SY
I realize you are going for troll king.
But the discussion started from if there is change in jitter from generational copies. Jan in his interpretation measured where jitter in the data stream caused perceived changes. As I recall he found that it needed around 200ns of jitter for him to pick up an issue. This got narrowed down to where jitter can introduce errors and where data jitter is removed.
All of that has been nicely consistent with theory. It is still open if there is some yet unrecognized mechanism that might increase generational jitter.
Now variation in discrete time signal reconstruction can also be addressed. If the listening space is a room with a noise floor of 25 dba and the music peaks at 95 dba what amount of clock jitter do you think would matter and why?
If they're not perceivable, by definition it doesn't matter. First principles means the ONLY part that can POSSIBLY matter is the clock at the DAC. And that has to be seriously screwed up to be perceivable, at a level that could only be achieved by a garage design. Published tests show it to be on the order of 5-10 nanoseconds. You have to work pretty damn hard to cause that much jitter in a DAC clock.
This is a total non-problem (assuming the design wasn't executed by a retired dermatologist or former hi-fi salesman).
So what is the clock error in the dac that is causing jitter? Is jitter a discrete shift in frequency to one side or the other of the clock frequency or is it now a range that becomes wider as the quality of the clock goes down?
I would think that the frequency of the crystal oscillation would be high enough that it wouldn't ever cause timing errors, is this not the case?
I would think that the frequency of the crystal oscillation would be high enough that it wouldn't ever cause timing errors, is this not the case?
No, it's not the case, and the spectrum can vary, depending on the specific cause. Jitter can result from poor layout between the oscillator and the DAC, or poor crystal quality. It's easy to get right, i.e., below a nanosecond or so, and the vast majority of DACs and CD players do just that.
Wow! That was a great dinner. Great Hungarian red. Followed by a swing through the main battle area of town with the convertible open and the stereo blasting. She liked it (and thus so did I).
Back to jitter. My DCX2496 started to throw out garbage at around 50nS pk-pk jitter; but as bcarso (iirc) noted that is probably the limit of the S/PDIF receiver rather than the DAC.
Jan
Kindhornman,
If you think of it more as a slow phase drift in the clock, you can see how the DAC is, in effect, slowing down and speeding up as it drifts around the target frequency. (It's %age driven more than absolute clock frequency, given you're affecting all the intermediate steps)
http://www.ti.com/general/docs/lit/getliterature.tsp?baseLiteratureNumber=slaa566&fileType=pdf
If you think of it more as a slow phase drift in the clock, you can see how the DAC is, in effect, slowing down and speeding up as it drifts around the target frequency. (It's %age driven more than absolute clock frequency, given you're affecting all the intermediate steps)
http://www.ti.com/general/docs/lit/getliterature.tsp?baseLiteratureNumber=slaa566&fileType=pdf
Interestingly, the word that most accurately describes jitter is 'jitter'. Really. Check your dictionary.
Envision a current steering DAC, where each byte value activates a specific level of output current. Say your audio sample value is '6'; then the activated current source combination output current is '6'. Next value is, say, 32, and thus a current '32' is output.
An I/V converter is basically an integrator and all current values charge a cap to get the audio signal voltage, and if all sample values are correct and of the same duration, the conversion is perfect.
Now assume that the point where the '6' sample is finished and the '32' sample starts, in time, is slightly shifted, just a few nano-seconds, because of some noise in the system. Thus when the '32' sample stays active just a bit too long, the audio signal will be just a bit too high. This error is caused by that time shift and is called, you guessed it, jitter.
Jan
Last edited:
Generational copies, when copied in the digital domain, do not cause any increase in jitter. Cannot cause any increase in jitter.
If you make generational copies by going through DA and AD every time you do a copy, all bets are off of course - the result is probably not nice.
Jan
Adventurous one, hey ?
Your views on sample rate (upsampler) and jitter level (reduction) ?
Ha!
I love cats and when I saw that a couple years ago, under "hipster cat" it put me on the floor. Haven't had a chance to use it until today.
se
With some cross-channel feeds. But I thought I had recalled that he intentionally added noise. Other than that arising from the extra circuitry, I don't see it now in the online free service manual.
I may have trusted too much what I was told by the guy who gave me his sample because he thought his newer CD player didn't need it anymore. But if it did add noise, Carver should have called it the Carver Stochastic Resonator.
Oh well, missed opportunities.
Last edited:
Here's something I came across last week I should probably just use for my avatar here.
Hmm. Don't know why it's getting squished vertically when it displays here.
se
An externally hosted image should be here but it was not working when we last tested it.
Hmm. Don't know why it's getting squished vertically when it displays here.
se
Last edited:
That reminds me of what a music professor friend used to say, if someone expressed concern that she or he had hurt his feelings. "My feelings?? Oh, I had those taken out."Here's something I came across last week I should probably just use for my avatar here.
se
For those who dont know REG.... Robert E. Greene --- he has a PHD in mathematics. And is professor of mathematics at UCLA. Best known as the inventor of Wavelet theory.
[He recently retired.] He is a very astute listener and I would not assume he isnt familiar with DBLT and the math behind all the 1's and 0's theory of digital systems.
We've known each other a very long time. I will be talking to him about the subject. .... especially,jitter.
THx-RNMarsh
Last edited:
- Status
- Not open for further replies.