Head-Fi.org › Forums › Equipment Forums › Computer Audio › Jitter measurement
New Posts  All Forums:Forum Nav:

Jitter measurement

post #1 of 20
9/21/09 at 10:35am
Thread Starter 
I think it would be interesting to compare the outgoing jitter of our digital sources, in order to understand which device does better in this field and if clock mods can bring measurable benefits.

This is the procedure:
  1. Set 44.1/16 as default audio format in the option of both the playback and recording device
  2. Put a Loopback cable between the input and the output of your soundcard/DAC and do a SNR test using RMAA @ 44.1/16 and check that the noise floor is below -120dB in the range > 6 khz.
    If this is not the case, you have a AC power too dirty and the jitter measurement won't be reliable.
  3. Download SpectraPLUS - Audio Spectrum Analyzer, install it (the 30dd demo version) and set correctly the input-output devices.
  4. Play this wav with your favorite player, with the eq disabled (flat), and check what's going on with SpectraPlus (click on RUN button)
  5. Check the peaks in the range 6 khz - 20 khz, focusing the windows in that range. The highest peaks will determine the jitter amount, according to this conversion table:
    -100 dB = ~500 ps
    -110 dB = ~180 ps
    -120 dB = ~60 ps
    -130 dB = ~20 ps
On the LessLoss site you find further details.
Share your results!
Reply
post #2 of 20
9/21/09 at 10:38am
Thread Starter 
Corsair hx520
Gigabyte GA-EG45M-DS2H
Intel c2d E4400
Teamgroup 2Gb ram DDR2
Asus Xonar Essence STX

result: 108 dB = 220ps

Reply
post #3 of 20
9/21/09 at 11:41am
I have doubts about this method.
Reply
post #4 of 20
9/21/09 at 1:12pm
looks good, thanks!

but the ADC on my HD2 is rated for 102dB..to my ears its analog out jitter is better than the Asus Essence, the stereo image is just so much less of a mess....but I won't be able to measure it I guess, except through another card's line-in..
Reply
post #5 of 20
3/29/10 at 4:13am
Asus Xonar Essence ST with XO changed (24576 kHz "FXO-BCF" "FULL"; T/HC 5В -40~85C 50ppm 10T/HC30pF 45/55%)

riderforever, isn't it worth to measure with 24 bit sample? If yes would you make this sample please?
Reply
post #6 of 20
3/29/10 at 10:20am
Quote:
Originally Posted by riderforever View Post
I think it would be interesting to compare the outgoing jitter of our digital sources
Did you mean to say "jitter of our analog sources"? I don't see how measuring the analog output/input has anything to do with the jitter on the digital output.
Reply
post #7 of 20
3/29/10 at 2:06pm
Quote:
Originally Posted by mojave View Post
Did you mean to say "jitter of our analog sources"? I don't see how measuring the analog output/input has anything to do with the jitter on the digital output.
If something on the analog output did not change due to jitter, then jitter would not be important. Jitter causes frequency modulation in the analog waveform.

Steve N.
Empirical Audio
Reply
post #8 of 20
3/30/10 at 7:38pm
Hmm, there is no reason you can't do the whole measurement with a high quality ADC (also connected to the computer) instead of the soundcard's ADC, right?
Reply
post #9 of 20
6/19/10 at 11:14am

One problem with this measurement technique is that it includes the DAC jitter contribution.  The PCI card that is used also has a questionable impact.  Most PCI cards dont get anywhere close to 120 dB noise floor.

 

I am now using a TDS7704B 7GHz Tek scope with jitter analysis software to characterize digital outputs and clocks.  Unfortunately, it is out of reach of most hobbyists.  List price is over $67K.  This is what it takes to do it right IMO.  I also use an hp 2.6GHz Spectrum analyzer.  I typically spec the full P-P jitter in addition to jitter below 1MHz.  20kHz is too low for a cut-off IMO.  My Pace-Car USB S/PDIF output for instance has P-P jitter below 1MHz of <27psec.

 

Steve N.

Empirical Audio

Reply
post #10 of 20
6/19/10 at 11:23am

The Empirical Audio Pace-Car Async USB convertor is brilliant.  I would love to be able to afford one!


Edited by SP Wild - 6/19/10 at 11:24am
Reply
post #11 of 20
6/19/10 at 11:51am

I will have a more affordable version of the Off-Ramp 3 with this technology by end of summer.  In the $700 range.

 

Steve N.

Empirical Audio

Reply
post #12 of 20
2/25/12 at 7:15am
Quote:
Originally Posted by shimm View Post

Asus Xonar Essence ST with XO changed (24576 kHz "FXO-BCF" "FULL"; T/HC 5В -40~85C 50ppm 10T/HC30pF 45/55%)

riderforever, isn't it worth to measure with 24 bit sample? If yes would you make this sample please?


Although it is an old thread, it's subject is now more interesting than ever, because the importance of the eliminated or minimized jitter is widely recognized.

 

So here is the measurement for ST 24 192: (Standard xo).

 

jriver192jitter.jpg

 

http://www.soundtrackforum.net/viewtopic.php?f=4&t=325

 

 

Just 10 - 20 ps!

 

Now a question for you mr. shimm, hoping that you will receive this message:

 

Is there a measurement for your ST's jitter before the changement of xo in 16 bit?

 

Because there is said that ST has only one clock for the 24 bit family and does not clean up  jitter for the 16 bit.

 

So how much worse is in 16 bit without the  modification?

 

Because as we can see for 24 bit, it is much better than 31%, (stx's=180 (16 bit) ps-31%),  which is given by ASUS!

 

Thank you very much.

 

 


Edited by hellenic vanagon - 2/25/12 at 7:39am
Reply
post #13 of 20
2/25/12 at 8:59am

I am not sure why 16 bit would be much different, as the card probably just plays it as 24 bit (filling the lowest bits with 0's). But measuring jitter with the card's own ADC may not be reliable, since it is driven by the same clock, so you can get a result that is better than it really is.

 

Reply
post #14 of 20
2/25/12 at 9:26am
Quote:
Originally Posted by stv014 View Post

I am not sure why 16 bit would be much different, as the card probably just plays it as 24 bit (filling the lowest bits with 0's). But measuring jitter with the card's own ADC may not be reliable, since it is driven by the same clock, so you can get a result that is better than it really is.

 



1)But you see that even with the modified clock, jitter rises to 50 ps in 16 bit mode. Or there is a case to be more than that with the standard,  since malfunctioning is occured  for any reason?

 

2)Another opinion says that in such case jitter is worst, because ADC's jitter is added to DAC's jitter.

 

Please, can you elaborate a little bit on this?

 

 

smile.gif

 

 

 

 

 

 

 

Reply
post #15 of 20
2/25/12 at 10:58am
Quote:
Originally Posted by hellenic vanagon View Post

1)But you see that even with the modified clock, jitter rises to 50 ps in 16 bit mode. Or there is a case to be more than that with the standard,  since malfunctioning is occured  for any reason?

 

Is the 24 bit jitter still better if you use the 16 bit test signal converted to 24 bits ?

 

Quote:
Originally Posted by hellenic vanagon View Post
2)Another opinion says that in such case jitter is worst, because ADC's jitter is added to DAC's jitter.

 

I meant there is a possibility in theory that jitter can be subtracted by the ADC, since it is not uncorrelated. Although it may or may not be an issue in practice.

 

Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Computer Audio
Head-Fi.org › Forums › Equipment Forums › Computer Audio › Jitter measurement