Question about DAC's/Jitter (for Jon?)

Collapse
This topic is closed.
X
X
 
  • Time
  • Show
Clear All
new posts
  • Brandon B
    Super Senior Member
    • Jun 2001
    • 2193

    Question about DAC's/Jitter (for Jon?)

    Posted what I thought was a simplified explanation about sound quality in DD/DTS vs. PCM. Curious where I'm off vs. maybe the guy correcting me.

    My post:

    In PCM audio (what CDs use) the bits are sent over the wire, and not only does the value of the bit (1 or 0) determine the signal, but the timing between the bits is part of the equation as well. Since the bits are not actual on/off, but an electrical signal that rises and falls in voltage quickly or sharply, there is some room for error.

    If your digital cable allows interference in, or because of its electrical properties sort of smears out the rise or fall of the voltage over time, the receiving equipment may have difficulty resolving the transition from carrier signal to bit signal at the exact correct instant to interpret the signal without any distortion. In effect this jitter (shifting back or forth in time the arrival of the bit by a small amount) can sort of subtly blur or distort the shape of the analog signal derived from the converted bits. Hence the claim that some digital cables sound "better" than others. You probably need a fairly good system to hear this difference.

    Now Dolby Digital and DTS bitstreams are NOT handled this way. Data is sent in packets, grouped together when the correct number of packets are received and then transformed into an analog signal. Much the way info comes over the internet. So there is no issue with whether the bits arrive at the exact right time, as long as they arrive nearly so, all intact and in the right sequence. When they do not, you get things like loss of digital lock, or dropouts and such, not continuous audio with degraded sound.
    The response:

    There is only ONE source of jitter in the digital (PCM) chain that is relevant: at the digital-to-analog conversion (i.e. the DAC in your receiver, preamp, etc). Jitter anywhere else has no relevance on the sound. As long as the DAC gets it right it makes no difference. If the DAC is dependent upon such timing the designer of the DAC should be taken out and publicly humiliated for such an incredibly stupid design.

    DD/DTS IS in fact handled the same way as PCM. The only relevant difference is DD/DTS is (usually) buffered more heavily due to the variable bit rate (i.e. there may be periods when the incoming data rate is slower than the output).

    Some actual examples of "high end" audio gear that mucked things up despite the fact there were proper implementations DECADES before digital audio was even a wet dream:
    1. Used the incoming bitstream as a reference clock.
    2. Poor mixed signal layout leading to noise coupling to the clock.
    3. And the best example (from Richard D. Pierce):

    quote:
    --------------------------------------------------------------------------------
    In the process of testing, I connected a Tascam DA-30 DAT recorder to a "highly regarded and favorably reviewed high-end DAC from a prestigious company. With certain cables, the residual noise floor on the output of the DAC was simply AWFUL. Why?

    The reason was simple: you had two utterly incompetent implementations. The DA-30's SP/DIF output had miserable current drive capability: load it up with capable capacitance, and the output went into slew-rate limiting and failed to meet the rise-time requirement. The DAC, for all of it's thousands of pretentious dollars, had the most MISERABLE design for clocking around. The result was that, given the right cable with enough capacitance, the resulting output had oodles of jitter in it.

    Now, here's the ironic thing: the DAC used was considered by many to be very "transparent" and was one of the few that, it was said, COULD reveal differences in digital cables.

    Well, that's NOT what was really happening. In reality, the idiot who designed the reclocking circuitry in the DAC simply got it wrong in a seriously stupid fashion: the designer simply made a DAC clock recovery circuit that was SO sensitive to small changes in input conditions, that its performance was all over the map.

    Yet, members of the high-end press praised this piece of crap for its "transparency. "It's obvious," it was said, "that anything that DOES show such large differences in cables MUST be transparent and high-resolution," when, in fact, precisely the opposite was the case.

    Far less price- and name-pretentious DACs, those with FAR better clocking circuits did NOT exhibit this ridiculous sensitivity to the cable: they were immune to the variation that the more expensive DAC simply could not handle properly.

    So, if the problem exists, it's not because some of us are using DACs that are immune to these errors, it's because some of you AREN'T :-(.

    Dick Pierce
    Professional Audio Development
    My followup:

    Understand that. But my understanding of PCM was that the signal clock is derived from the arrival timing of the incoming bitstream. Clock data is NOT included in the digital data itself. Is this not so?

    If it is so, would not anything that distorted the electrical signal in the digital cable distort the result of that clock timing calculation the DAC performs? And hence the result of the transform which yields the analog waveform?

    I know some good DACs buffer and reclock the signal for PCM, but these tend to be nicer ones($1K range and up), certainly not the ones in even otherwise very decent receivers.

    "DD/DTS IS in fact handled the same way as PCM. The only relevant difference is DD/DTS is (usually) buffered more heavily due to the variable bit rate (i.e. there may be periods when the incoming data rate is slower than the output). "

    This buffering and reclocking is what should make this type bitstream more immune to this problem. No?

    So are you saying I am incorrect about PCM jitter, or incorrect about DD/DTS immunity to jitter, or both? Or just incorrect that most DAC stages in receivers and CD/DVD players do not buffer and reclock the bitstream for PCM?
    Some more expert insight into my misunderstanding? The other poster is an analog signal engineer, so I am guessing he knows more about this than me.

    Thanks.

    BB
  • JonMarsh
    Mad Max Moderator
    • Aug 2000
    • 15259

    #2
    Both of you are partly correct, and both are overlooking or interpreting in odd ways some of the things that can happen.

    First, with SPDIF outputs, the clock must be "recovered" from the bitstream. Unlike some other digitial interfaces which send the clock separately (and hopefully, differentially), signal quality can be a serious issue for SPIDF and recovered clock jitter. There are ways to attempt to minimize this problem, including using multiple phase locked loops (combing a fast PLL (which locks on quickly and responds to short term timiing errors) with a slow one (for stability of the recovered clock output to the converter. If this sounds a bit like a kluge, it's only because, inherently, it is. The best way is to have a single central clock located very close to the DAC chips which is then also distributed to the transport; this is why really good "one piece" players can ultimately have better performance than any player using SPDIF or the balanced equivalent.

    Now, regarding is statement about the TASCAM DAT recorder, let's get one thing straight. If you have a true 75 ohm source, and if the cable is 75 ohm characteristic impedance, it doesn't matter AT ALL what the cable capacitance is! Longer cable runs merely delay the signal more before it gets to the DAC. This sounds like an example where the source gear was not actually 75 ohm.

    Beyond that, a good transformer isolated receiver should also be carefully implmented to not inject non-linearity into the sending source- for example, many reciever chips need to be buffered instead of driven directly on their RX/TX inputs, because they use schmitt trigger logic to improve noise immunity, but this presents a nonlinear load which distorts the received waveform, and hence the timing, in a signal correlated way. (yes, folks, that's called jitter).

    Now, even through out the DAC itself, you can run into many problems, and with standard logic families, the best solutions I know of is relentless and extensive local bypassing and regulation of every digital power pin. Otherwise, variations in Vdd, which correlate with signals as well as local clocks, cause variations in gate logic threshold, which, guess what, causes jitter.

    You'd be amazed at how much better you can make a cheap Philips CD transport mechanism sound just by going through it and replacing every electrolytic cap with an OSCON or similar low ESR/ESL cap, and bypassing every chip right at the component pins with a low ESR Y5V ceramic cap. Also, place some thin copper foil sheilds over all the IC bodies; these reduces propagation into near by circuitry.

    There are many other factors, but it comes down to good digital design in a more "analog" sense, paying attention to the details, and even terminating some of the high speed digital busses on the PCB properly- somethign which is rarely done.

    Regards,

    Jon




    Earth First!
    _______________________________
    We'll screw up the other planets later....
    the AudioWorx
    Natalie P
    M8ta
    Modula Neo DCC
    Modula MT XE
    Modula Xtreme
    Isiris
    Wavecor Ardent

    SMJ
    Minerva Monitor
    Calliope
    Ardent D

    In Development...
    Isiris Mk II updates- in final test stage!
    Obi-Wan
    Saint-Saƫns Symphonique/AKA SMJ-40
    Modula PWB
    Calliope CC Supreme
    Natalie P Ultra
    Natalie P Supreme
    Janus BP1 Sub


    Resistance is not futile, it is Volts divided by Amperes...
    Just ask Mr. Ohm....

    Comment

    • Brandon B
      Super Senior Member
      • Jun 2001
      • 2193

      #3
      Thank you sensei.

      BB

      Comment

      Working...
      😀
      😂
      🥰
      😘
      🤢
      😎
      😞
      😡
      👍
      👎
      Searching...Please wait.
      An unexpected error was returned: 'Your submission could not be processed because you have logged in since the previous page was loaded.

      Please push the back button and reload the previous window.'
      An unexpected error was returned: 'Your submission could not be processed because the token has expired.

      Please push the back button and reload the previous window.'
      An internal error has occurred and the module cannot be displayed.
      There are no results that meet this criteria.
      Search Result for "|||"