[doctorcilantro]

So you are saying the PC contaminates the digital output at the sound card? even with a high quality clock on card....

You do external clocking in different applications that have no bearing on playback, what we are interested in.

You mean if I clock my sound card to an external source, and then send digital out to a DAC. Just trying to clarify what you mean with regard to no bearing on playback.

The Wadia technique seemed akin to adding an external word clock to me; they seperate the two. Seemed as though their diagram indicated jitter creeping on the actual digital audio signal.

DC

Yes, PC contaminates the digital output at the sound card, that is why I am talking about addressing jitter at the source.
I tested this as well, I moved the same sound card between PCs with different HW enabled and SW installed, and consequently with different OS footprint, there were clear differences, PC that I build specifically for playback purpose, the completely fanless one that that had all non-essential HW and SW disabled, had less negative impact on playback.

And no, by different application I mean multiple devices in a digital chain, something you would find in production. Not in playback where DA conversion occurs.

So essentially, if I read you right, a sound card with a very accurate word clock is still susceptible to jitter because of the contaminated (emi) environment?

So would Toslink be preferred over coax?

Conversly, does pro recording of all the high-res and quality audio we enjoy escape this problem running ADCs into computer-based recorders? I use to run a Lavry AD-10 and sync the EMU card to it, then output to a Lavry DA-10. This as opposed to using the internal sync on the EMU; it is the same as adding the aformentioned external word clock.

Here are GR's comments on his usb implementation:

"First we don't fix jitter.... second there is no way to feedback the clock to the computer so with Async we do one better. We control the clock locally. This is what makes it so much better than any product and it does not require drivers or any special hardware. Everything is included....

Word Clock jitter is the specification for which all are based.

When Stereophile tested our Cosecant which uses the same clocks as the Proton they said it was the lowest jitter they ever tested.

All of our products are tested with Wavecrest DTS jitter analysis tool as well as the Prism dScope III and our new TEK MSO.

Believe me it is really low, much lower than most products and 100x better than any adaptive based USB products."

I believe I have some papers/resources on how jitter actually gets into the digital output; if you have any ways it does or solutions please share (still not clear why syncing to an accurate word clock would be adverse - a literal delay as the word clock source isn't directly on the card?).

DC

[doctorcilantro]

Sasha - this seems to pertain to what you mentioned earlier:

"The digital interface has two major functions: It carries the digital data and it carries the sampling clock. Both data and clock are transmitted on the same physical electrical conductors or optical fiber link: data are encoded using a pulse-modulation scheme, and the clock is embedded in the pulse edges. This system requires that the clock and data be separated in the digital processor, a function performed by the digital processor's input receiver. The "recovered" clock then serves as the processor's master clock. Consequently, jitter in the interface data stream produces clock jitter at the DAC. This is the mechanism by which transports affect a digital system's sonic performance. Moreover, the quality of the digital interface implementation greatly affects the amount of jitter in the recovered clock."

http://www.stereophile.com/reference/1093jitter/index1.html

And this...yikes!

As with noise, jitter has multiple causes. One of them is electrical noise picked up on the interface cable. Noise causes the zero crossing points to shift slightly?which is, by definition, interface jitter. If the digital interface signal has an average slope of 20V/?s (a typical value in many S/PDIF implementations), just 10mV RMS of noise, for example, will introduce 500ps RMS of jitter.

I'm looking into the Proton. I'm on my third sound card capable of 96kHz and trying to isolate a sync problem with the Peachtree NOVA. It is a 96kHz upsampler and feeding it native or PC-upsampled 96khz material is resulting in digital artifacts every now and then. I never had any issues when I was using a Lavry DA10 which was not an upsampler.

DC

[Sasha]

So essentially, if I read you right, a sound card with a very accurate word clock is still susceptible to jitter because of the contaminated (emi) environment?

Yes, EMI/RFI and dirty power

So would Toslink be preferred over coax?

No. It is jittery as any other digital output, galvanic isolation does not help you with jittery signal out of sound card.
I addition, I compared SPDIF to TosLink on multiple devices that had both, and TosLink was always inferior.
Measurements also confirm this.
It seems electro-optical converters  make it more jittery.

Conversly, does pro recording of all the high-res and quality audio we enjoy escape this problem running ADCs into computer-based recorders? I use to run a Lavry AD-10 and sync the EMU card to it, then output to a Lavry DA-10. This as opposed to using the internal sync on the EMU; it is the same as adding the aformentioned external word clock.

It seems to me that pro recording would not have nasty PC in the recording chain, and it does not play such a role anyway in AD conversion, you are already in digital domain by the time you start manipulating digital signal on PC, as long as you remain in digital domain you have no issues with jitter.

Here are GR's comments on his usb implementation:

"First we don't fix jitter.... second there is no way to feedback the clock to the computer so with Async we do one better. We control the clock locally. This is what makes it so much better than any product and it does not require drivers or any special hardware. Everything is included....

Word Clock jitter is the specification for which all are based.

When Stereophile tested our Cosecant which uses the same clocks as the Proton they said it was the lowest jitter they ever tested.

All of our products are tested with Wavecrest DTS jitter analysis tool as well as the Prism dScope III and our new TEK MSO.

Believe me it is really low, much lower than most products and 100x better than any adaptive based USB products."

I am not interested in product that has limitations (96kHz sampling rate), nor in product that uses transformers in analog section.
Also, if all that was true, different PCs, different USB controllers, different USB cables would not impact performance, yet they do, it is widely reported. I do not see how USB is any better.
I prefer to stay with products from companies that have good engineering talent, Bryston being one of them.
Bryston digital products show verifiable performance.
If Bryston developed reasonably priced version of Linn DS or Transporter it could be a great step forward in performance.
There is a lot of nonsense in advertisement and on forums, it is hard to figure out what represents real breakthrough and what is just another flavor of one the same thing.
You see, I sold my Wadia and got BDA-1, it is not to say that BDA-1 fed from whatever source I have is better performer, it is certainly not, but it offers features I am interested in, in some aspects it rivals Wadia, and it is true value for the money.
To put it into the perspective of the subjects we discussed, the difference between BDA-1/BCD-1 and Wadia 581 is of the same magnitude as the difference between the top CDPs and best executed PC/DACs.
I would have to spin disks to get top performance, yet I want music to be more readily available. Since I have to compromise Wadia with jittery source what do I need Wadia for?
I want to take PC based transport to the level that would give me the performance of tray in Wadia.
That is why I am not interested in those gizmos and limited use devices, regardless of how much someone may enjoy them and be ignorant of top performance that such devices are so far from.
That is why I started this tread to see if James had opportunity to hear Amarra, to get a sense of how real the fuss is. I do not want to get engaged in discussion on how MAC/TosLink/iTunes are great, I am interested in much higher level of performance.
Hope this explains all my driveling.

[Sasha]

Sasha - this seems to pertain to what you mentioned earlier:

"The digital interface has two major functions: It carries the digital data and it carries the sampling clock. Both data and clock are transmitted on the same physical electrical conductors or optical fiber link: data are encoded using a pulse-modulation scheme, and the clock is embedded in the pulse edges. This system requires that the clock and data be separated in the digital processor, a function performed by the digital processor's input receiver. The "recovered" clock then serves as the processor's master clock. Consequently, jitter in the interface data stream produces clock jitter at the DAC. This is the mechanism by which transports affect a digital system's sonic performance. Moreover, the quality of the digital interface implementation greatly affects the amount of jitter in the recovered clock."

http://www.stereophile.com/reference/1093jitter/index1.html

And this...yikes!

As with noise, jitter has multiple causes. One of them is electrical noise picked up on the interface cable. Noise causes the zero crossing points to shift slightly?which is, by definition, interface jitter. If the digital interface signal has an average slope of 20V/?s (a typical value in many S/PDIF implementations), just 10mV RMS of noise, for example, will introduce 500ps RMS of jitter.

I'm looking into the Proton. I'm on my third sound card capable of 96kHz and trying to isolate a sync problem with the Peachtree NOVA. It is a 96kHz upsampler and feeding it native or PC-upsampled 96khz material is resulting in digital artifacts every now and then. I never had any issues when I was using a Lavry DA10 which was not an upsampler.

DC

Well, you found answers to all of your questions....

I am not sure I understand the sync problem, can you elaborate a bit more?

[doctorcilantro]

Benchmark's claim of "immunity to jitter" may be referring to whether their converter will lose lock or not, as versus the effects of jitter on the reproduced digital audio material. Many of our DA10 customers were former DAC-1 owners, and expressed that they felt the DA10 offered better sonic performance. This corresponds to the opinion of independent reviewers, as well. So yes, you can design an input circuit that is capable of dealing with a large amount of jitter and not lose lock; but this does not mean that there is no other effect on the over-all performance of the DA converter.

Lavry on jitter.

I think the NOVA loses lock on the Coax S/PDIF 96kHz material I'm sending via the EMU1616M. I have to do more tests and it could end up being driver related as I am using Windows 7 x64; I could have a bad 1616M as well, it was dropped once 

The BDA-1, if fed a 192kHz sample rate with upsampler turned on..does what? Would the upsampler simply need to be turned off? I'm unclear how upsamplers handle the same format they are upsampling to (in the case of the NOVA).

thanks,
DC

[Sasha]

OK, if I understand correctly your E-MU supports ASIO, and you installed the latest driver for the platform?
Have you made appropriate selection in whatever player you use?
Did you install card into appropriate PCI slot (some cards require specific slots)?
Try to adjust buffer settings in player.
If still have loss/pops try to free up memory and processor by unloading unnecessary drivers, stopping unnecessary processes.
Search for posts from cics on cplay and PC optimization in Computer Audio Asylum on http://www.audioasylum.com, there is plenty of information that could help you.
But if it all worked with Lavry, your Nova may be the culprit?
BTW, when I read something like ?patented jitter reduction circuit re-clocks the digital signal to almost 0 jitter? I simply walk away, it is utter rubbish.
Or "The tube maintains a high level of detail while taming harsh digital files"?
It seems to me Nova is yet another one of those PC gizmos, good as an all in one inexpensive solution for PC sound, but far from decent performance, let alone the performance they try to convince you of (they compare Nova with DACs in $3,500-$10,000 range).
Upsampler on BDA-1 does nothing when you bring 192kHz, and you do not need to turn it off.

[doctorcilantro]

Thanks for the tips. I've been messing with Windows for audio playback since 2000 (and before that when you had to cross your fingers when burning a CD - some still ended up with massive artifacts) when we really needed to tweak things big time on that limited PCI bus, so this has been a real pain; feels like going back a decade in time.

I'm slowly adjusting things and now listening to the NOVA headphone output sending 96kHz from onboard Realtek Toslink. If I don't hear any issues using this device, I'm going to point the finger at the EMU card first. This is not the exact same card or OS as my last one (Windows XP & EMU1212M).

I wouldn't right it off completely. While their PR campaign is devoid of specs (manual even says: detailed specs were not available at time of printing!) and definitely non-audiophile, Scott Nixon does a lot of design for them, and did the Decco before this. I agree with the rather large claims being annoying but it is a class A tube stage. Just because it has the Sabre chip doesn't mean it is guaranteed to be implemented well, but so far it sounds great to me.

I'm still waiting for my Omega speakers and I realize that these powered monitors may even be the culprit; maybe the digital artifact is actual a psu issues on the monitors!

DC

[doctorcilantro]

OK, if I understand correctly your E-MU supports ASIO, and you installed the latest driver for the platform?
Have you made appropriate selection in whatever player you use?
Did you install card into appropriate PCI slot (some cards require specific slots)?
Try to adjust buffer settings in player.
If still have loss/pops try to free up memory and processor by unloading unnecessary drivers, stopping unnecessary processes.
Search for posts from cics on cplay and PC optimization in Computer Audio Asylum on http://www.audioasylum.com, there is plenty of information that could help you.
But if it all worked with Lavry, your Nova may be the culprit?
BTW, when I read something like ?patented jitter reduction circuit re-clocks the digital signal to almost 0 jitter? I simply walk away, it is utter rubbish.
Or "The tube maintains a high level of detail while taming harsh digital files"?
It seems to me Nova is yet another one of those PC gizmos, good as an all in one inexpensive solution for PC sound, but far from decent performance, let alone the performance they try to convince you of (they compare Nova with DACs in $3,500-$10,000 range).
Upsampler on BDA-1 does nothing when you bring 192kHz, and you do not need to turn it off.

THE BRYSTON SOLUTION

Bryston delivers superb sonic performance by re-sampling and re-clocking the digital input in order to reduce jitter. The result, a significant reduction in jitter (1/1000 of a nanosecond). But it isn?t enough to just get the bits right; those bits have to be converted back into music with the same timing reference as when the music was first digitized. The input signal of the BDA-1 is re-clocked and re-sampled to reduce any possibility of jitter affecting the sound quality. Even the input receiver and the sample rate converter serve to further reduce jitter.

UR?

[Crimson]

That is why I started this tread to see if James had opportunity to hear Amarra, to get a sense of how real the fuss is. I do not want to get engaged in discussion on how MAC/TosLink/iTunes are great, I am interested in much higher level of performance.

Come on over to the Apple Core and read about it. Folks that are using Amarra (myself included) are not using TosLink. You, OTOH, may be limited to Toslink or adaptive mode USB with the BDA-1.

[Sasha]

On ?taming harsh digital files?, you do not address problems of such nature in pre-amp, but again at the source, your digital frontend should provide decent analog signal to the extent possible considering the limitations of technology, and assuming the problem was not with the recording in the first place. Pre-amp should be as transparent as possible, not to ?tame? anything. I am not seeking colorations.
On jitter reduction, there is no 0 jitter, never will be, not in this universe.
Claims of 1ps are stretched as well, if that was true I believe it would present a value below audible threshold, yet you hear differences between different sources.
If this attenuation method works so well and no matter what signal you bring at the input you end up with 1ps, would you not think that all sources should sound the same?
You got to be a skeptic when wondering through all the marketing material in search for the right equipment.

[James Tanner]

Hi All,

I have been searching hi and low for any scientific info on the detectability of Jitter in digital music playback and so far other than the Dolby study (see link above) there is not much out there. If anyone knows of any valid research on this subject please post.

The Dolby study claims anything below 4NS is not detectable. That seems a little on the high side to me but Dolby is no feather weight when it comes to research.. Most sources I have spoken too (all though antidotal) seem to feel below 400PS is required for most not to hear the negative affects of Jitter?

Also, there are many ways of measuring jitter. What's the measurement bandwidth? Is the jitter "white" or signal-correlated? Is the measurement RMS or peak? There's no standard spec.
 
james

[doctorcilantro]

On ?taming harsh digital files?, you do not address problems of such nature in pre-amp, but again at the source, your digital frontend should provide decent analog signal to the extent possible considering the limitations of technology, and assuming the problem was not with the recording in the first place. Pre-amp should be as transparent as possible, not to ?tame? anything. I am not seeking colorations.
On jitter reduction, there is no 0 jitter, never will be, not in this universe.
Claims of 1ps are stretched as well, if that was true I believe it would present a value below audible threshold, yet you hear differences between different sources.
If this attenuation method works so well and no matter what signal you bring at the input you end up with 1ps, would you not think that all sources should sound the same?
You got to be a skeptic when wondering through all the marketing material in search for the right equipment.

The "tame digital harshness" is of course Bose-esque PR for the masses. It's a Class A tube preamp, end of story.

I was referring to the fact that Bryston uses the same terms: re-clocking, low jitter, etc.

You seem to reaching for something truly elusive. How long does a piece of equipment stay in rotation in your system? 

What equipment are you using now? I would still like to hear how you ABX'd various OS/PC sources; how many at a time? Or was it from memory (subjective). You have made a lot of empirical claims, yet it's not clear what other equipment/validity controls were used in your purported random experimental design.

DC

[Crimson]

Hi All,

I have been searching hi and low for any scientific info on the delectability of Jitter in digital music playback and so far other than the Dolby study (see link above) there is not much out there. If anyone knows of any valid research on this subject please post.

The Dolby study claims anything below 4NS is not detectable. That seems a little on the high side to me but Dolby is no feather weight when it comes to research.. Most sources I have spoken too (all though antidotal) seem to feel below 400PS is required for most not to hear the negative affects of Jitter?

Also, there are many ways of measuring jitter. What's the measurement bandwidth? Is the jitter "white" or signal-correlated? Is the measurement RMS or peak? There's no standard spec.
 
james

James,

Another thing to consider is that jitter has become the current catch-all-du-jour for all things inherently wrong with digital: EMI, RFI, switching power supplies, etc, etc, etc, .......ad nauseum.

The fact is, as you are probably fully aware, that jitter is only introduced into the signal chain during the A-D (recording) and D-A (playback) processes at the hardware level, i.e. during a physical medium transition. No more, no less.

Jitter is not a function of the other factors inherent in the signal chain that could potentially add noise and/or artifacts to the signal, yet we audiophiles have become obsessed with............jitter.

As to it's audibility, I agree with AES paper.

[James Tanner]

Hi Crimson,

I agree and I have been saying for a long time that the reason we have been getting such excellent response to our CD Player and BDA-1 DAC (not excluding the very low jitter rates) is because of the attention we pay to other just as important details in the design of a product.

Things like independent power supplies for analog and digital circuits, independent ground plans for digital and analog circuits, discrete Class A analog output section, transformer coupled inputs etc. go a long way to maintain signal integrity throughout the product.

james
 

[Sasha]

On ?taming harsh digital files?, you do not address problems of such nature in pre-amp, but again at the source, your digital frontend should provide decent analog signal to the extent possible considering the limitations of technology, and assuming the problem was not with the recording in the first place. Pre-amp should be as transparent as possible, not to ?tame? anything. I am not seeking colorations.
On jitter reduction, there is no 0 jitter, never will be, not in this universe.
Claims of 1ps are stretched as well, if that was true I believe it would present a value below audible threshold, yet you hear differences between different sources.
If this attenuation method works so well and no matter what signal you bring at the input you end up with 1ps, would you not think that all sources should sound the same?
You got to be a skeptic when wondering through all the marketing material in search for the right equipment.

The "tame digital harshness" is of course Bose-esque PR for the masses. It's a Class A tube preamp, end of story.

I was referring to the fact that Bryston uses the same terms: re-clocking, low jitter, etc.

You seem to reaching for something truly elusive. How long does a piece of equipment stay in rotation in your system? 

What equipment are you using now? I would still like to hear how you ABX'd various OS/PC sources; how many at a time? Or was it from memory (subjective). You have made a lot of empirical claims, yet it's not clear what other equipment/validity controls were used in your purported random experimental design.

DC

Using terms is not an issue, I am not saying that reclocking and jitter attenuation are not happening, just that claims of complete jitter removal, or removal to such low levels are not objective.
You have to ask yourself why sources sound differently if jitter was removed to such low level.
And how do you measure such low level of jitter?
Stereophile review that you posted link to talks about resolution limit of the measuring equipment used.
What I am looking for is not elusive, what technological obstacle is there to bring the performance of PC based source to the level of best CDPs? None, it will happen sooner or later.
Here is again how I arrived to the conclusion that PC based transport does not match CDP.
Rig consisted of PMC IB2, Bryston BP26, Wadia 581i SE (tried both into the pre-amp and directly into amps), PC based on Zalman TNN-300 (no fans at all, just heatpipes and heatsinks), 2.5? low power consumption HD for OS, 3.5? low power consumption HD for wave storage, Lynx L22 soundcard, native ASIO support, all non-esential HW disabled (USB, Fireware, NIC, etc.), no CD/DVD (ripping done on another PC), all non-essential services disabled, the lowest achievable XP OS footprint, Steinberg Wavelab used for playback. PC 100% dedicated to just one function.
So I take good recordings, good quality CDs, rip them, transfer files to PC.
Play the same track from Wadia tray, then from PC into Wadis?a digital input.
Tray is better.
Burn the same track on CDR, play it from Wadia tray.
Tray is better.
End of story.
The only explanation is jitter, again read info on clocklink, and on double PLL implementation on Wadia I/O board.
As much as it is desirable to do DA conversion outside of hostile environment so is generation of digital signal.
In theory squeezebox should work like miracle, but is utter crap, and only due to poor engineering and execution.
Then different digital interfaces, tested in the same manner, CD player with both coax and Toslink into Wadia, Toslink not as good, TV receiver with coax and Toslnk, again Toslink is not good.
Universal player with both coax and Toslink into yet another DAC, Toslink again not as good.
Those that claim the opposite (PC surpasses CDP, Toslink is the way to go, all sources sound the same, etc.) actually never did any comparison, or if they did it was done in different systems, or based on memory.

[Sasha]

Jitter audibility threshold of 4ns is nonsense.
In what context, what application?
In IP telephony maybe, it is true you cannot distinguish jitter measured in ms, but what is your bandwidth, what is the codec, what is the resolution of you HW in such application?
EMI, RFI, switching power supplies, etc. is exactly what causes jittery digital signal and plays havoc during DA conversion in this application.
Can any of the naysayers offer at least a theoretical explanation of why for example different digital sources sound differently in exactly the same chain, or why different protocols sound differently in exactly the same chain?
There is only one variable you are changing, so if it is not the resulting jitter, what is it?

[James Tanner]

Jitter audibility threshold of 4ns is nonsense.
In what context, what application?
In IP telephony maybe, it is true you cannot distinguish jitter measured in ms, but what is your bandwidth, what is the codec, what is the resolution of you HW in such application?
EMI, RFI, switching power supplies, etc. is exactly what causes jittery digital signal and plays havoc during DA conversion in this application.
Can any of the naysayers offer at least a theoretical explanation of why for example different digital sources sound differently in exactly the same chain, or why different protocols sound differently in exactly the same chain?
There is only one variable you are changing, so if it is not the resulting jitter, what is it?

Hi Sasha,

Based on your experience at what level do you feel you can detect jitter?

james

[doctorcilantro]

Okay, disclaimer, the following thoughts are friendly.

Plays havoc?

I mean seriously, you make PC audio sound BAD BAD BAD.

Your standard is set rather high. Changing speakers is likely to have a larger affect on the fidelity than the source.

Usability must be balanced with fidelity. I'm simply not going to cue up single .wav files in Wavelab. I use it for archiving/mastering vinyl, not playback. Maybe you can hear an extra 200ps of jitter as another word being said in a song  I understand why you used it and maybe it seems off topic of me to mention it, but....

Your whole endeavor while possibly scientifically valid (did you ABX or just AB), is in some ways fruitless, in other ways, not so. You say "tray is better". What does that mean? People change power cords and say "it sounds better". Did you take notes since you did such an elaborate comparison? It's still subjective unless you ABX it.

Let's assume there is a marked difference, however minute, maybe an overall clearer quality or something - bottom line, as you stated earlier, there is a level of convenience with PC-audio and in this case, possibly a sacrifice of fidelity. However, I don't switch on my system and cringe, and I don't give up the ability to easily compare songs, control playback with iPhone, tag and display my meta-data, stream lossless audio to my office over WAN, etc. You made this point before, and what comes up again and again in this strange future of PC-audio is usability, however indirectly related to sound quality, it is often a primary concern. I'm not trying to same PC-audio is better than CDP, but  what constitutes better? If it means given up all the conveniences of PC-audio, then a CDP is not better; if a truly audiophile pursuit then maybe CDP is better.

I will do what's reasonable to improve my system starting from the foundation of PC audio; maybe you have kept one system that has a PC source since you put in all the effort? I'd still like to hear what the actual difference were.

Thanks for your patience and input Sasha.

DC

[Sasha]

Jitter audibility threshold of 4ns is nonsense.
In what context, what application?
In IP telephony maybe, it is true you cannot distinguish jitter measured in ms, but what is your bandwidth, what is the codec, what is the resolution of you HW in such application?
EMI, RFI, switching power supplies, etc. is exactly what causes jittery digital signal and plays havoc during DA conversion in this application.
Can any of the naysayers offer at least a theoretical explanation of why for example different digital sources sound differently in exactly the same chain, or why different protocols sound differently in exactly the same chain?
There is only one variable you are changing, so if it is not the resulting jitter, what is it?

Hi Sasha,

Based on your experience at what level do you feel you can detect jitter?

james

I do not know, and mainly because you cannot trust published figures, if you can find them in the first place.
I am not saying the unsubstantiated information is intentionally published as such, take for example the Stereophile review of players where Ayre was the best scoring one. Charles is saying that the figure should be much lower (I think figure was below 100ps but I could be wrong) and that it was due to the resolution limit of the measuring equipment that Ayre did not score even better. Yet Charles said openly that he could not measure and confirm the claimed figure himself as he did not have such sensitive equipment (yet), and that his figure was based on calculations.
The only figures I am certain of at this time is that the difference between 400ps and 800ps is very audible and you do not need some very resolute system to hear it, nor golden ears.
But also consider that it is not all in numbers, but what the receiver does, and how jitter gets transformed and translated into analog signal during conversion.
For example, take DAC A and B, A shows audibly much smaller differences between different sources or between different inputs. Does it mean that A is better than B because it employs some better jitter attenuation technique? No, my experience so far has been that in such instances some monkey business is going on, and it is the component that masks differences, sometimes caused by multiple factors (their implementation of upsampling, poor analog section, poor PSU, etc.) Things like that eventually come to the light as more and more people use given equipment and objective measurements appear.
Finally I do not have limitless funds to compare sufficient number of players and draw conclusions from it such as the audible threshold for jitter.
Everything else is speculation.
Take for example the Lynx cards. I have seen some fantastic numbers claimed for digital output.
But they do not say under what circumstances.
So what does it mean when I hear the difference between PC/Lynx card and tray, that I can hear such low jitter or that the publisher is not telling the truth for whatever reason?

[Sasha]

Okay, disclaimer, the following thoughts are friendly.

Plays havoc?

I mean seriously, you make PC audio sound BAD BAD BAD.

Your standard is set rather high. Changing speakers is likely to have a larger affect on the fidelity than the source.

Usability must be balanced with fidelity. I'm simply not going to cue up single .wav files in Wavelab. I use it for archiving/mastering vinyl, not playback. Maybe you can hear an extra 200ps of jitter as another word being said in a song  I understand why you used it and maybe it seems off topic of me to mention it, but....

Your whole endeavor while possibly scientifically valid (did you ABX or just AB), is in some ways fruitless, in other ways, not so. You say "tray is better". What does that mean? People change power cords and say "it sounds better". Did you take notes since you did such an elaborate comparison? It's still subjective unless you ABX it.

Let's assume there is a marked difference, however minute, maybe an overall clearer quality or something - bottom line, as you stated earlier, there is a level of convenience with PC-audio and in this case, possibly a sacrifice of fidelity. However, I don't switch on my system and cringe, and I don't give up the ability to easily compare songs, control playback with iPhone, tag and display my meta-data, stream lossless audio to my office over WAN, etc. You made this point before, and what comes up again and again in this strange future of PC-audio is usability, however indirectly related to sound quality, it is often a primary concern. I'm not trying to same PC-audio is better than CDP, but  what constitutes better? If it means given up all the conveniences of PC-audio, then a CDP is not better; if a truly audiophile pursuit then maybe CDP is better.

I will do what's reasonable to improve my system starting from the foundation of PC audio; maybe you have kept one system that has a PC source since you put in all the effort? I'd still like to hear what the actual difference were.

Thanks for your patience and input Sasha.

DC

I agree with almost everything you said, speakers definitely have a larger affect, but I would not go that far and say larger than the source in general, rather quite larger than the differences being discussed here.
And yes on usability balanced with fidelity, after all you can see that I kept the dreaded PC and got rid of the tray, Wavelab was mentioned just so that questions on SW player were not raised.
Now, on the subject of ?sounding better?, there is different sound and better sound, when you consider the reference, which for me is the sound of live unamplified acoustic instruments and human voice.
So far I was talking about better sound, simply because one knows how the reference sounds, is familiar with specific recordings, even has recording of live performance witnessed at the time of the recording.
When I say tray is better, I mean it is closer to the reference.
My main reason for all this discussion is to find out if someone has come across PC based solution that is equal to CDP, reasonably priced (we are again in agreement). For example, in theory I may be convinced that Linn DS is the ticket, but I am not going to spend that amount of money just to try it.
I am not trying to convince anyone to get rid of whatever they have and pursue something else, or anything along those lines, I am only asking if someone in the same pursuit as I am has come across potential solution. Sometimes the discussion gets derailed by naysayers who want to impose their view on me and convince that what I am hearing is my imagination.
Then we start going in all directions with discussion.