1. Dear Guest, if you haven't already... enter to WIN Samplitude Pro X4!
    Dismiss Notice

clock jitter

Discussion in 'Converters / Interfaces' started by audiokid, Apr 18, 2010.

  1. Chris Perra

    Chris Perra Active Member

    Joined:
    Jul 13, 2014
    Location:
    Toronto
    Home Page:
    I suppose some people would have the ability to hear the difference. I would suspect most wouldn't. For me I can hear my fingers rubbing together sitting in my studio while listening to those examples out of phase. I can hear my fingers but not the jitter without cranking the snot out of the gain. For me,.. that's enough not to worry about jitter.

    It's definitely there though.... I just can't hear it with those examples.
     
  2. Chris Perra

    Chris Perra Active Member

    Joined:
    Jul 13, 2014
    Location:
    Toronto
    Home Page:
    Here's B vs A 40 tracks of each cancelling each other out. This is audible. I wonder if Jitter stacks like this or not though It's pretty noticeable.

     

    Attached Files:

  3. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    What I notice most problematic in music,
    • why I choose to replace specific tracks in order to improve the imaging of busy mix,
    • what has to do with consistent acoustic reflections throughout all the tracks summed
    • what we tend to sacrifice for volume on a track per track.Example moving volumes from track to track and forgetting about the silent sections we think is dead and useless information....
    • what happens when we overdub and break tracks into pieces, I'm seeing the image of invisible sky scrapers of dead space left in a mix. So what we have are all sorts of levels of air, we get sloppy with.
    • digital editing....
    This is imho, why music today sounds boring, dead and worst case, all swirly.

    Inconsistent track to track acoustic space (room sound, reverberation, bleed from track to track, bouncing, overdubs) all effects the sum image. This is the "accumulating" that depicts a "good, better best, stellar" sounding recording. Which I assume why people invest in $6000 master clocks to try and help sync their problematic rats nest of external gear, why we invest in better, more stable converters, which I would go so far as including, why a 2 channel DA AD apposed to 8 or 16 or 32 one rack converters are more stable.

    I'm thinking PSU and the ability for multiple channels to remain stable starts looking why most of the top end converters don't go past 8 channels on a single PSU.

    Most of us only listen to the face of an instrument and pay less attention to the dead space around it. I think its safe to say, the dead space has a lot of detail that gets us in trouble. Kind of like have higher ceilings in one track , opened windows in another and mixing them all up at various levels, all out of sync.

    I'm thinking this is related in this thread somewhere?
     
    Brien Holcombe likes this.
  4. Chris Perra

    Chris Perra Active Member

    Joined:
    Jul 13, 2014
    Location:
    Toronto
    Home Page:
    This 40 track example shows what happens if you were to use the same material and run it round trip 40 times. I'm not sure who does that wanting to keep the track pristine at the end. In a practical situation you would be round tripping to outboard gear with the intent to change the sound. So unless you hear horrible things from the hardware you are using, I'm not sure it would be a concern.

    I also wonder if say you recorded a whole band all at once whether or not the jitter would stack the same way or be the same.

    I suppose if you had 24 tracks going. 3 sets of 8 channels per Interface then you would have only 3 potential jitter differences.
    If you recorded say drums over 2 units A and B and used A for the rest of the overdubs you'd only have 1 jitter difference.

    I'm not sure if jitter in that situation would be an audible concern.
     
  5. MrEase

    MrEase Active Member

    Joined:
    Jan 19, 2006
    Location:
    West Suss ex, UK
    It is important to recognise that clock jitter causes slight errors during the sampling process on each individual track. Once digitised these errors are fixed within that track. Provided each individual track does not have excessive errors that are audible then there should be no cumulative effect whenever mixing in the box as each track is independent whatever processing is used. Compare this to a track where you add some overdrive to a guitar. You would have no expectation of this overdrive affecting other tracks within the mix or causing a cumulative effect.

    Clock jitter only becomes a concern whenever you convert A-D or D-A as this is the only stage where jitter can have an effect. For instance, if you mix outside the box then you are passing the mix through D-A converters and then back into (possibly another DAW's) A-D. This process will add two stages of clock jitter error to the equation. Balance this with the fact that several members here are convinced that mixing OTB and using an independent DAW to record the mix yields improved results and we must conclude that either there is something the OTB mixer adds or could it possibly be that the clock jitter helps??? I'm not drawing any conclusions here but really I'm pointing out that this really shouldn't be an issue.

    Just a point about the tests given in the Cranesong links. Whenever you have to significantly boost difference signals, you are also losing resolution. In another forum some years ago I contributed to a thread about dither where I intentionally limited the resolution of a sine wave to just a few bits - which sounded awful when boosted. Using dither on the same process produced a much more acceptable sine wave at the same resolution. This is really to point out that the boosted difference signals are not necessarily representative of the "true" distortion from clock jitter and will easily add extra distortion due to limited resolution. Does anyone remember the older digital synths with 8 or 10 bit resolution? They generally sounded OK(ish) but I frequently found the "buzzy, distorted" tails on some decaying sounds really annoying! Again, that was just down to limited resolution.
     
  6. Chris Perra

    Chris Perra Active Member

    Joined:
    Jul 13, 2014
    Location:
    Toronto
    Home Page:
    So, unless something is totally noticable noise, buzz etc, jitter isn't a concern?
     
  7. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    For the quoted reasons you point out I will share some brief insight into my madness:
    I choose to use two uncoupled DAW's for a variety of reasons.
    These would include:
    • tracking @ higher samples rates than the destination, monitoring SR conversion in real time,
    • increasing computer performance (dedicated DAW's to do specific tasks),
    • avoid bouncing,
    • monitoring strategic order of specialized outboard gear,
    • monitoring sum and master
    • monitoring the harvesting and export of audio, especially when you are in the creative process of a mix/sound replacing/ sound designing/master and you are flipping between two DAW's .
    The list of why I do what I do is long and little to do with the simple change that may happen (add value) using a D A D pass, but, it is one of the reasons as well!


    Personally I find ITB sounds bigger and fuller to what any analog console does for me. Once ITB, stay ITB. The less converter, the better.
    But, One box kind of sucks in comparison to 2. My madness is all about the workflow and monitoring the workflow.
    DAW1 is for tracking and mixing, DAW2 is for capturing the sum. DAW 1 is designed only for tracking and mixing, DAW 2 is designed for summing, mastering and export.
    To my knowledge, this is not possible on one computer, at least not to the level I require.

    I do hear an improved sum when I track at example: 96/24 and capturing the mixdown sum to 44.1/24 apposed to bouncing on one box like the mass do.
     
  8. MrEase

    MrEase Active Member

    Joined:
    Jan 19, 2006
    Location:
    West Suss ex, UK
    Hi Chris (audiokid),

    I wasn't making a comment as to why you might do something a particular way and certainly not saying I disapprove of your methods or whether you are mad or not! :)

    Hi Chris (Perra),

    That's sort of putting words in my mouth but in essence, true. What I was really trying to say is that clock jitter only comes into play whenever you move from analogue to digital or vice versa. If you're happy with the recorded sound of individual tracks then the mixing process will have no impact whether tracks had high jitter when tracked or not. There should be no accumulative errors due to jitter and the mixing process. Note that when mixing, you are constantly playing back and the playback clock jitter comes into play. If you really had a clock jitter problem then it should be noticeable then, even if the tracking was "perfect" (which it can never be, sadly).

    I guess it's worth noting that even if you're just listening to a single track, you have two lots of jitter involved, once when recorded and again whenever you play it back.

    To all, I only came back here as I was interested in the samples provided in the link given earlier and had some comments I thought people should be aware of. I hope it helps!
     
    Chris Perra likes this.
  9. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    no worries, but I am Mad for getting into the business. Somehow I'm able to function remotely well on the outside world. :)

    I have more questions for you, just need to think about the wording. Thanks as always for your contribution here.
     
  10. Chris Perra

    Chris Perra Active Member

    Joined:
    Jul 13, 2014
    Location:
    Toronto
    Home Page:
    For me.. listening to this examples nulled and hearing super exaggerated jitter I'd treat it like part of your sound like any other part of the chain. You either like or don't notice the sound or notice and dislike what you hear. If it sounds good it's good if it sound bad it's not.
     
  11. MrEase

    MrEase Active Member

    Joined:
    Jan 19, 2006
    Location:
    West Suss ex, UK
    Up to now I've treated this whole thread almost entirely from the (electronic) engineering standpoint. I agree with Mr. Perra though.

    The bottom line is that I've re-recorded so many things in the past due to a botched performances (frequent in my case!), trying different mic's & positions to get better sound and lots of other reasons. I have never ever re-recorded because I thought there was too much jitter! Other than swapping master's on my two sound-cards, there's nothing else I could do anyway. The key here is what I said earlier - I doubt anyone can clearly state what clock jitter actually sounds like as it can arise from such diverse sources.

    I doubt anyone hereabouts can ever say they re-recorded any track solely due to jitter problems. Drop-outs sure but not clock jitter per se.

    Can anyone here say for sure that they have ever encountered a problem with clock jitter?
     
  12. DonnyThompson

    DonnyThompson Distinguished Member

    Joined:
    Nov 25, 2012
    Location:
    Akron/Cleveland, OH
    Home Page:
    Unequivocally no. Until I heard the examples posted, I had no idea what clock jitter even sounded like. I knew from reading various articles over the years that clocking issues could be a concern, but I've never re-recorded or re-mixed anything because I heard clock jitter.

    That's not to say that I might not have heard something amiss as a result of clock jitter, I suppose that it's possible that this has happened - but as far as positively identifying it or knowing for a fact that those problems were caused by jitter, no... I've never once heard something "off" on a track or mix, and said "Yep... That's clock jitter!" And, taking into account that it never seems to sound the same, presents an even greater difficulty in identifying it. We heard, what, four examples of intentionally represented jitter? And, no two of them sounded alike; so how many other possible audible examples of jitter are there? 50? 100? 1000? Maybe even limitless? If it never sounds the same, if it never has the same sonic fingerprint, if it's always "random" and presents itself differently every time, then how can it be identified as such in a mix? I'm not telling ... My question isn't rhetorical, guys ... I'm sincerely asking here...

    I've certainly re-tracked and re-mixed due to certain issues I've heard - things like maybe too much "room" on a vocal track, or phase issues, or other noise(s) - times where I heard noise due to low output mics and a cheap pre gained-up in response, or hum on a guitar track, or sibilance, etc. Those issues are identifiable. You know phase issues when you hear them, you know the sound of a noisy pre, or the sound of too much "room" on a vocal or on an acoustic guitar track...

    But never was it because I was able to actually hear and definitely attribute the problem to jitter.

    IMHO of course. :)

    -d.
     
  13. MrEase

    MrEase Active Member

    Joined:
    Jan 19, 2006
    Location:
    West Suss ex, UK
    Sorry for the slow reply Donny. The simple answer to your non rhetorical question is that you're highly unlikely to ever have jitter induced noise that you could ever hear. If you look way back in the thread I gave calculations of the worst possible case scenarios for jitter induced artifacts which (I hope) show how unlikely this is to ever be a problem.

    If you consider the Cranesong tests, they only serve to confirm that. They have deliberately introduced very large and discrete signals to the jitter and then had to enormously ramp up the gain on the difference files to even give a demo of a jitter "sound". The real world is quite different and even a relatively high jitter clock would normally be made up of predominantly pink(ish) noise (once again with caveats regarding poor design!). This would not have given Cranesongs much to go on in trying to demo the "sound" of jitter as it would only impart a pink(ish) noise to the recorded signal at very low levels. Ramping up the difference gain on this would only yield pink noise that should be indistinguishable from "normal" noise.

    All told this means you have no way of having a generalised "how can it be identified". In all honesty I sincerely believe the "prophets of doom" in regard of clock jitter have got it wrong. All our A-D, D-A and DAW processes rely heavily on mathematical procedures. It follows that the same procedures can show why clock jitter should not be a problem in our audio systems. This is what I have attempted to show in this thread giving some real numbers to work with. It makes no sense (to me at least) to discount the simple maths I've presented here while inherently accepting that all the other maths our signals are processed with are OK.
     
  14. Boswell

    Boswell Moderator Distinguished Member

    Joined:
    Apr 19, 2006
    Location:
    UK
    Home Page:
    Jitter-induced artifacts and jitter-induced distortion are two separate things. As MrEase said, the artifacts are almost never encountered on a properly clocked system, as they would manifest themselves as double-sample, spikes or similar very audible (and visible) things.

    On the other hand, distortion due to sampling uncertainty is present every time you convert from the analogue domain to the digital domain or back again, as the sampling instant is never precise, at least at the picosecond level. A shift in sampling instant results in a change in converted amplitude of an audio waveform, and this shows up as a distortion in the conversion process. Any digital processing assumes the waveform is sampled at exactly the clock instant, and when it is not, the sampled value is wrong by a small amount. This is not visible at the DAW level.

    What you have to evaluate is whether the sampling uncertainty caused by the inevitable clock jitter in a system results in an acceptable level of audio distortion. This type of jitter may not cause any audible artifacts, but the waveform does deteriorate, and, as the jitter increases, you will start to hear it first on certain types of programme material, solo piano being a common example.

    It's not easy to generate demonstration files of this type of distortion. Maybe the same piano recorded using a low-end converter and a top-quality low-jitter converter would illustrate something, but it would depend heavily on the listener's reproduction equipment whether any difference between the two could be detected, and that would have to be after other more obvious differences in the quality of the conversion itself had been allowed for. Simply degrading the external clock going into a top-end converter would not produce the required effect, as modern high-end boxes use an incoming clock as a long-term frequency correction of their own internal conversion clock rather than clocking the conversions directly from it.
     

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice