Head-Fi.org › Forums › Equipment Forums › Dedicated Source Components › An interesting paper on jitter audibility
New Posts  All Forums:Forum Nav:

An interesting paper on jitter audibility - Page 6

post #76 of 112
12/21/06 at 7:58pm
Quote:
Originally Posted by hciman77 View Post
I have ordered a copy of

Nora Jones - come away with me (best)

from Amazon, sadly it wont be hear until after Christmas, how much do you want for shipping ?

This should be interesting. I have two transports feeding a DAC, I will get the wife to load the two CDs into the transports without telling me which is in which and I will switch betwen the two blind. Alternatively I have a changer and I can get the missus to load two adjacent slots but this will introduce a delay.

I will listen to the commercial CD a few times first to get ued to the sound.

I must be honest and say that Jazz really is not one of my normal listening choices, I generally listen to "Classical" and lately a lot of Lute music from John Dowland and contemporaries. I offer this in the spirit of disclosure, though if the effect of jitter is pronounced this should not be an issue.

Why did you choose that particular CD incidentally ? - I would have gone for Mehta's 1975(?) Mahler 2 which has a tremendous dynamic range and is a superb recording for such an old performance.

Regarding your jitter-free copy, I could not get the same reduction from a 1X speed secure rip using EAC and 1X burn and a good quality CD-R ? - just asking.
I have a lot of great classical as well, but the Nora Jones really shows the jitter improvement well. It is a disk that many think was a poor recording. It just has high jitter, that's all.

Copying using EAC is fine, but the writer and media that you use is critical to getting more readable pits and more accurate placement.

Shipping to PA should be $2-3 for parcel post. Dont worry about it. Just email your shipping address to me at:
nugent@empricalaudio.com

BTW, I believe track 10 is the best one for this experiment, the one about the painter.

Steve N.
Reply
post #77 of 112
12/21/06 at 8:21pm
Steve,

thanks for the offer but I am unsure what I would receive. I own both discs and have them both in my hard drive library as well. I have experimented with rewriting CDs in the past and I do have a lot of choice on writers. On My system I could never detect a difference an yes my system is rather resolving. However rather than doing these experiments with disc writers which could introduce all sorts of issues let's focus on the key question here.

Do you agree that most modern CD players use chipsets that drive their digital output directly from a crystal oscillator and do not depend on deriving their clock from the laser pickup or the disc pits in any way shape or form?

See the description of one of the Philips CD chipsets below.

http://www.nxp.com/acrobat_download/.../SAA7806_1.pdf

Then we can just move away from this topic and focus on areas where this is highly relevant.

Cheers

Thomas
Reply
post #78 of 112
12/21/06 at 8:58pm
Thread Starter 
Quote:
Originally Posted by audioengr View Post
nugent@empiricalaudio.com is my PayPal address

Steve N.
Steve,
sorry, your email address bounces with a Domain name not found error. I will try a PM
Reply
post #79 of 112
12/22/06 at 1:43am
Originally Posted by drarthurwells: Your system's source jitter, and other colorations in your sysem, could mask the differences making discrimination difficult between the jitter-free and jitter-containing materials.

Quote:
Originally Posted by hciman77 View Post
Ah, the your stereo isnt good enough to hear the difference clause, I was wondering when that would rear its head again and bang on cue. Of course if I come back and say I do hear jitter you wont say - no you couldnt your stereo isnt good enough.

If added jitter is a problem any competent system should show it up, if it isnt the best system in the world will not. You are obviously worried that I wont hear any difference and are getting the excuses in early, next will be my hearing ability, my taste in music and my training.

Lets recap the study

The experimenters used a computer based system - such systems supposedly less prone to the jitter inducing issues plagued by normal CD systems, thus they could produce a signal with very very low levels of jitter Audioeng himself uses such systems and implies that they are in fact better in jitter terms. His 3:4 test is predicated on this assumption.

Then the experimenters added high levels of jitter to determine at what point it became audibly perceptible. Using their own kit and music they know nobody (inc musicians and audio pros) in blind tests could detect added jitter until it was between 250 and 500 ns. Are you serioulsy suggesting that anybody has discriminative abilities that are 2500 times better i.e enough to detect 100ps. By that premise the Wow and flutter in turntables of even 0.02% i.e several hundred microseconds would be grotesquely unlistenable for these listeners.
I was referring to your system's ability to convey differences in two data sources and not the experimental condition - what is the relevance of your reference to the experimental condition?

I did not say your system was incapable of conveying such differences, just that his was apossiblity that needed to be taken into account.

I distrust any AB comparison on a brief basis because of the unrelaibility of human perception I described previously.

Let both A and B each be daily listening for a week, alternating A and B over two months, and you might find that after such experience a very low level of jitter difference could be discerned and attrributed properly to the two sources. Let A and B be a minute and it will take very large jitter differences to be detected.
Reply
post #80 of 112
12/22/06 at 2:12am
Quote:
Originally Posted by hciman77 View Post
http://www.jstage.jst.go.jp/article/ast/26/1/50/_pdf

The study provides evidence that audible limits on some forms of jitter, they used a random jitter pattern, may actually be somewhat higher than even the Benjamin and Gannon study which indicated a threshold of about 20ns. In the study above no subjects detected jitter of 250ns and only 25% could detect jitter at 500ns.
I think jitter is overrated..
Reply
post #81 of 112
12/22/06 at 3:33am
Quote:
Originally Posted by thomaspf View Post
Steve,

thanks for the offer but I am unsure what I would receive. I own both discs and have them both in my hard drive library as well. I have experimented with rewriting CDs in the past and I do have a lot of choice on writers. On My system I could never detect a difference an yes my system is rather resolving. However rather than doing these experiments with disc writers which could introduce all sorts of issues let's focus on the key question here.

Do you agree that most modern CD players use chipsets that drive their digital output directly from a crystal oscillator and do not depend on deriving their clock from the laser pickup or the disc pits in any way shape or form?
I agree that a crystal oscillator that drives a PLL is the clock source for the output of most CD players.

I looked at this datasheet. I read that there is a PLL for reading the data from the disk and servo control

Also:

The I2S-bus back-end logic runs on this clock. Bclk is also output as part of the I2S-bus
interface. In audio slave mode this clock needs to be programmed exactly at
44100 ´ 2 ´ 16/24/32 Hz (depending on I2S-bus-mode), to get a 1 ´ data rate to the audio
DAC.

I also notice no description of a FIFO buffer for output data flow, just for frame decoding. There is no specification for output jitter either. Seems to me that if it is independent of the jitter on the disk, that they would spec the output jitter, given a perfect clock.

What did I miss?

Steve N.
Reply
post #82 of 112
12/22/06 at 11:52am
Thread Starter 
Quote:
Originally Posted by drarthurwells View Post
Originally Posted by drarthurwells: Your system's source jitter, and other colorations in your sysem, could mask the differences making discrimination difficult between the jitter-free and jitter-containing materials.
Quote:
I was referring to your system's ability to convey differences in two data sources and not the experimental condition - what is the relevance of your reference to the experimental condition?

I did not say your system was incapable of conveying such differences, just that his was apossiblity that needed to be taken into account.
That stills sounds an awful lot like your system isnt good enough, I take no offence by the way but I like to clarify things.

Quote:
Let both A and B each be daily listening for a week, alternating A and B over two months, and you might find that after such experience a very low level of jitter difference could be discerned and attrributed properly to the two sources. Let A and B be a minute and it will take very large jitter differences to be detected.
I have no problem with this method either as long as the jitterred and jitter-free sources are not identified upront as this introduces the inevitable human bias, sorry but it just does.

In fact in a follow up email from the authors they told me before the 2nd experimen they allowed listeners access to the jitter and jitter-free material in advance to allow them to improve their discrimination ability by training. So the results shown in the paper are from listeners who have been allowed some sighted practice for the task.
Reply
post #83 of 112
12/22/06 at 11:58am
Thread Starter 
Quote:
Originally Posted by audioengr View Post
There is no specification for output jitter either. Seems to me that if it is independent of the jitter on the disk, that they would spec the output jitter, given a perfect clock.
Few audio-related companies bother to specify jitter measurements at all, I think it possible that they dont because they just feel it is a non issue
Reply
post #84 of 112
12/22/06 at 1:35pm
Quote:
Originally Posted by hciman77 View Post
Few audio-related companies bother to specify jitter measurements at all, I think it possible that they dont because they just feel it is a non issue
Maybe they are hiding something.
Reply
post #85 of 112
12/22/06 at 4:04pm
Quote:
Originally Posted by Audioengr
I also notice no description of a FIFO buffer for output data flow, just for frame decoding. There is no specification for output jitter either. Seems to me that if it is independent of the jitter on the disk, that they would spec the output jitter, given a perfect clock.

What did I miss?

Steve N.
Hi Steve,

what you missed is the point of these new controllers. There is no PLL involved in the digital output path whatsoever. The outout is clocked with no PLL directly from the clock input.

What happens to the laser assemby and drive mechanics is completely asynchronous. There are probably many PLLs on this chip but none of them anywhere in the digital output path. This chip will actually reread a sector if you drive over a bump.

I apprecite your work on low jitter DAC connections but the pit jitter thing is just not there anymore.

Since the clock is external the resulting jitter is obviously a function of the quality of the clock you supply to the chip. However unlike your modded Transits there is no PLL in the path from the clock input to the bclk output since this clock is the master abd the rest of the chip has to get the data at the rate it is being old by thid clock.

And yes the coupling between the synchronous clock domain of the digital output and the drive assembly is asynchronously via a buffer. If the the buffer runs low the chip speeds up the rotation of the disc, it also compensates for driving around corners and over bumps. That has way more impact on the data rate from the disc than the pit variances but given this design it still does not matter.

Hope this clarifies.

Cheers

Thomas
Reply
post #86 of 112
12/22/06 at 4:59pm
Quote:
Originally Posted by thomaspf View Post
Hi Steve,

what you missed is the point of these new controllers. There is no PLL involved in the digital output path whatsoever. The outout is clocked with no PLL directly from the clock input.

What happens to the laser assemby and drive mechanics is completely asynchronous. There are probably many PLLs on this chip but none of them anywhere in the digital output path. This chip will actually reread a sector if you drive over a bump.

I apprecite your work on low jitter DAC connections but the pit jitter thing is just not there anymore.

Since the clock is external the resulting jitter is obviously a function of the quality of the clock you supply to the chip. However unlike your modded Transits there is no PLL in the path from the clock input to the bclk output since this clock is the master abd the rest of the chip has to get the data at the rate it is being old by thid clock.

And yes the coupling between the synchronous clock domain of the digital output and the drive assembly is asynchronously via a buffer. If the the buffer runs low the chip speeds up the rotation of the disc, it also compensates for driving around corners and over bumps. That has way more impact on the data rate from the disc than the pit variances but given this design it still does not matter.

Hope this clarifies.

Cheers

Thomas
This sounds like it does the job, at least in car and portable stereos. I have not found this type of behavior in home equipment yet. Maybe its there, but the jitter is a result of indirect coupling of the mechanical parts or the PLL? This is what Altmann suggests.

BTW - I dont mod Transits anymore. I sell my own designs now.

Steve N.
Reply
post #87 of 112
12/22/06 at 9:55pm
Hi Steve,

I am glad we can move beyond this discussion. I won't drag out the point whether the current Philips chips sets for home players would go back to an antiquated jitter prone scheme as long as we all agree that a CD controller with an asynchronous assembly decouples the output clock from pits, street bumps, or even rereads.

I like the work you are doing on the USB converters.

Cheers

Thomas
Reply
post #88 of 112
12/24/06 at 1:00am
Ever notice the ones objecting the most are selling things that this scientific study basically claims are useless. The spam on this forum is getting out of hand.
Reply
post #89 of 112
12/24/06 at 1:47am
Quote:
Originally Posted by regal View Post
Ever notice the ones objecting the most are selling things that this scientific study basically claims are useless. The spam on this forum is getting out of hand.
yes, your spam is getting out of hand. many people with a financial interest can still be objective, especially when they brings facts and data to the table. they are welcome, while those who have nothing to contribute are not.
Reply
post #90 of 112
12/24/06 at 2:11am
Quote:
Originally Posted by regal View Post
Ever notice the ones objecting the most are selling things that this scientific study basically claims are useless. The spam on this forum is getting out of hand.
There's an interesting and respectful discussion going on. If you don't have anything to add (and clearly you don't), don't post.
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Dedicated Source Components
Head-Fi.org › Forums › Equipment Forums › Dedicated Source Components › An interesting paper on jitter audibility