How high is the jitter level on HDMI vs S/PDIF on a High-End Processor? - Page 6 - AVS Forum
Forum Jump: 
Reply
 
Thread Tools
post #151 of 164 Old 03-19-2010, 01:58 PM
AVS Special Member
 
krabapple's Avatar
 
Join Date: Apr 2003
Location: in a state bordered by Kentucky and Maine
Posts: 5,247
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 88 Post(s)
Liked: 154
Quote:
Originally Posted by DonH50 View Post

Nick,

Good point about the PC noise (and the rest, natch). When the paper said the listeners used their own DACs, I took my experience out of context and assumed they'd provide a clean clock and data path (e.g. double-buffering the data, or loading the samples and then executing from a clean copy independent of the PC). Re-reading, if anything the opposite appears true. To my mind that almost completely invalidates the test, as such a high noise floor would likely mask any of the things they were trying to hear.

Cheers - Don


Why do you assume the noise of the PC -- noise in the soundcard output, really -- was high enough to be a problem?
krabapple is offline  
Sponsored Links
Advertisement
 
post #152 of 164 Old 03-19-2010, 01:59 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 17,824
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 542 Post(s)
Liked: 340
Quote:
Originally Posted by krabapple View Post

I see your amirm and raise you a James Johnston.

Since JJ used to work in my group, not sure where that leaves you .

Amir
Founder, Madrona Digital
"Insist on Quality Engineering"

amirm is offline  
post #153 of 164 Old 03-19-2010, 02:07 PM
AVS Special Member
 
krabapple's Avatar
 
Join Date: Apr 2003
Location: in a state bordered by Kentucky and Maine
Posts: 5,247
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 88 Post(s)
Liked: 154
Quote:
Originally Posted by dknightd View Post

Sorry in advance for taking this off topic (but related to this post)

I've tried this test - well not quite, but a related one. Years ago when I was deciding how to put music on my computer I compared mp3 at various bitrates to lossless.
I picked 10 songs more or less at random, but ones I thought would be a good test, and that I could stand listening to over and over.
128 vs lossless was easy - I stopped after "passing" on the first three songs.


All mp3 codecs are not created equal -- nor are codecs from years ago necessarily comparable to ones available today. Even 128kbps vs lossless stopped being easy some time ago, if you use the best codecs.


Quote:


Occasionally I think I can hear problems at 320, so I rerip lossless and compare again. So far I have not been able to pass an abx at 95% confidence. Maybe that is because I subconsciously do not want to hear a difference - it would mean lots of work reripping. Maybe it is because I don't know what to listen for, or maybe I'm using the wrong tools.

Or maybe you're normal -- it's not typical to 'pass' an ABX of 320kbs vs lossless. MP3 codecs are really quite good.
krabapple is offline  
post #154 of 164 Old 03-19-2010, 02:08 PM
AVS Special Member
 
krabapple's Avatar
 
Join Date: Apr 2003
Location: in a state bordered by Kentucky and Maine
Posts: 5,247
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 88 Post(s)
Liked: 154
Quote:
Originally Posted by amirm View Post

Since JJ used to work in my group, not sure where that leaves you .

I know he did, amirm. It would be most interesting to get him here to talk about the audibility of jitter, and how one would verify it, as well as the audibility of high-bitrate lossy. He does post to AVSforum.

I'd also be interested to know about the selections you used in your high-bitrate AAC listening tests, what version of the codec you used, and how the ABX tests were performed and the statistics used.
krabapple is offline  
post #155 of 164 Old 03-19-2010, 02:38 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 17,824
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 542 Post(s)
Liked: 340
Quote:
Originally Posted by krabapple View Post

All of which seems irrelevant to this thread, really. How would I go about doing an ABX to test my sensitivity to correlated jitter, amirm?

"Small moves Elli... Small moves!" So says the (alien) father to Jodi Foster's character in the movie Contact which she tries to get the answer to everything in the universe .

I will say more when I have time....

Amir
Founder, Madrona Digital
"Insist on Quality Engineering"

amirm is offline  
post #156 of 164 Old 03-19-2010, 02:41 PM
AVS Special Member
 
krabapple's Avatar
 
Join Date: Apr 2003
Location: in a state bordered by Kentucky and Maine
Posts: 5,247
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 88 Post(s)
Liked: 154
Quote:
Originally Posted by amirm View Post

"Small moves Elli... Small moves!" So says the (alien) father to Jodi Foster's character in the movie Contact which she tries to get the answer to everything in the universe .

I think my request was rather smaller-scale than that.

Quote:


I will say more when I have time....

Looking forward to it.
krabapple is offline  
post #157 of 164 Old 03-19-2010, 04:12 PM
AVS Special Member
 
DonH50's Avatar
 
Join Date: Feb 2010
Location: Monument CO
Posts: 5,932
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 102 Post(s)
Liked: 238
Quote:
Originally Posted by krabapple View Post

Why do you assume the noise of the PC -- noise in the soundcard output, really -- was high enough to be a problem?

Actually, the PC's sound card was not used. The testers provided their own DACs, amp, and speaker or headphones. The authors provided a PC and interface to the DAC using whatever output the DAC box required. I have measured jitter from PC busses to range from perhaps 100's of ps to well into the ns region, but have not measured them recently (might be an interesting thing to try in the lab). I do not know the specs of the interfaces they used between PC and DAC, nor the PC they used (the names etc. not stated in the paper). So, properly speaking, I assumed there was insufficient information in the paper to properly validate the tests. Nor do I know that they are invalid (throwing out the baby with the bath water), but in any event they do not address the original question I had about correlated vs. coherent/deterministic jitter levels, either present at the DAC's clock input or their threshold of audibility. In another thread the point was made that, for such low levels of jitter (few ns), it is unlikely that anyone could hear it if it were correlated or not, a valid point.

Honestly, for me this falls deep into the "sorry I asked" category. I was curious and hoped to get an answer, not start (continue) a huge squabble or get embroiled in semantics. There are obviously people with ears far more golden than mine, and credentials far greater than I'll ever have in this area, on both sides of the fence. I plan to sit on the sidelines (obviously not my strong suit) and continue to follow up the reading material various folk have provided before I get thrown off this board. Or perhaps that would be best...

I am tempted to sign off with a tag a friend uses" "Have a nice day, unless you've made other plans". It's clearly time for me to bow out.

Besides, I've got a foot of snow to shovel so my wife can get back in the garage when she gets home from work (spring in the Rockies).

G'day - Don

"After silence, that which best expresses the inexpressible, is music" - Aldous Huxley
DonH50 is offline  
post #158 of 164 Old 03-19-2010, 11:36 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 17,824
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 542 Post(s)
Liked: 340
Quote:
Originally Posted by krabapple View Post

I think my request was rather smaller-scale than that.

Well, I have learned that nothing about jitter is "small scale" as evidenced by your post claiming that what I said had nothing to do with the topic. Everything turns into a cat fight sooner or or later, sad to say. But I will go ahead and provide an answer, hoping you are true to your word that a simple response is all that was required.

I think we can all agree that jitter distortion is small and most people can't hear it. I hope we can also agree that one can arbitrary increase jitter until it becomes audible to everyone. In that sense, it completely mirrors compression artifacts. Everyone can hear compression artifacts at 64 kbps, but scant few can do so at much higher data rates such as 384kbps.

It also then follows that if you are not able to hear small compression artifacts which expert listeners can detect (i.e. proving that artifacts do exist), then you are not going to hear jitter either at similar levels that others could. In that sense, the opinion expressed that jitter must be inaudible because that person can't hear it, can be tested with something much easier, namely compression artifacts.

Second point was that hearing digital artifacts requires learning for many people. People don't know what small artifacts sound like and hence, are shooting in the dark as they go about testing themselves to hear it. By training yourself on compression artifacts, you learn to use your ear as an instrument rather than something that provides listening pleasure. Once there, I assure you that you can hear other small impairments such as jitter.

As I implied with my quote, think of my suggest as a "small move" toward becoming an expert listener. Keep working at it and I assure you that soon, you will have hearing acuity well above where you started and better than 99.9% of the general public.

Amir
Founder, Madrona Digital
"Insist on Quality Engineering"

amirm is offline  
post #159 of 164 Old 03-20-2010, 12:12 AM
AVS Special Member
 
krabapple's Avatar
 
Join Date: Apr 2003
Location: in a state bordered by Kentucky and Maine
Posts: 5,247
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 88 Post(s)
Liked: 154
Quote:
Originally Posted by amirm View Post

Well, I have learned that nothing about jitter is "small scale" as evidenced by your post claiming that what I said had nothing to do with the topic. Everything turns into a cat fight sooner or or later, sad to say. But I will go ahead and provide an answer, hoping you are true to your word that a simple response is all that was required.

I think we can all agree that jitter distortion is small and most people can't hear it. I hope we can also agree that one can arbitrary increase jitter until it becomes audible to everyone. In that sense, it completely mirrors compression artifacts. Everyone can hear compression artifacts at 64 kbps, but scant few can do so at much higher data rates such as 384kbps.

It also then follows that if you are not able to hear small compression artifacts which expert listeners can detect (i.e. proving that artifacts do exist), then you are not going to hear jitter either at similar levels that others could. In that sense, the opinion expressed that jitter must be inaudible because that person can't hear it, can be tested with something much easier, namely compression artifacts.

That doesn't follow. Just because two different kinds of subtle artifacts become readily audible when you increase their level, doesn't mean they become audible at the same level. That someone is sensitive to compression artifacts doesn't determine they will be equally sensitive to jitter.


Quote:
Second point was that hearing digital artifacts requires learning for many people. People don't know what small artifacts sound like and hence, are shooting in the dark as they go about testing themselves to hear it. By training yourself on compression artifacts, you learn to use your ear as an instrument rather than something that provides listening pleasure. Once there, I assure you that you can hear other small impairments such as jitter.

Again, small impairments are not necessarily alike in either magnitude or audibility by virtue of being 'small'.
krabapple is offline  
post #160 of 164 Old 03-20-2010, 09:04 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 17,824
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 542 Post(s)
Liked: 340
Quote:
Originally Posted by krabapple View Post

That doesn't follow. Just because two different kinds of subtle artifacts become readily audible when you increase their level, doesn't mean they become audible at the same level.

Who said they became audible at the same level? That has nothing to do with the points I made. Are you able to answer the question put earlier on how one becomes a trained listener to hear compression artifacts better?
Quote:


That someone is sensitive to compression artifacts doesn't determine they will be equally sensitive to jitter.

But the fact that you learn the skills necessary and scientific approach used to get there certainly does. To wit, when we test new methods in audio compression, expert listeners always rate much higher than general population in detecting its artifacts. By your logic, since the artifacts are different in new methods, they should not be able to detect any better than general public yet they clearly show better ability to detect them.
Quote:


Again, small impairments are not necessarily alike in either magnitude or audibility by virtue of being 'small'.

The commonality is not the distortion itself but your ears being able to tune to digital artifacts in general. That starts by knowing what artifact can be there, what selection of content reveals it, and then learning to search for it.

Amir
Founder, Madrona Digital
"Insist on Quality Engineering"

amirm is offline  
post #161 of 164 Old 03-20-2010, 09:12 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 17,824
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 542 Post(s)
Liked: 340
Quote:
Originally Posted by dknightd View Post

I would LOVE to be able to abx 320mp3 from lossless. It really bugs me that I can throw away that much signal and NOT be able to tell. Maybe I'm using the wrong source material - maybe I should try doing it on speakers instead of headphones - maybe I just have not learned how to tell the difference.

Just wanted to let you know that I have not forgotten about your question . I will address it later. For now, you are absolutely doing the right thing by using headphones. It makes the testing simpler.

Second point, MP3 at 128kbps rolls off the high frequencies (at least most encoders do). That makes it easy to tell the difference from the original that way, meaning that you may not be looking for the right thing as at higher data rates, the roll off is not there.
Quote:


Do you have any tips on how I might be able to reliably tell a 320 mp3 from a lossless file? Can you actually do this?

Yes, but the logic is rather inverted . I will explain later.

Amir
Founder, Madrona Digital
"Insist on Quality Engineering"

amirm is offline  
post #162 of 164 Old 03-22-2010, 10:01 AM
Advanced Member
 
jj_0001's Avatar
 
Join Date: Sep 2006
Location: Kirkland, Wa
Posts: 653
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Ok, I just came into this at the very end. I have no idea what the argument is about, but the first thing one must consider is the frequency spectrum of the jitter.

After all, sinusoidal jitter at .5555...Hz is called "decentered LP". Sinusoidal jitter at a high frequency will alias all sorts of things into your face.

Content with mostly high frequency content will be much more sensitive to jitter in general than low-passy signals. The mathematics is quite simple, error goes as the square of slew rate of the desired signal.

But I have no idea what you guys are arguing about.

James D. (jj) Johnston
jj_0001 is offline  
post #163 of 164 Old 03-22-2010, 12:59 PM
AVS Special Member
 
AndreYew's Avatar
 
Join Date: Jan 2000
Location: Southern California
Posts: 1,670
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by jj_0001 View Post

Ok, I just came into this at the very end. I have no idea what the argument is about, but the first thing one must consider is the frequency spectrum of the jitter.

It is said that one needs to repeat something 6 times before people remember. Consider the importance of the jitter spectrum repeated twice now:

http://www.avsforum.com/avs-vb/showt...2#post11641082



Quote:


But I have no idea what you guys are arguing about.

I think it's like professional wrestling. The conflict is the draw, not the resolution or closure.

--Andre
AndreYew is offline  
post #164 of 164 Old 03-22-2010, 01:43 PM
Advanced Member
 
jj_0001's Avatar
 
Join Date: Sep 2006
Location: Kirkland, Wa
Posts: 653
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by AndreYew View Post

It is said that one needs to repeat something 6 times before people remember. Consider the importance of the jitter spectrum repeated twice now:

http://www.avsforum.com/avs-vb/showt...2#post11641082




Has anyone noticed yet?
Quote:

I think it's like professional wrestling. The conflict is the draw, not the resolution or closure.

--Andre


"Soon the Gypsy Queen
In a blaze of vasiline
Will perform Guillotine
What a Scene
What a Scene"

I think that about says it.

James D. (jj) Johnston
jj_0001 is offline  
Reply Ultra Hi-End HT Gear ($20,000+)

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off