Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
my system
#21
(29-Jan-2021, 04:25)bernardl Wrote:
(28-Jan-2021, 16:27)Jean-Marie Wrote:
(24-Jan-2021, 22:36)bernardl Wrote: That assumes that there is no way to best the Roon replay and internal D1000 DAC combination, which isn’t obvious.

I am of course aware that the analog signal coming out of the Esoteric will be converted to digital again in the D1000 but still.

And then there is of course the subjective preference of sound converted one way or another.

Anyway, so far your forecast matches my perception. :-)

Cheers,
Bernard

Bernard,

Yes I am making two assumptions: 
1) Since all 4 options are using the internal DAC of the D1000 (you cannot do otherwise with the Devialet anyway), the simpler way to bring in bit perfect source to this DAC is the better
2) Roon/RAAT is bit perfect

Cheers,
Jean-Marie

Those are reasonable assumptions, 

Out of curiosity, do you think that all DAC sound the same? Also, what are the KPIs you consider relevant to assess the ability of a DAC to do its job?

Cheers,
Bernard

From my perspective, the chain is the following: Low Pass Filter (Anti Aliasing)-> ADC -> DAC -> Low Pass Filter (Reconstruction).

If you are semi-competent in your design and implementation the ADC -> DAC stage is nowadays transparent and the noise floor and distortion is pretty much well below the hearing threshold.

So what remain are the filters and I think they are the most important characteristics of the DACs and the highest contributors to their difference in sounding

The Devialets are a bit of a special situation because DAC is intimately bounded to the amplifier and in fact parts of the amplifier work in the digital domain. Also, I may be wrong, but I believe the amplifier is part of the reconstruction filter, contributing to it's exceptional transparency (at least in my opinion as well as measurements)

Please note that the transparency of the ADC/DAC operation explains why a Devialet accurately propagates the personalities of external DACs that you can put in front of it, even if it will go into the internal ADC and DAC of the Devialet.

So to answer your initial question, the most important characteristic of a DAC is its analog part (the reconstruction filter and the circuitry that will drive its analog outputs). And I'm convinced that any difference of sound between DACs is primarily coming from this part.

Jean-Marie
Reply
#22
If we are getting technical now.....

I recall the Miller Audio Research measurements of the D170.  These indicated a signal to noise ratio of 118dB for the S/PDIF inputs, which is quite good.

The analogue input measured 93.2dB, which is a rather disappointing number in comparison.

Interestingly, the USB input measured at 97.7dB, not massively better than the analogue input.

Paul Miller has not performed equivalent measurements for the Pro, but Stereophile said this about the 140 Pro:

While the individual FFT bins in this graph lie at or below –80dB, it suggests that the 140 Pro's analog inputs are limited to 14-bit resolution.

Then we had this, which indicates that the Pro is indeed not that good via analogue inputs:

https://www.stereo.de/hifi-test/produkt/...-dual-1807

What the above suggests is that no mater how transparent anyone might think the analogue inputs are (and I think they sound pretty good), definitively this is not the best route to getting optimum performance from a Devialet.  14bits of resolution via analogue, whereas the 118dB s/n ratio figure is equivalent to about 20 bits.

One thing I have never seen is equivalent measurements for the CI Board.  Although this is asynchronous, whereas S/PIF is synchronous.  So maybe this is a reason that some of prefer the sound via AES/EBU, using kit like the MC3+USB and similar?  At least, I presume the CI Board is effectively asynchronous, but I am only assuming this because asynchronous is the "A" in AIR.  Does anyone know more about how the CI Board actually works and interfaces with ADH?

Of course, just because both USB and the CI Board are both asynchronous does not mean they would measure or perform the same.  Note the USB input is integrated into the CI board on the Pro, so this will almost certainly measure differently to the original Expert measurements above.

Lots of waffle here, but what I am suggesting is that @bernardl might find that using the MC3+USB + clock might produce some pleasing results (or he might not, of course), and quite possibly their might be a valid technical reason for this.  What is key to using the AES/EBU synchronous route is that the quality of the clock feed matters because this is extracted by the amp.

Per Bernard's post on the proceeding page, both options 1 and 2 are "bit perfect", but maybe some bit perfects are more perfect than others?
1000 Pro - KEF Blade - SOtM sMS-200Ultra Neo - tX-USBultra - Mutec REF10 - MC3+USB - Pro-Ject Signature 12
Reply
#23
Thanks Jean-Marie and confused for these interesting inputs.

So far my listening confirms that the Esoteric makes for a very very good “normal” system while with Raat it’s “abnormally” good. Still the best system I’ve ever heard.

Cheers,
Bernard
Room: Gik Acoustics | Vibration: Townshend pods | Power: Shunyata Omega XC + Everest + Sigma NR v2 + Sigma ground cables | Source: Mojo Audio DejaVu EVO linux server running Roon core (Raat) | Ethernet: Sonore Optical module + Melco S10P with dedicated LPS + Shunyata Omega Ethernet x 2| Synchronous: Mutec MC-3 + USB (Paul Hynes SR7T LPS) + Cybershaft OP21A (Shunyata Omega USB, AES/EBU, clock cables) | Dac/Pre/Amplification: Devialet D1000 Pro CI (Chord Sarum T RCA-RCA link) | Speakers: Chord Sarum T cables + Wilson Benesch Act One Evolution P1
Reply
#24
The Esoteric N-01XD is possibly the best streamer on the market… but it isn’t as good as Air streamed through a Melco S10 with Shunyata Omega Ethernet cables.

And Air isn’t as good as a Mutec MC-3 + USB clocked by a Cybershaft OP21A and powered by a Paul Hynes LPS.

So the morale of the story is that the Esoteric N-01XD has found a new owner who is, rightfully so, very happy with it.
Room: Gik Acoustics | Vibration: Townshend pods | Power: Shunyata Omega XC + Everest + Sigma NR v2 + Sigma ground cables | Source: Mojo Audio DejaVu EVO linux server running Roon core (Raat) | Ethernet: Sonore Optical module + Melco S10P with dedicated LPS + Shunyata Omega Ethernet x 2| Synchronous: Mutec MC-3 + USB (Paul Hynes SR7T LPS) + Cybershaft OP21A (Shunyata Omega USB, AES/EBU, clock cables) | Dac/Pre/Amplification: Devialet D1000 Pro CI (Chord Sarum T RCA-RCA link) | Speakers: Chord Sarum T cables + Wilson Benesch Act One Evolution P1
Reply
#25
Final tweaking for the foreseeable future, I have replaced all the cables connected to the Mutec by Shunyata Omega (USB, AES/EBU and clock cables).

I consider this done. No idea how to improve any of the components in the system at this stage.

And the sound is... not too bad. :-)
Room: Gik Acoustics | Vibration: Townshend pods | Power: Shunyata Omega XC + Everest + Sigma NR v2 + Sigma ground cables | Source: Mojo Audio DejaVu EVO linux server running Roon core (Raat) | Ethernet: Sonore Optical module + Melco S10P with dedicated LPS + Shunyata Omega Ethernet x 2| Synchronous: Mutec MC-3 + USB (Paul Hynes SR7T LPS) + Cybershaft OP21A (Shunyata Omega USB, AES/EBU, clock cables) | Dac/Pre/Amplification: Devialet D1000 Pro CI (Chord Sarum T RCA-RCA link) | Speakers: Chord Sarum T cables + Wilson Benesch Act One Evolution P1
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)