The topic of jitter often comes up when discussing pc based audio,
but I have never personally experienced its negative effects firsthand.
It's possible that I might be unaware of its presence.
How do you all perceive and confirm the existence of jitter on hearing in music and in what situations?
but I have never personally experienced its negative effects firsthand.
It's possible that I might be unaware of its presence.
How do you all perceive and confirm the existence of jitter on hearing in music and in what situations?
Last edited by a moderator:
Someone will correct me if I've got this wrong, but I don't think you 'hear' jitter (or noise, particularly at my age) which make both hard to register. What you do hear is the absence of both, because the former smears and the latter masks detail and the feeling of presence that you only get with a clean signal. My digital system is by no means the last word in high end audio (I choose not to spend that much) but it will almost certainly be my last. It consists of a Denafrips Iris+AresII in one location and Iris+PontusII in another, both fed by an HP laptop. Depending on your current source, you may get some idea by trying Fidelizer software or auditioning an Iris or similar DDC.
Thanks davidjt
Well, yes, I started PC audio when I turned 60, so it has been almost 20 years since then.
Perhaps my aging ears couldn't perceive what I couldn't hear.
I can understand the mechanism behind the occurrence of jitter based on reasoning,
but it is difficult to perceive it audibly in reality.
To enjoy music through audio,it could be seen as a kind of happiness in a way.
Well, yes, I started PC audio when I turned 60, so it has been almost 20 years since then.
Perhaps my aging ears couldn't perceive what I couldn't hear.
I can understand the mechanism behind the occurrence of jitter based on reasoning,
but it is difficult to perceive it audibly in reality.
To enjoy music through audio,it could be seen as a kind of happiness in a way.
Last edited by a moderator:
That seems like a pretty good way of putting it. There may also be a long answer but probably not very helpful to go into it all here.
So far I'm too cloth-eared to notice the effect(s) of jitter but my wife noticed it when I fed one of my multibit DACs from a USB adaptive interface (CM108). She said it sounded like the individual players didn't sync up their contributions very well, like they were playing slightly out of time with one another. She doesn't have the usual audiophile vocabulary but if she did then my guess is that would translate as 'PRaT' degradation.
I reckon the audible effects of jitter will depend on what kind of DAC you're using. Sigma-delta DACs with oodles of noise-shaping will be likely to fold-down high frequency noise into the audio band when jittered and hence raise the noisefloor in signal-dependent ways.
Okay, there is 'jitter' (corresponds to 'far-out phase noise'), and there is 'close-in phase noise.' They tend to have different perceptual effects, at least if any effects at all happen to be noticed.
One resource: https://www.analog.com/media/en/technical-documentation/application-notes/AN-756.pdf
For further reading if interested, e.g. pages 3, 15 of attached.
One resource: https://www.analog.com/media/en/technical-documentation/application-notes/AN-756.pdf
For further reading if interested, e.g. pages 3, 15 of attached.
Last edited:
If you're happy I'd stop right there and stay off the treadmill.
I'm still increasing my enjoyment now and then, plus I enjoy building stuff - but that's a different hobby altogether.
I agree. I really notice the difference. By default I do not have W10 configured to minimize the jitter generated by the CPU (ALU <-> registers) so sometimes I play music without having it activated, immediately noticing the difference (with good recordings).
Thanks abraxalito
My DAC is a delta-sigma Type, and I can clearly hear the 'PRaT'.
So, even if there is, the jitter is not so significant enough to be noticeable.
Last edited by a moderator:
You can not hear jitter as 'noise' coming out of your loudspeakers.
The effect of jitter can only best be heard when it is absent. The effect of jitter is a slight nervousness to the sound. Without jitter the is no such thing as having more detail, what is does is it let's you relax more into the music ands in that state you are able to discern more little details and spaciousness. Measurably it stays the same however, so it's all very subjective and because it has the most influence on you being able to relax into the music it is very hard if not completely impossible to hear any difference in direct AB comparison. But don't go chasing for the absolute lowest jitter specs, there is a certain point where further reduction doesn't make any sense anymore.
The effect of jitter can only best be heard when it is absent. The effect of jitter is a slight nervousness to the sound. Without jitter the is no such thing as having more detail, what is does is it let's you relax more into the music ands in that state you are able to discern more little details and spaciousness. Measurably it stays the same however, so it's all very subjective and because it has the most influence on you being able to relax into the music it is very hard if not completely impossible to hear any difference in direct AB comparison. But don't go chasing for the absolute lowest jitter specs, there is a certain point where further reduction doesn't make any sense anymore.
This thread and video may be useful:
https://www.audiosciencereview.com/...audio-measurements-and-listening-tests.21115/
https://www.audiosciencereview.com/...audio-measurements-and-listening-tests.21115/
Deterministic jitter from SPDIF, versus random jitter/close-in_phase-noise from clocks should be considered separately. Jitter can sound like a lot of different things depending on the particular jitter and the particular dac. There is no simple answer to what it sounds like.
This is a very crude analysis of the theoretical effects of jitter in the timing of samples in a DAC:
Consider a 16 bit digital signal at 48kSPS, with a full volume 1kHz sinusoid, at zero crossings it will be changing at about 4300 counts/sample in order to manage a swing of +/-32767 counts over 48 samples.
A 1-bit error on that slope would thus be 1/4300 of the signal. Each sample period is about 21µs, and 21µs/4300 = 5ns, roughly. So a 5ns timing error will affect the effective slope of the waveform by the same amount as a 1 bit error (for a 1kHz sinusoid).
Thus 5ns jitter could be equated with 1 bit lost resolution at that frequency, ie reduction from 16 to 15 bits. Or 10ns jitter loses 2 bits, etc. (Very roughly)
On a 2kHz signal the slope is twice as large, so that 5ns jitter loses 2 bits or so, or rather it raises the noise floor (when a strong signal is present) - in fact the noise floor raises more at for higher frequencies when random jitter is present. As a corollory to this bass is unlikely to be affected unless the jitter is really dire - so jitter should be a non-issue for a subwoofer channel for instance, as slopes are low and most HF noise is filtered out anyway.
If the jitter isn't random, but correlated with the signal, there will be spurious harmonically-related tones produced as well as noise, as this is a kind of FM modulation that gives harmonic distortion that rises with signal level and signal frequency.
Consider a 16 bit digital signal at 48kSPS, with a full volume 1kHz sinusoid, at zero crossings it will be changing at about 4300 counts/sample in order to manage a swing of +/-32767 counts over 48 samples.
A 1-bit error on that slope would thus be 1/4300 of the signal. Each sample period is about 21µs, and 21µs/4300 = 5ns, roughly. So a 5ns timing error will affect the effective slope of the waveform by the same amount as a 1 bit error (for a 1kHz sinusoid).
Thus 5ns jitter could be equated with 1 bit lost resolution at that frequency, ie reduction from 16 to 15 bits. Or 10ns jitter loses 2 bits, etc. (Very roughly)
On a 2kHz signal the slope is twice as large, so that 5ns jitter loses 2 bits or so, or rather it raises the noise floor (when a strong signal is present) - in fact the noise floor raises more at for higher frequencies when random jitter is present. As a corollory to this bass is unlikely to be affected unless the jitter is really dire - so jitter should be a non-issue for a subwoofer channel for instance, as slopes are low and most HF noise is filtered out anyway.
If the jitter isn't random, but correlated with the signal, there will be spurious harmonically-related tones produced as well as noise, as this is a kind of FM modulation that gives harmonic distortion that rises with signal level and signal frequency.