way, share your experiences. They are true as they are yours. But avoid the pitfall of jumping to conclusions as to the underlying causes, often enough things have more behind than what first meets the eye, sorry the ear.

    Gunnar
     
  9. Opus2000

    Opus2000 Well-Known Member

    Lightpipe is NOT as good as a word clock, SPDIF or AES in any way shape or form....

    Lightpipe by nature has more jitter than anything due to the light reflections within the cable. This is a known fact.

    Jitter can be measured indeed but you're right most engineers at a "studio" will not have one of those machines! You'd need a Audio Precision Machine or a Jitter tester.

    Indeed a T Connector is good but only when you are clocking multiple devices.

    If the external clock source is much cleaner than the sound card then you would want to use the external clock as that will help the DA sound even more so. Each design of the internal clock, whether they use a crystal or not, is different and may not have as good a buffering stage as another.

    Opus :D
     
  10. ghellquist

    ghellquist Member

    Hi Opus,
    I am learning things all the time in this area. Maybe you could help me by adding a bit to your earlier answer.

    >>Lightpipe by nature has more jitter than anything due to the light reflections within the cable. This is a known fact.

    Could you walk us through the math there. I cannot make it out really, as the signal travels with lightspeed inside the cable. I would believe the maximum difference in signal propagation, or the jitter, to be less than double cable length divided by the speed of light in the cable. What have I missed?

    >> Jitter can be measured indeed but you're right most engineers at a "studio" will not have one of those machines! You'd need a Audio Precision Machine or a Jitter tester.

    As you seem to know, what brand and type of machine have you been using? As it is discussed quite often I would not mind renting some equipment and do some real world tests.

    >>Indeed a T Connector is good but only when you are clocking multiple devices.

    You loose me here. I have been taught that all word-clock cables needs a termination when the cable is long enough. How can I terminate the cable without a T-connector, unless the termination is built-in to the sound card? Maybe you can walk me through the math at how long the cable can be without introducing noticeable problems?

    >> If the external clock source is much cleaner than the sound card then you would want to use the external clock as that will help the DA sound even more so. Each design of the internal clock, whether they use a crystal or not, is different and may not have as good a buffering stage as another.

    You are probably right. But what effect does the PLL in the sound card have on this?

    Gunnar
     
  11. Opus2000

    Opus2000 Well-Known Member

    Audio Precision makes a software/hardware testing equipment that can measure THD for AD and DA as well as FFT tests and other stuff.

    http://www.audioprecision.com/index.php

    Prism makes a similar box but can do jitter testing through AES. I like the Prism boxes!

    http://www.prismsound.com/test_measure/test_products.html

    The main jitter tester name eludes me for the moment. If you do a google search for word clock jitter tester there are several brands..of course not cheap either!

    When I was working for Apogee as their technical support rep I had to do many testings and researching on jitter and so forth.

    I'm not a math guy in any way so I'm not going to go that route. Math really isn't needed for lightpipe...as I mentioned reflections within the lightpipe cable cause the jitter...length of cable can definitely cause more jitter of course. If you're going to use lightpipe as the source the best way is to use glass fiber and not the regular plastic cable. Regardless of the cable, lightpipe is not a solid clocking source as any other digital source.

    Indeed, you are correct in that if the device is non terminated you should terminate it...sometimes this eludes my brain but to be honest if all you are doing is one device it's not going to be the end of the world if you don't terminate it.

    If a sound card is clocked externally it will use the clock from the external device and not it's internal clock. Thus bypassing any PLL or crystal it may be using...and can improve the overall stereo imaging and sound quality. Again, this really depends on the quality of clock and it's design in the long run.

    Also remember that a PCI card with converters inside a computer are more prone to have interference than an external box.

    Again, from my past experience and from talking to more experienced techs, optical is not as good as other digital formats and word clock itself.


    Opus :D
     
  12. Opus2000

    Opus2000 Well-Known Member

    Another note:

    If you do some tesing you'll find that doing THD testing will show you these results more so than jitter...

    If you rent an Audio Precision(AP 1 or 2) do a AD test and have it record it and you'll see that with optical it will vary more so than SPDIF or AES. This is due to the nature of lightpipe...

    Opus :D
     
  13. Hawkeye

    Hawkeye Active Member

    I second the idea that jitter is indeed measurable. Many of the audio magazines publish jitter specifications for D/A converters, CD Players, and there used to be a whole market for jitter reduction devices that would sit between the CD disc drive's SP/DIF output and the D/A converter's input (Sonic Frontiers Ultra Jitterbug, Monarchy Audio etc.).

    There has been a correlation reported between low jitter and better sonic characteristics. I tend to believe this.
     
  14. ghellquist

    ghellquist Member

    I´ll work my way through measuring jitter. It is slowly dawning on me that things are not quite as simple. But you lost me on two things, could you please help explain.

    I was under the impression that lightpipe was quite OK for clocking, but if measurements show differently I am accepting it at once. I cannot in any way make sense of your explanation though.

    >> ..as I mentioned reflections within the lightpipe cable cause the jitter...

    Could you help me understand how this happens.

    >> If a sound card is clocked externally it will use the clock from the external device and not it's internal clock. Thus bypassing any PLL or crystal it may be using...and can improve the overall stereo imaging and sound quality. Again, this really depends on the quality of clock and it's design in the long run.

    You have to help me here as well. I was under the impression that the PLL is only used when you connect an external word-clock. I thought that all modern sound-cards use overclocking, often 64 times, and that to reach that you have to synchronize the internal clock to the external word clock and this is done using a PLL in most cases. Hence the word-clock is not used to directly trigger the sampling. And I was told that PLL-s are notoriously difficult to get right. And by comparison, creating a very stable signal from a fundamental mode crystal inside the box is childs play.

    Funnily enough people like Dan Lavry and Bob Katz seems to say time after time that the best way if possible is to run from the internal clock, while the sellers of Big Ben say something different.

    Gunnar
     
  15. Opus2000

    Opus2000 Well-Known Member

    Think about light...if it reflects off of a surface that is blocking the main target it doesn't get to all the places it was aimed at...thus parts of the light are missing...minutes it may be within the cable, but enough to cause some slight slew in the data...thus giving a slightly less reliable clock. It's just a bad idea to truly clock from optical. Again, if that's the only clock source you can use, then it must be used. If you have the chance to use word clock...do indeed use it.

    It really depends on the design on the converter and clock circuitry. Also depends on what TTL it goes through as well. Yes, indeed it may go through a crystal but basically what will happen is that the external clock may be cleaner than what the design of the device being fed the clock thus giving you less of a chance that a poor clock design will degrade it...

    As I'm not an electronics designer I can't really explain it in terms that would truly make sense! Indeed what Lavry says is true as well as Apogee..confusing?! YES!

    Everyone has their own opinions in this argument. The Big Ben is truly a remarkable device due to the design on the clock itself. It's truly one of the cleanest clock sources you'll find out there...almost unmeasurable in the jitter realm...feed that to anything and indeed you will notice a difference.

    The bell curve(thus why the name Big Ben-since Big Ben is the CLOCK in the tower!) is soooo smooth that it makes the difference in the long run. It doesn't deviate from the curve or jump around like most PLL based clocks.

    More later as I have to go back to my 4 month baby boy!! :D

    Opus :D
     
  16. droid

    droid Guest

    I was thinking about buying an external word clock unit, instead of using the internal one in my MOTU 24I/0 is that something that will make my recordings sound better or tighter, I heard it was a good piece of gear to upgrade, any comments?

    andrew
     

Share This Page