Chord Qutest

[ElviaCaprice]

2 hours ago, barrows said:

I work with JS, and I have never heard him say anything negative regarding the proper implementation of a USB interface in a DAC.  Perhaps you have mis-interpreted him...

Is possible?  Like I said, I have no listening experience to suggest it doesn't work properly.  But JS did do some extensive posting on GI and how it can differ greatly in actual application.  There have been other postings from Alex on GI in application with the ISO Regen and it's issues.  Also there have been posts from Roy in consultation with SOtM engineers and how they view GI negatively and thus you never see them use GI but instead prefer filtering.

[barrows]

20 minutes ago, ElviaCaprice said:

Various postings from John Swenson here on the forum.

I work with JS, and I have never heard him say anything negative regarding the proper implementation of a USB interface in a DAC.  Perhaps you have mis-interpreted him...

[ElviaCaprice]

2 hours ago, barrows said:

I work with JS, and I have never heard him say anything negative regarding the proper implementation of a USB interface in a DAC.  Perhaps you have mis-interpreted him...

Great, then you can ask him directly next time you see him.  I am only drawing on his past postings here on the forum which I find difficult to search for.  Arggg

 Also alot of what Alex has posted on GI concerning the Iso Regen would be information coming from JS.

[barrows]

17 minutes ago, ElviaCaprice said:

Is possible?  Like I said, I have no listening experience to suggest it doesn't work properly.  But JS did do some extensive posting on GI and how it can differ greatly in actual application.  There have been other postings from Alex on GI in application with the ISO Regen and it's issues.  Also there have been posts from Roy in consultation with SOtM engineers and how they view GI negatively and thus you never see them use GI but instead prefer filtering.

What I am talking about has no relationship to ISO-Regen, I am talking about an isolated USB receiver section in a DAC: in such an interface the USB signal is not running through isolators, the I2S signal is.  Here is the technical description (for example our USB receiver will be an XMOS chip):

 

USB input-XMOS chip, powered from USB 5VDC-I2S-isolators-Reclocking and masterclock(s), powered from dedicated "clean" power supply-I2S-DAC chip.

 

In this example, the jitter inherent in the the output of the XMOS chip is removed on the "clean" side, by direct reclock from the masterclock, right before the I2S lines input to the DAC chip (DAC chip for example only, could be discrete DAC, whatever).  This approach insures jitter at the DAC chip is as low as possible.  This approach also reduces processor noise from the XMOS chip and USB source, keeping it away from the masterclocks and DAC.

[ElviaCaprice]

And that gets to the crux of the matter.  Does the reclocking by the masterclock remove all the negative effects caused by the GI?  Or is there some left over effects that are passed on and another reclocking would be needed?  I wouldn't be so sure that it does.

[barrows]

1 minute ago, ElviaCaprice said:

And that gets to the crux of the matter.  Does the reclocking by the masterclock remove all the negative effects caused by the GI?  

Yes.  That is the entire point.  The only "negative effect" is jitter.  By re-clocking direct from a local, free-running, masterclock jitter at the DAC chip is reduced to that which is inherent in the masterclock, plus slight additions for parasitic elements etc. (which would be present no matter what).

If you do not believe in the effectiveness of re-clocking, you might as well not believe in USB audio either, as it relies on clocking the samples out of a buffer via a clock: if you believe this cannot be done well, then basically forget about digital audio entirely.

[ElviaCaprice]

2 hours ago, barrows said:

Yes.  That is the entire point.  The only "negative effect" is jitter.  By re-clocking direct from a local, free-running, masterclock jitter at the DAC chip is reduced to that which is inherent in the masterclock, plus slight additions for parasitic elements etc. (which would be present no matter what).

If you do not believe in the effectiveness of re-clocking, you might as well not believe in USB audio either, as it relies on clocking the samples out of a buffer via a clock: if you believe this cannot be done well, then basically forget about digital audio entirely.

Oh, I believe strongly in the effectiveness of clocking/power/isolation, but I also believe that there are properties that have yet to be scientifically explained that effect an audio stream.  Like I said, I have no testing proof that the GI doesn't do all that is claimed by Chord.  In fact, it is probably one of the better implemented GI's.  Wouldn't stop me from buying one with GI.  But I shall remain skeptical and i gave my reasoning.

[barrows]

In the I2S data there are only two things, the data, and the jitter.  The data is either present or it is not.  If data is lost there will be an audible "tic" or dropout.  As far as the I2S data is concerned that is the only possibility (if there is a "problem").  as long as one does not have a "tic" or dropout, the data is not corrupted.

Jitter is a separate issue and has already been discussed.

There is no "subtle data integrity loss" which could result in a sonic change, you either have continuous music, indicating no data problem, or you have "tics" or dropouts, nothing in between.

[ElviaCaprice]

It doesn't have to be either black or white, only.  I just don't agree with this mentality and all of yesterdays science won't convince me otherwise.  We can argue all day and it will come down to this.  

[Audio_ELF]

On 11/01/2018 at 7:00 PM, beerandmusic said:

 

Agree...a Chord network player....wishful thinking!

Why! They’ve done it before (okay maybe not the best implemented network players).

 

However I suspect if you think of something more a cross between the Blu 2 and the Poly then you’re probably on the right lines to add network streamer / player to the Qutest (and Hugo2 / TT and even DAVE). 

[barrows]

1 hour ago, ElviaCaprice said:

It doesn't have to be either black or white, only.  I just don't agree with this mentality and all of yesterdays science won't convince me otherwise.  We can argue all day and it will come down to this.  

Yes it does, this is not a "mentality" it is a technical fact.  I will not argue it, as it is inarguable.  Sorry no "alternate facts" here.

[ElviaCaprice]

2 hours ago, barrows said:

Yes it does, this is not a "mentality" it is a technical fact.  I will not argue it, as it is inarguable.  Sorry no "alternate facts" here.

Are you really that full of yourself?  The concept of zero, a set, a point and two parallel lines never intersect  are not technical facts.  They are accepted truths, unproven.  

I look forward to innovative scientific findings, not back seat coaching.

[barrows]

21 minutes ago, ElviaCaprice said:

Are you really that full of yourself?  The concept of zero, a set, a point and two parallel lines never intersect  are not technical facts.  They are accepted truths, unproven.  

I look forward to innovative scientific findings, not back seat coaching.

I am in the front seat.  Just because we may not understand all things in audio from a technical perspective is no reason to dismiss the things we actually do fully understand.

[ElviaCaprice]

2 hours ago, barrows said:

Just because we may not understand all things in audio from a technical perspective is no reason to dismiss the things we actually do fully understand.

I don't disagree with that.  

[beerandmusic]

6 hours ago, barrows said:

There is no "subtle data integrity loss" which could result in a sonic change, you either have continuous music, indicating no data problem, or you have "tics" or dropouts, nothing in between.

 

This is very interesting to me.  I wasn't aware of that.  Are you suggesting that if i am playing native dsd 256 and that if i don't hear any "tics or dropouts" that I can assume that every bit of information was processed correctly?

 

I always thought that bits could be lost and we will miss some actual sonics, details, depth, or other "non-perfect" playback.  I didn't know that if bits were lost due to clocking, noise, or whathave you, that you would always here a tic or something....i have never ever heard any "tics" or dropouts, so i guess that is a good thing.

 

Is what you are saying "known as fact" that all engineers agree with, or is that just your belief?

 

[barrows]

1 hour ago, beerandmusic said:

 

This is very interesting to me.  I wasn't aware of that.  Are you suggesting that if i am playing native dsd 256 and that if i don't hear any "tics or dropouts" that I can assume that every bit of information was processed correctly?

 

I always thought that bits could be lost and we will miss some actual sonics, details, depth, or other "non-perfect" playback.  I didn't know that if bits were lost due to clocking, noise, or whathave you, that you would always here a tic or something....i have never ever heard any "tics" or dropouts, so i guess that is a good thing.

 

Is what you are saying "known as fact" that all engineers agree with, or is that just your belief?

 

Yes, if you do not hear any little "tics" or dropouts, there is no lost data.  Almost all computer audio products in common use by computer audiophiles will not ever have any data loss.  The same cannot be said, for say, Bluetooth (most of us who have ever used Bluetooth music transfer have heard "tics" and dropouts, right?)  Same is true for Ethernet and USB.

 

Sound quality differences are not about data loss, or data integrity problems, all sound quality differences are related to noise issues affecting digital audio clocks, DAC chips, and the analog sections of DACs.

 

Addendum:  It does not mean that the data was processed correctly, if what you mean by processed is converted to analog.  It means that all the data got through the interface to the DAC, where sonic problems happen is in the conversion to analog.

[beerandmusic]

12 minutes ago, barrows said:

Yes, if you do not hear any little "tics" or dropouts, there is no lost data.  Almost all computer audio products in common use by computer audiophiles will not ever have any data loss.  The same cannot be said, for say, Bluetooth (most of us who have ever used Bluetooth music transfer have heard "tics" and dropouts, right?)  Same is true for Ethernet and USB.

 

Sound quality differences are not about data loss, or data integrity problems, all sound quality differences are related to noise issues affecting digital audio clocks, DAC chips, and the analog sections of DACs.

 

damn that noise (grin)!

 

kind of funny when you think about it...noise ruins music.

 

I always thought before, that the noise could affect the accuracy of the data before it reached the flip-flops...I figured the data would be accurate after the flip flops, but was concerned that the data could be corrupted somewhere between the receiver and the actual dac circuitry.

 

I knew the circuitry up to the receiver had to have been perfected long ago, otherwise we would have some serious banking problems.  But not sure what happenned between the reciever and the time it reached the D-A circuitry.  I would "hope" that any DAC engineer could ensure the data integrity after the reciever, but nothing would surprise me.  My logic tells me that we "should be able to have perfection all the way to the output of the DAC chip itself, and that only the output section would really make a difference in SQ.  But i am a 1's and 0's kind of guy, and don't know the first thing about analog.

 

 

[barrows]

Any data problem, dropped samples, will result in an audible "tic" or dropout.  There is not subtle degradation of sonics caused by data integrity problems, as there is no error correction: any missing sample will be missing, and result in a "tic" or dropout.

The very earliest USB audio interfaces had these problems, and dropouts were common, but these problems were solved long ago.

[beerandmusic]

2 minutes ago, barrows said:

Any data problem, dropped samples, will result in an audible "tic" or dropout.  There is not subtle degradation of sonics caused by data integrity problems, as there is no error correction: any missing sample will be missing, and result in a "tic" or dropout.

The very earliest USB audio interfaces had these problems, and dropouts were common, but these problems were solved long ago.

 

That said, above you stated that the noise causes issues with the clocking....

So, hypothetically speaking, if there was NO noise, then there would be no issues with clocking?

The only reason we need better clocks is to deal better with noise?  There isn't any issues with cheaper clocks if there wasn't any noise?   And the power problems are all related to noise as well?  A cheap power supply would be fine, if we had no noise?  The only reason for a LPS over a cheap power supply is because they are less succeptable to voltage variances due to noise?

Are all of these true statements?

 

If so, you would think that more effort would be put into eliminating noise?  And noise really isn't anything more than variances in voltage.  I would think we should be able to produce "perfect voltage" by now....

 

I guess i should stick to 1's and 0's.

[beerandmusic]

^^^ The more i think about this, the more it drives me crazy to think we haven't achieved "sound perfection up to the output stage....

 

I know we can get perfect data up to the DAC reciever, otherwise our banking system would fail.

Timing I can see as a potential  issue, but should be easily overcome with any buffer.

 

We should be able to accomplish perfection with a USB interface, regardless of any noise on the line, again otherwise the banking system would have failed. 

 

With a correct buffer, and the pc usb ground isolated, the only real PS we need to be concerned with is the PS used for the dac. 

 

We should be able to have a near perfect USB dac up to the DAC's output stage with just a simple good quality PS.

 

hmmmm.....

 

 

 

[barrows]

No.  The better the masterclock (specifically the lower the phase noise of the clock) at the DAC chip, the more accurate the conversion.  For any given masterclock, any noise present will degrade its performance, so will vibration.

So best performance will be realized with the best clock, powered by the best power supply, with 0 noise present, and no vibrations, and perfect coupling to the DAC chip's clock input pin.  Everything matters, and of course, there is no such thing as perfection.  All clocks have some phase noise, all power supplies have some noise, every circuit layout has some parasitic problems, and some vibration is always present.  Best performance is realized by reducing all of these factors.

[beerandmusic]

^^^^ Ok, let's leave the ANALOG out of the equation for a bit longer.

 

Assuming a good DAC PS, a good buffer, and an isolated ground from the PC, the signal should be perfect and the same for all DACs up to the DAC chip....so all dacs could potentially sound the same that use the same dac chips, up to the DAC chip.

 

Besides the analog output stage, the difference could/should be only based on the DAC chip or FPGA (which when you think about it, is "house sound").  I am content in not having "house sound" even though the FPGA algorithm may sound "subjectively better" than an off the shelf dac chip.  The ess9038 is such high resolution, that I would probably be more comfortable with it (for myself anyway, even though i like the house sound of the mcintosh amps)....and not willing to pay for some designer's concept and applied algorithm of what sounds good...plus an off the shelf ess9038 should be notably cheaper....

 

That leaves just the analog output stage....

[beerandmusic]

10 minutes ago, barrows said:

No.  The better the masterclock (specifically the lower the phase noise of the clock) at the DAC chip, the more accurate the conversion.  For any given masterclock, any noise present will degrade its performance, so will vibration.

So best performance will be realized with the best clock, powered by the best power supply, with 0 noise present, and no vibrations, and perfect coupling to the DAC chip's clock input pin.  Everything matters, and of course, there is no such thing as perfection.  All clocks have some phase noise, all power supplies have some noise, every circuit layout has some parasitic problems, and some vibration is always present.  Best performance is realized by reducing all of these factors.

 

Ok, I don't disagree with any of what you are saying....i am just living in a hypothetical world at the moment, coupled with factual 1's and 0's....again assuming NO NOISE.

[barrows]

The point is, reducing noise is exactly what high end audio is about, it is not easy, or affordable to do so.  Same thing with clocks, even a $1500 NDK DuCoLon ovenized clock still has significant phase noise.  everything matters, the more attention which is paid to these details, the better performance gets, but it costs more to do so.

Besides these factors, you have the design of the digital filter(s) and the dAC conversion it self is never perfect, and then you have the analog output section which is also far from perfect.

[beerandmusic]

^^^^ Oh poop, i thought we were in the thread about "best source", not realizing we were in thread for qutest....i will stop here....sorry for interrupting with my mental diarrhea....

 

carry on with qutest and my apologies.