|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
206.190.70.135
It seems to me that with higher sampling rates the bit duration is shorter and so the jitter would be inherently less. But I just calculated that bit durations are us whereas jitter of ns can be heard apparently, so I suspect that transports that can do 192 kHz do not necessarily have less jitter?
Follow Ups:
Doubling the clock frequency will, all else being equal, double the uncertainty in the sample value due to timing errors (this might be much less than any uncertainty due to amplitude errors). But, because you have doubled the sample rate the timing error per sample will be averaged over two of the smaller sample periods so, in terms of the original clock frequency, you are back where you started.
Regards
13DoW
Two people who have probably forgotten more than most people know about digital audio reproduction disagree on something as fundamental as jitter.
Nobody really knows unequivocally what is exactly going on in regard to jitter..... Maybe the biggest question mark of all being the jitter characteristic of the original A/D transcription.... It is NOT zero, yet with this never-ending quest to completely eliminate jitter, it's as if we presume the original jitter signature is zero......
My opinion of jitter performance differs from the norm.... I believe the best sonic performance is attained by preserving the original A/D jitter signature, not by eliminating jitter.
.
Given there is jitter on the clock in the A to D conversion; if you want to preserve that, and I don't see any possible reason why you would if you could get rid of it, then you would want to have zero jitter on the D to A end of it.
I've never heard anybody say there was no jitter at the A/D end, but why in the world would you want to add more jitter on the D/A end, which is exactly what you are saying you do.
Turning off jitter suppression = adding more jitter than you have to.
.
.
You mentioned this before, which implies that that information is contained in the recording somehow? Or how is it done? Re-clocking? I doubt that. Or just a good idea that cannot be done?
♬♬ ♬
When playing CDs, I use players or DACs that do NOT re-clock the digital signal. And when I rip or burn CDs, I leave the jitter suppression option turned off. Provided the DAC is good to begin with, I generally get the most satisfaction in the playback with as little internal alteration of the signal (aside from the digital filtering) as possible.
If the CD manufacturers would record another clock track or tracks on the CD itself, then it could be done. The jitter from the original recording clock could be removed for the most part.
.
"you're making a joke right?"
Perhaps not. In some cases jitter can be undone. Wow and flutter are nothing but the analog form of jitter and there are techniques that can be used to remove these because analog tape recordings do have a "clock track" in the form of residual bias noise.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Hi Tony,
I have heard the results of this system on some Greatful Dead CDs. It has a live feel like nothing I have heard before. I wish they would put out some jazz and classicial with this system.
Dave
.
yea, I can see if you had a perfect replica of the original clock you could compensate for it, but tell me how you could add that data to a CD.
AND... if it is possible to correct for A to D jitter why not do it on the other end? Why not correct the data going on the disc instead of when you take it off?
.
.
It is a very complex topic with many ways to define it and many ways to measure it. Since there is no jitter standard in audio and no way to directly correlate what you measure to what you hear it is inevitable that there will be disagreements about it.
.
Sbrians,
Actually with every division by 2 you gain 3dB better jitter performance. While this is a maximum it is pretty easy to obtain. So 44.1 would be 3dB better than 88.2 and so forth. Mixing and upsampling do not apply here!
But just thinking like this is crazy stupid. Look jitter is what it is and is NOT all what you are hearing. There are a ton of variables to think about when you design digital audio equipment. This is only one of them.
Thanks
Gordon
J. Gordon Rankin
If you are adding dividing stages, how does the jitter go down?
the jitter added by the divider is much smaller than the 6dB reduction from doubling the clock period. In this case, the amount of jitter doesn't really get smaller since you are still modulating the clock edges by the same amount, but the number of clock edges is cut in half, so the jitter energy vs time period is halved. This is often how phase noise is measured.
This has little to do with audio quality IME. Its how much the edges move relative to the ideal, regardless of the sample-rate.
If differences in SQ between 44.1 and 88.2 are small (they are if your DAC does decent digital filtering), then I think the edges jittering are a bigger factor than the improvement due to SR.
Steve N.
RE: Jitter is often expressed as a percentage of clock period ..."
Steve, I believe we are in agreement. Let me put it a different way as it may make it clearer to some others.
Unless there's noise shaping, jitter depends on the high frequency content of the music, not the sampling rate. So long as the sampling rate is high enough to capture essentially all the audio content, going to a higher sampling rate isn't going to affect jitter. This is most obvious when looking at a DC (constant) signal, where there will be no effect of jitter whatsoever in the absence of noise shaping. If there is noise shaping, then jitter will appear as a form of noise modulation. Expressing jitter as a percentage of clock rate may be of interest to a communications engineer as it relates to the eye pattern and may affect bit error rate. However, for music reproduction this measure is irrelevant. What matters is the effect of jitter on the analog output of the DAC, which appears as phase modulation of the audio signal and is largely independent of the sampling rate.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Tony, Steve;
Remember audio related jitter error happens at really low frequecies like below 10Hz or 1Hz. While yes this error will effect higher frequency content because they are made up of less samples. The over all jitter will tend to go up as the sample rate goes up.
As slider said this is the way most phase noise measurements are made.
It's actually pretty easy to see by looking at the phase noise of the bit clock coming out of the divider.
Now the big questions is does any of this matter? Really if you look at current DAC chips internally none of this matters. As they simply have what is called a Serial Input Module which takes the data input and outputs parallel samples and also give some indication of the current sample rate then the rest of the dac works off of parallel samples. Really with parts like this the Master Clock becomes the key to really true low jitter output.
Thanks
Gordon
J. Gordon Rankin
As far as I know, no DAC manufacturer has yet to deliver a DAC that makes it unnecessary to tweak the PC or USB cable. So presumably, something really matters. Whether it is jitter or something else, who can say?
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
The clock jitter we most need to reduce is the data-correlated type which is so prevalent in many of the transmission interfaces, and so it doesn't really matter how you measure it in relation to the clock period since it isn't related to clock timing. That said, clock averaging can be a very effective method of eliminating that incoming jitter, and it is based on dividing the high rate clocks down to very low frequencies, and then using the diff signal to regenerate "clean" sample clocks from a low phase noise VCXO.
It is possible that jitter can be slightly lower depending on the design because the lower frequency sample-rate clocks require more dividing down. Less flip-flops used for the higher sample rates can mean slightly lower jitter.
These differences are usually very small, so they may not be audible.
Steve N.
Jitter is still an issue - ans "solved" or mitigated the exact same way regardless of sample rate.
"Knowledge is knowing that a tomato is a fruit. Wisdom is knowing not to put it in a fruit salad"
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: